This PDF file contains the front matter associated with SPIE Proceedings Volume 12708, including the Title Page, Copyright information, Table of Contents and Conference Committee lists.

Congestion status evaluation is a vital link and tool in urban traffic congestion management. To timely warn and evaluate the congestion status of urban hot spots, a new congestion situation visualization model based on vector field is proposed. The calculation method of congestion status evaluation index is simplified based on this method. At the model building level, combining the advantages of OD flow diagram and OD matrix diagram, a dynamic visualization model of traffic congestion situation monitoring is established by constructing grid division and extracting action quantities. A method for identifying the trend of traffic congestion characteristics in urban areas is proposed. Simplify the algorithm level, combine the model visualization and the calculation process of the congestion evaluation index, and calculate the average actual travel time according to the vector field data structure, to realize the rapid calculation of the congestion index. The research shows that the model has the ability of trend analysis based on the visual expression of many basic traffic flow parameters such as flow, flow direction, speed, and spatial distribution, and realizes effective identification and early warning of hotspot congestion areas. The simplified calculation results of congestion index Compared with the traditional traffic index algorithm, the average error is 0.016. While the calculation efficiency is optimized, the effectiveness of the regional congestion index is guaranteed. This method can apply the research results to urban traffic congestion monitoring and management.

In order to accurately analyze the causes of accidents in highway reconstruction and expansion project and grasp the law of traffic safety state change, a traffic safety situation analysis method based on dynamic Bayesian network model is proposed. Firstly, on the basis of summarizing the existing research results and expert experience, eight key influencing factors reflecting state evolution were selected, and the dynamic Bayesian network structure was established. Secondly, based on fuzzy evidence theory, an improved mixed interval evidence synthesis rule was designed to solve the negative impact of high conflict evidence on parameter learning. Finally, by using Markov theory and time factor management method, the state transition probability is determined, and the dynamic Bayesian network analysis model of traffic safety situation of highway reconstruction and expansion project is constructed. The empirical analysis is carried out by taking the traffic accident occurred during the construction of a highway reconstruction and expansion project as an example. The results show that the traffic safety situation analysis method based on dynamic Bayesian network model can accurately predict and describe the changes of safety situation with time before and after traffic safety accidents. The analysis of the sensitivity of each factor under different situation levels shows that the traffic composition (the proportion of large vehicles) is the most sensitive. Through THE DIAGNOSIS and analysis of the cause of the traffic accident, it is shown that the main reason of the accident is that the temporary traffic safety facilities are not installed in time, which leads to the great difference of vehicle speed, which is consistent with the actual situation at the scene when the accident occurs. The research results provide a scientific basis for taking the corresponding safety control measures.

In the context of smart city, the propagation loss prediction is essential to predict the strength of signal and interference. The Standard Propagation Model (SPM) is a widely used propagation model in wireless communication. The accuracy and unique of SPM depend on the values of parameters. In this paper, a calibration method for SPM based on the genetic algorithm is proposed and the measurement campaigns are conducted in urban areas in the city of Dalian. The experiment results show that the proposed method is feasible and has high accuracy that reduces the mean error and standard deviation effectively. It is more suitable for measuring signal propagation loss in urban environment than the original unadjusted model.

The safety warning function of intelligent and connected vehicles is of great significance to improve driving safety. Firstly, based on the coverage of application scenarios, this paper builds an application scenario database which based on different sources, with multi-dimensional scene distribution characteristics and a uniform format. Secondly, the application scenario screening model is constructed to determine the application scenario of the safety warning based on connected technology. Then, the application scenarios are classified from different dimensions. Based on the classification method proposed in this paper, considering extensibility and backward compatibility, an safety warning application scenario database is constructed that conforms to China's traffic patterns and reality.

The measured torque signal of the vehicle half-axle is a periodic signal with rotational frequency and harmonics as the main components and interspersed with burr interference. Therefore, it is necessary to remove burrs from the measured signals for noise reduction before transforming them into bench load spectra based on the measured road signals. In this paper, we propose a method based on the improved morphological component and apply it to the burr removal of the measured signals of rotating components to achieve the separation of the actual half-axis load signals from the burr signals. The morphological component analysis based on fast adaptive step iterative shrinkage and P-index threshold noise reduction is proposed to address the drawback of slow convergence and poor noise reduction of the MCA algorithm based on iterative soft threshold shrinkage method. The simulation and the calculation results of a pure electric vehicle half-axle measured torque signal show that the improved morphological component analysis is significantly better than the traditional morphological component analysis in terms of convergence speed, and can effectively separate the burr components in the half-axle load signal.

Under the uncertain conditions and time-varying operating environment, dynamic planning and scheduling have become the key problem, which must be solved to improve the program feasibility, production efficiency, and reducing construction costs during the ship intelligent manufacturing process. A scheduling optimization method based on digital twin is proposed. Firstly, the real-time and dynamic requirements of segment construction scheduling are analyzed, and the technical framework of subsection construction scheduling based on digital twin is established; Secondly, a digital twin model for subsection construction of ships is constructed, and the twin data-driven scheduling method which is based on the IOT system and rescheduling method is established for subsection construction. Finally, through the scheduling process optimizing in the shipyard subsection construction, the effectiveness of the proposed method is proved.

In order to improve the efficiency of warehouse outbound delivery and avoid local congestion of AGVs, this paper considers both strategies of AGVs path optimization and tasks balancing in the warehouse. In this paper, the A * algorithm is improved. Different estimation functions are used in different areas to more accurately depict the distance relationship between each point pair, and efficiently calculate the shortest feasible path from each storage location to each picking station. At the same time, the outbound tasks are planned reasonably. By establishing a 0-1 integer programming model and using Gurobi to solve it, tasks are allocated regularly to each picking station in the warehouse to maintain the workload balance among the picking areas. The experimental results show that the above strategies reduce the maximum working time of each picking area from 1793t seconds to 1511t seconds, and effectively reduce the total mileage of AGVs and the probability of AGV local congestion.

With the rapid development of Intelligent and Connected Vehicles technology, all kinds of cybersecurity risks and problems of intelligent connected vehicles are followed. Systematic protection measures of multiple dimensions are proposed for the main security risks in this scenario. Based on intelligent made car multidimensional cybersecurity defense concept, in-depth analysis of intelligent made car networking security threats and risks, to "risk assessment – cybersecurity protection - test evaluation" as the core, to study and put forward a set of "cloud - communication link - car end" technology protection system is an organic whole, and to test the vehicle key elements, guiding the enterprise application.

Indoor network model as the basis of indoor navigation research, how to automatically generate indoor network model has become a research hotspot in recent years. For the problem that road network model is not fit enough for a complex environment when it is extracted automatically, this paper proposes a method to automatically extract indoor floor plan network model, the indoor space is divided into public space and exclusive space, accordingly forming public space path and exclusive space path, and connecting two kinds of space through the transition points. Finally, the public space path, exclusive space path and connection path constitute the floor plan navigation network model.

With the continuous growth of urban motor vehicle ownership and the increasing complexity of the road traffic environment, dynamic evaluation by analyzing the arrival characteristics of vehicles in the inlet lane of a single intersection is a prerequisite for achieving adaptive control at intersections. This paper establishes a new evaluation system based on the Traffic Smooth Index (SI) by comparing the three indexes of occupancy ratio, speed, and queue length with various weight coefficient determination methods. At the same time, a mathematical set-pair analysis (SPA) model based on the new evaluation system is constructed to reduce or eliminate the subjective influence and establish the five elements' connection number from the same, different, and opposite perspectives. Finally, taking an intersection in the Yanjiao area of Hebei Province as an example, the cluster analysis method and the set pair analysis method were respectively used to distinguish the traffic status level of this section under different time intervals, which proved the superiority of the form and the characteristic that the more subdivided time interval, the more sensitive the changing trend.

Clothing is an important commodity category for e-commerce sites. As the number of consumers and clothing on the e-commerce site continues to grow, the "sparsity" and "cold start" issues of consumer rating data have affected the accuracy of collaborative filtering recommendation algorithms. To solve the above problems, an improved collaborative filtering algorithm is proposed. Based on the classification attributes of clothing, the algorithm weights and combines the clothing category preference similarity and consumer feature similarity to obtain a comprehensive similarity and uses this to conduct the final personalized recommendation. Experiment results show that the algorithm not only optimizes the selection of nearest neighbors, but also alleviates the problem of data sparsity, and achieves good recommendation results.

To improve the operating efficiency of the bus transport system and service level to the public, the scheduling optimization of single-line bus route for express bus and shuttle bus was carried out during the rush hours. Based on the relevant basic assumptions, this paper established the parking judgment model and passenger ride selection model of express bus and shuttle bus. Meanwhile, taking passenger ride experience and bus company operating cost as the objective function, genetic algorithm was used to optimize the bus departure interval. In the optimization process, two schemes including discrete scheme and continuous scheme were provided. The models and optimization methods were applied to an actual single line bus route. The results show that, the discrete scheme has faster convergence speed and better optimization effect. The value of objective function decreases by 12.38%. Although the bus operation cost has increased slightly, considering the public commonweal of the bus transport system, the optimization results are still in line with the essence of its operation. The models and its optimization methods can provide a reference for the scheduling optimization of single-line bus route with the introduction of express bus and shuttle bus.

The management of Internet of Things (IoT) devices is becoming increasingly complex. One of the reasons is that IoT device manufacturers are different, and there are different degrees of heterogeneity in service, technology, protocol and other aspects. Accurate identification IoT devices connected to the organization network is an effective way to maintain the organization network security. In this paper, we propose a two-stage machine learning method to identify IoT devices by analyzing network traffic. This method uses the futures of the original traffic advance to classify the types of IoT devices. The method is tested on two public datasets, our method classified IoT devices with an accuracy rate of over 99%.

Compared with traditional network modelling methods, hypergraph has superiority in simplifying the structure of multi-modal network and improving the efficiency of traffic assignment and can solve the common line problem in public transit network without virtual nodes. In this study, the modelling of multi-modal hypergraph is provided. A new storage structure based on hyperedge list is put forward, which reduces the storage space of network significantly. An effective path search algorithm suitable for hypergraph is designed. A numerical experiment proves method feasibility. This study provides a meaningful supplement and promotion for the theory of transportation network modelling and assignment.

BIM MEP models are the data support for visualising building information in "smart cities". To address the key visualization issues such as inefficient data scheduling of BIM MEP models, this paper proposes a data organization method for BIM MEP models that combines instantiation technology. The method firstly generates instance groupings based on the instance information of the electromechanical model components, secondly spatially divides and hierarchically clusters the massive component models by using KD trees according to the spatial distribution characteristics of the BIM electromechanical models, and finally combines with instantiation techniques to batch render the BIM electromechanical models. The experimental results show that the framerate of the optimised BIM MEP model is about 50 fps, and the lowest framerate is stable at about 30 fps, which can better ensure the visualization effects of the massive MEP model.

To meet the actual operational needs of public security traffic control departments, we propose an intelligent construction method for traffic control scenarios and introduce its construction process and application value. In this study, traffic control scenarios are split into several constituent elements. The associated elements are found and combined into scenarios based on actual data. To meet the computational efficiency challenges due to the rapid update of massive traffic violation data in the real world, this study applies ordering points to identify the clustering structure algorithm to achieve fast and efficient clustering of latitude and longitude data. Furthermore, the frequent pattern growth algorithm is improved by splitting the database and node exchange based on a Hash table to improve the association rule mining efficiency. The results reveal that both algorithms significantly improve efficiency and accuracy compared with traditional algorithms. By obtaining the refined influence range of traffic violation control scenarios through the clustering method, correlating the environmental factors of scenarios with control behaviors through association rules, and analyzing the intrinsic causes of scenarios based on identified indicators, the efficiency and accuracy of public security traffic control departments in finding and solving problems are effectively improved.

In order to maximize the operational efficiency of intersections, this paper first establishes a multi-objective function model with the maximum traffic capacity, minimum average number of stops, and minimum intersection delay as the optimization goals by studying the traffic data of instance intersections and their change laws. Then, the multi-objective optimization model is solved by heuristic algorithms such as the genetic algorithm and particle swarm algorithm in Python language, and the above algorithms are compared and analyzed to find the optimal intersection signal control scheme. Finally, taking the intersection of Zhongyang East Road and Yijing Street in Siping City as an example, the genetic algorithm and particle swarm algorithm reduced the average delay of the intersection by 32.3% and 31.4%, and the average number of stops decreased by 4.8% and 6.0%, respectively. The results show that the signal control scheme optimized based on the heuristic algorithm can reduce the delay level and average number of stops at the existing intersections, which proves the feasibility of the proposed heuristic algorithm in this paper.

In order to realize the background of large-scale data analysis and control of power equipment development based on Internet of Things, a design idea of power equipment management system and large-scale data system analysis is proposed. This paper firstly introduces the design of power industry control and control system application. The open source electronic information service platform includes customer and interactive terminal. Designing an architecture design system. Using mathematic model and analytic method to analyze, analyze and determine the major data of power supply enterprise. From the conclusion of data analysis, with the system management theory and scientific management method, the inherent regularity of enterprise management and management, the operating process and management result are revealed, which widens the evaluation in the area of energy service, and there is plenty of room for improvement.

Internet of Things (IoT)has numerous applications in the industry and society, thanks to its ability to achieve automation and connectivity in a range of activities. Despite its great potentials, IoT is susceptible to physical and cyber-attacks, which causes security threats (e.g., financial risk and leakage of privacy). To address this problem, an approach for attack prediction is proposed for IoT. Aiming at a high degree of flexibility, an intelligent model is designed to construct knowledge graph by integrating equipment information CPE, vulnerability information CVE and attack pattern information CAPEC disclosed by the National Institute of Standards and Technology (NIST) and the security organization MITRE. Based on the knowledge graph, the safety analysis and operation analysis of many IOT information are carried out. To conclude the possible attack, knowledge representation learning method that fuses the triple information and semantic path combination information of the knowledge graph (FTSPC) was employed. We transform the attack prediction task into the link prediction problem. The suggested method is evaluated on a public dataset and our dataset, the results demonstrated that the method can predict the attack of IoT infrastructure, providing rich IoT security knowledge to security researchers and professionals and a useful reference for active defense.

Obtaining the traffic operation state of the road network can help the traffic management department to optimize management measures, provide reference for travelers on travel modes and travel routes, effectively alleviate the congestion situation and avoid the vicious circle caused by blindly building more roads. Floating car method has become an important method for intelligent transportation system to collect road network operation information, and the collected data has the advantages of large quantity, wide coverage, low cost and great representativeness. This paper starts with GPS data collected by taxis in Dongguan, China, extracts traffic parameters from them, and uses clustering algorithm to identify traffic conditions. Finally, according to the correlation analysis method, the temporal and spatial variation law of the running state of each section in the road network is mined, which provides suggestions for improving traffic conditions.

As the railway network structure becomes more complex and capacity resources become tighter, risk factors such as equipment disruption, bad weather and human interference pose greater challenges to the on-time operation of high-speed trains. The study of the correlation between primary delay and risk factors plays an important role in improving the punctuality of trains and real-time dispatching command. Firstly, the disruptions were divided into 13 categories; secondly, a Bayesian network structure was constructed based on expert experience, and the structure was modified using a greedy thick thinning algorithm, and the expectation maximization algorithm was used for parameter learning. Finally, the TF-IDF algorithm was used to complete the keyword extraction of the text data of the emergencies that triggered delayed trains from 2016-2019, and the structured and processed 0-1 matrix data were fed into a Bayesian network for inference studies using the joint tree algorithm. The results show that the more likely types of disruptions when trains have a primary delay are, respectively, foreign object strikes or foreign invasion, contact network fault, track circuit fault, shaking or noise. The results can provide useful and valuable information for the optimization of inspection and maintenance schemes for key disruptions.

Traditional signal control at intersections is mainly based on the maximum capacity as an indicator, which leads to long signal cycles at many single intersections, but instead aggravates the average delay, stopping frequency rate, and tailpipe emissions of vehicles. From the perspective of improving intersection efficiency and environmental protection, the paper establishes a multi-objective function model with the average delay, stopping frequency, capacity and exhaust emission of vehicles as the optimization indicators, and proposes an improved genetic particle swarm algorithm (GAPSO) solution for intersection timing optimization. The solved results are compared with the original timing scheme, the GA and PSO solved timing schemes, in conjunction with real-life case studies. The results show a significant optimization in terms of average vehicle delays, exhaust emissions and stopping frequency, thus demonstrating the effectiveness of the improved GAPSO for timing optimization at single intersections.

In order to meet the needs of the visually impaired to learn Braille and improve the efficiency of teachers teaching Braille, this paper uses Internet of Things technology and embedded technology to design a new type of Braille learning machine to optimize the existing Braille teaching. The Braille teaching system is designed through the MQTT protocol to realize the network control of multiple Braille learning machines. A micro motor is used to control the lifting of a single braille bump, and a dual CPU control architecture is used to control the lifting of the braille bump, which improves its reusability compared with traditional braille boards. We tested the accuracy of Braille conversion, and the experimental results were in line with expectations. This research can better serve the visually impaired and bring convenience to their Braille learning.

As the world is undergoing rapid changes, the IoT device products and corresponding technologies are also developing rapidly. Therefore, security control of IoT devices should be carried out in order to unify the management of the vulnerabilities that exist. While the current mainstream research side of IoT device identification is to identify and predict devices using http messages, few studies and discussions have been conducted on IoT devices using FTP messages. A clustering tree based FTP message recognition method is proposed here, which first automatically extracts the keywords used for recognition in the messages, and then generates a binary clustering tree with training and test samples and calculates the similarity degree. The experimental results indicate that the method achieves high performance measures for identification, including 99.5% accuracy, 99.6% precision, 99.5% recall, and 99.5% F1 score.

With the continuous development of 5G technology and big data processing technology, the Internet of Things has been applied to every aspect of human life. At the same time, due to the deployment of a large number of intelligent terminal equipment, massive Internet of Things traffic is generated and sent to the cloud platform storage. In the Internet of Things network environment, there are many high-dimensional, complex data characteristics. These characteristics lead to the unsatisfactory application effect of traditional intrusion detection technology in IoT security protection. Therefore, establishing a complete IoT intrusion detection system has become essential to ensuring IoT security. Since deep learning has powerful data processing and feature learning capabilities, this paper uses deep learning technology to build the Internet of Things intrusion detection model to provide the regular operation of the Internet of Things. The main research work and innovation are aimed at the problems that the sparse stack encoder (SSAE) model is challenging to learn practical features and slow convergence due to its low sparsity in dimensionality reduction of high-dimensional data in the Internet of Things. This paper improves the SSAE model and proposes a BR-SSAE model.

It has become mainstream in the industry to introduce a variety of power distribution intelligent sensing devices into the traditional power network, and collect information in real time to provide support for the intelligent monitoring platform. Such an Internet of Things is called power distribution intelligent network. This paper proposes an Identity authentication scheme for power distribution network device using dynamic tokens and identity-based cryptograph. Its advantage is that it does not require a real-time online public key management system, and terminal devices can complete identity authentication, secret key negotiation and encrypted communication by themselves. The analysis and simulation experiments of this scheme show that this scheme has good performance and can adapt to the terminal access scenario of power distribution equipment.

During vehicle driving, many problems will occur in the communication process of VANET, including but not limited to node conflict management, time delay, time slot division management, etc. The Media Access Control (MAC) protocol is specifically responsible for solving these problems. The existing MAC protocol with fixed time slot allocation cannot solve the problems of excessive node conflicts and low packet reception rate when current traffic is high. When designing MAC protocol for VANET, we must consider the speed and dynamic distribution of nodes, which is also the key to optimize the protocol. In this paper, we propose a new TDMA protocol, called SMDP-MAC (Adaptive Slot Management MAC Protocol Based on Dynamic Parameters). This protocol is a new protocol that divides time slots according to vehicle dynamic parameters to solve conflict problems. SMDP-MAC manages the time slot in the vehicle channel access and maintenance phase according to the vehicle's GPS information, driving direction, speed and acceleration to reduce the collision probability. In the SMDP-MAC protocol, the frame is divided into multiple timeslots based on TDMA, the timeslots in the frame are divided into two timeslot sets according to the direction of travel, and then the channel access timeslot is divided and selected according to the dynamic parameters of each vehicle (including GPS information, vehicle speed and vehicle acceleration information). After the network enters a stable state, the time slot occupation of the vehicle will also be dynamically adjusted according to the vehicle position.This paper presents the analysis and simulation results under the highway scenario and verifies the effectiveness of SMDP-MAC protocol by comparing with VeMAC, BMA and ASMAC protocols.

Ultra-wideband (UWB) has strong anti-interference capability, low power consumption and competitive price advantage, making it widely used in short-range Internet of Things(IoT) applications, particularly for positioning. However, there are still many problems in practical applications, such as clock synchronization between base stations and non-line of sight (NLOS). To solve the problem that NLOS errors caused by the complex indoor environment affects the positioning accuracy, this paper proposes a Kalman algorithm-assisted Taylor algorithm for indoor positioning. Firstly, the distance between the base station and the tag is measured by Symmetrical Double-Sided Two Way Ranging (SDS-TWR), which is filtered by Kalman filter to eliminate some of the errors, and the preliminary position estimation result is calculated as the initial value. Taylor algorithm uses the initial values calculation to achieve precise positioning. Results show that the method can effectively eliminate the interference of indoor complex environment to the positioning system and improve the positioning accuracy.

Group key agreement technology is one of the most popular solution to establish a reliable secure channel between communicating entities of a group. Group key agreement protocol with administrator (GKAA) can handle a large number of group operations in sequence to resolve disorder problem and it is widely used in practical applications. However, in most of existing protocol, the administrator has to compute materials for each participant in the group which brings huge calculate overhead. In this paper, we propose an efficient group key agreement protocol with Administrator. Our GKAA protocol is able to perform group operation in a high efficacy way with the help of a key tree. It is also flexible and scalable since it suits for any network communication environment like IoT, video meeting, social applications, etc. Our protocol satisfies the properties including forward secrecy, PCS and key control resistance.

Classification, segmentation, and detection are the most important tasks in computer vision, and target detection as one of them is a hot research topic in the field of computer vision, which is widely used in medical, traffic, surveillance, etc. YOLOv4 and R-CNN have excellent target detection performance, and an improved YOLOv4 target detection algorithm is proposed to improve the real-time detection of small targets for target recognition. A priori frames are designed using the K-means clustering algorithm for adapting to different small and medium sizes; a feature layer is extracted according to the size of small and medium-sized labeled objects and four different feature layers are fused for detection; the Mish activation function is applied to the neck of the detection model to improve the detection performance. The experimental results show that the improved algorithm can effectively improve the detection accuracy.

With the gradual improvement of people's demand for quality of life, the smart home industry is developing rapidly. A large number of smart home devices concentrate a large amount of data, so smart home systems mostly use data cloud outsourcing to manage data. Meanwhile, the centralized cloud servers used by traditional smart home solutions bring single-point failure problems and data privacy protection risks. To solve the above issues, we propose a blockchain-based public audit solution for smart home data integrity. The scheme uses blockchain and IPFS instead of centralized cloud servers, and supports dynamic attributes of data in cloud servers. Further, we extend the bulk auditing functionality. Finally, the security analysis and performance analysis show that our scheme outperforms existing scheme, and the audit computation overhead is reduced by 38%. When the number of users of the batch audit function is greater than 10, the average audit computation overhead per user is significantly reduced, which can effectively improve audit efficiency.

Driven by the development of new generation information technologies such as the Internet of things, mobile Internet and big data, the development of smart home has also made unprecedented progress. Especially since the Ministry of industry and information put forward the action plan for the development of "double Gigabit" network collaborative plan, the smart home has attracted the attention of the industry and promoted the maturity of terminal equipment. As an indispensable role in the smart home networking, the home gateway is responsible for information interaction, management control, wireless routing and other functions. Therefore, increasing its security in all aspects is the premise of protecting users' privacy information, ensuring family safety and providing more comprehensive services to users,the security of the home gateway is particularly important. The home network is the closest network to people's lives, and ensuring the security of the home gateway is the last line of defense to protect everyone's privacy. It is urgent to attach importance to and strengthen the research on gateway security.

This paper expounds on the concept of the Internet of Vehicles and studies the system architecture and popular technologies of the current Internet of Vehicles. The software section will present specific iterations of Wireless Communication Technology and the hardware section of Sensor platform will introduce the latest achievements in Vehicle camera and Vehicle radar. Based on this, innovative ideas are presented and simulated accordingly. In Wireless Communication, a PID algorithm is proposed to improve the power of the signal during transmission. In Sensor, an edge detection function is proposed to analyses the sound signal using the FFT algorithm. Both parts are simulated and shown to be well suited for use in the field of connected vehicles. Finally, by referring to the research status in the field of Internet of Vehicles, suggestions are put forward for the development direction of the Internet of Vehicles.

This paper investigates the recent progress of Locality-sensitive hashing (LSH) under different metrics and its adaptive performance in different application scenarios. LSH is an approximate nearest neighbor query algorithm, which aims to perform fast similarity finding in high-dimensional space. In recent years, the improvement directions of LSH can be divided into three categories: 1) constructing suitable function families under different metric indexes, 2) constructing detection sequences by perturbation, and 3) expanding the radius to improve the retrieval range. Since metric indexing is more widely used, this paper focuses on a comprehensive and schematic review of this approach. computational simplicity and efficiency of LSH, which is prominent in the fields of image, recommendation, and de-duplication, we review the latest research on LSH in different application scenarios to provide an advanced general framework and a better understanding of LSH in combination with other fields. In conclusion, various algorithms for LSH are evolving rapidly, and it is expected to see more diverse application scenarios for these methods in the future.

Studying the rational allocation of urban space and public transportation resources to meet the daily activity needs of commuters is a major topic to solve urban traffic problems for a long time. In response to the above problems, this paper designs a dynamic path planning algorithm for urban residents' public transportation aiming at the shortest commuting time. First, the shortest path between residential nodes and office nodes via bus lines is calculated through the shortest path algorithm; then the optimal bus ride scheme in the shortest path is calculated through the dynamic programming algorithm. The dynamic programming algorithm in this paper simulates the public transportation commuting of residents in the main urban area of Beijing. The simulation results show that the path planning algorithm can provide path planning for residents in the main urban area of Beijing to commute using public transportation. The path planning results include the bus stops and bus lines that residents pass through. In addition, the simulation experiment also obtained that the average commute time of the road network used by Beijing residents using public transportation is 47.24min, and the average waiting time of the road network is 19.44min, which can clearly and intuitively represent the status of the public transportation road network in the main urban area of Beijing.

In this work, we propose a Multi-authority (, ) t n Threshold Identity-based Matchmaking Encryption (MAT-IBME) scheme under the standard assumptions in the standard model, in which multiple authorities share the secrets for constructing users’ secret keys. The MAT-IBME scheme satisfies the requirement that the receiver can only get the decryption key with the "joint permission" of t ones among n authorities. Meanwhile, only when the identity of the sender satisfies the policy proposed by the receiver, and the identity of the receiver satisfies the policy proposed by the sender, the receiver can decrypt correctly. In this way, the scheme realizes precise ciphertext access control when deployed in cloud services. Security and performance analysis shows that MAT-IBME not only achieves authenticity security but also system-level robustness when no less than t authorities are alive in the system.

In order to protect family privacy, personal safety and property safety stably and efficiently, an application method of Lora technology in smart home of Internet of Things is proposed. Based on the overview of Internet of Things technology and smart home, this paper deeply explores the practical application of Internet of Things technology in smart home, and designs and implements an Internet of Things smart home security system by using LoRa communication technology. The results show that the system has good structural objectives and simple operation, which meets the needs of normal families. It will pave the way for the future family interconnection, the integration of urban Internet of Things and the construction of smart cities.

In recent years, China's electric power business has been booming, the scale of material procurement of electric power enterprises has been growing rapidly, and the intensity and complexity of electric power material management work have also increased significantly. In order to improve the overall level of material management of electric power enterprises, we should plan for the long term, actively introduce advanced technology and pursue high-quality development. This paper will discuss the feasibility of applying blockchain technology in the field of material management of electric power enterprises, and then propose a design scheme for a new electric power material management system based on blockchain technology for the industry of electric power material management to make reference to.

Based on the road network development of the urban agglomeration of Greater Bay Area of Guangdong, Hong Kong and Macao (GBAGHM) from 2014 to 2020, this paper focuses on evaluating and characterizing the spatial-temporal evolution and differentiation characteristics of road network density and its relationship with the urban form development, with the integration of the data of point of interest (POI), population, land, and economy, as well as the support of geographical information system (GIS). The result indicates that the weighted kernel density of road network in GBAGHM has significant characteristics of spatial clustering and regional differences. The urban core areas of the major cities of Guangzhou, Foshan and Shenzhen abnormal high aggregation of road network density. Furthermore, the spatial distribution pattern of road network is closely related to the spatial aggregation characteristics of urban form.

In this paper, according to the empirical trajectory data obtained from Yingtian Avenue in Nanjing, verifies whether the chain asymmetric behavior model applies to Chinese roads and the parameter adjustment when applied to Chinese expressways. The empirical trajectory proves that the basic assumptions of the chain asymmetric model are consistent with the actual situation. The authors also adjust the parameters of the model according to the data.

The merging behavior of expressways is regarded as one of the main sources of traffic congestion and delays on expressways and surrounding road networks. As the road narrows, the number of lanes decreases, the traffic capacity decreases, and vehicle conflicts increase. Comprehension of the joint characteristics between driving behavior and traffic conflict can help traffic engineers better estimate the characteristics of traffic flow. Therefore, this paper uses the idea of zoning modeling to evaluate the difference of traffic conflict in the merging area, describes the relationship between oscillation waves and traffic conflict, and puts forward a theoretical model to distinguish the impact of vehicle behavior on traffic safety when traffic oscillation occurs through the microscopic analysis of the influence factors of oscillation wave on traffic conflict.

The global population of visually impaired people is growing, but the guide rods that serve their daily travels are not intelligent enough. In this regard, this paper designs four core functions of intelligent guide rod, including voice obstacle avoidance reminder, turning voice prompt, night travel warning and GPS emergency help; Then, based on the single-chip Arduino development board and the Internet of Things sensing technology, the obstacle avoidance, steering and night warning modules are developed by using ultrasonic sensors, photosensitive sensors, Internet of Things modules and GPS modules. The GPS positioning in emergency situations is realized, and the location information is sent to the guardian mobile phone APP through the Internet of Things module; Finally, the prototype is tested in the practical scene of coffee shop to verify the usability and practicability of the guide rod.

In this paper, a complete simulation model of steering system of nose landing gear and towbarless traction is established based on LMS Virtual.Lab Motion. And the B737-800 aircraft was taken as the research object to simulate the steering condition of the towbarless traction. The load of the steering system of the nose gear are simulated and analyzed under various steering conditions. The simulation results show the influence of the changes in the steering angle, steering rate and traction rate on the steering system, and provide some reference for further research on the safety of towbarless traction.

Building pattern informs urban spatial structure understanding and modeling. However, previous studies showed limitations on the identification of building groups which has complex spatial distribution. Specifically, they usually use original spatial or non-spatial characteristics but omit certain complementary among multi-faceted features of buildings. In this paper, we give a novel method based on a multi-view clustering framework, which establishes a centroid distance view and a building attribute view to recognize building patterns accurately. The two views are based on both the spatial structure and non-spatial attributes. The similarity graphs on these views, which are obtained by the Gaussian kernel function, are executed graph diffusion and fusion process to obtain a unified graph. Then, the clustering results are obtained via the Normalized cut algorithm. We conduct experiments to recognize building patterns by four real-world community building footprint datasets of two cities in China: Wuhan and Chengdu. The experimental results show that our proposed method identifies building groups with grid, grid-like and unstructured patterns more effectively. Our method performs better in seven evaluation indexes compared with the baseline model.

This paper introduces a data transmission and processing system for the speed governor test-bed. And the design employed the Internet of things (IOT) technology, added the WIFI communication module on the traditional speed governor test-bed, as well as combined the database and WEB server, which enabled the data measured by the speed governor test-bed to be stored in the database in real-time. In addition, the experimenter can remotely retrieve and view historical experimental data and view the direct results of data processing. It solved the issue that the data and information in the process of governor testing can be transferred and transmitted efficiently and reliably among all levels and layers, and accordingly enhanced the ability of data storage and fast retrieval in the process of governor testing.

Aiming at the problem of position deviation and low precision of quadrotor UAV in fixed-point hovering in indoor environment, this paper analyzes the flight principle of quadrotor UAV, establishes a quadrotor UAV system model, and simplifies and linearizes the model. At the same time, according to the quadrotor UAV model, the PID control technology is used to design the attitude controller and position controller, and the feedback control system is realized. Finally, the software and hardware platform of the quadrotor UAV is built, and the PID control algorithm is realized. The experimental results show that the attitude angle error is about ± 3 °, and the position error is about ± 2.3cm/s, which achieves a good hovering effect and solves the problem of low indoor fixed-point hovering accuracy of the quadrotor UAV.

With the rapid development of information technology, the internal decision-making mode of enterprises has been greatly affected. Intelligent decision support system (IDSS) is a system supported by information science and modern management science, which can help executives make better decisions. Starting from data support and technical support, this article discusses the support framework of intelligent decision support system. After that, the article analyzes the composition and partial application of its constituent system. On this basis, the mechanism and implementation path of decision support system are introduced, and a digital and intelligent decision support platform is designed. This will provide corresponding reference for enterprises to build intelligent decision support system.

Based on building information modeling (BIM) plus Robotic process automation (RPA) technology is rapid development in smart building operations and maintenance (O&M). RPA is a workflow-based automated execution tool that is particularly well suited for replacing the manual execution of simple and repetitive tasks using computer systems. The use of RPA technology can save man-hours and increases productivity, which is important for the automation of building operations and maintenance operations. BIM technology, as an advanced technology in the construction industry, can provide new ideas and methods for building operation and maintenance based on its visualization characteristics. That’s integrating RPA with BIM technology which will enhance the digitalization and virtualization of smart building operation and management. This article will first introduce the features of RPA and BIM technology and then discuss the technical roadmap for remote automated operation and maintenance of smart building equipment rooms to implement a low-cost and highly efficient building equipment room operation and maintenance solution

Strain is an important parameter that reflects the safety state of bridge structure, at present, distributed strain measurement is one of the research hotspots in the field of structural health monitoring (SHM). The sensors such as strain gauge, vibrational chord strain gauge and fiber Bragg grating (FBG) strain gauge is difficult to network and realize distributed measurement. The optical fiber sensing technology based on Brillouin scattering can realize distributed strain measurement, but it’s always low in the signal-to-noise ratio (SNR) and measurement accuracy, so the application is limited. In this paper, a distributed strain measurement method for bridge based on weak FBG (WFBG) array is proposed and verified by experiments. The continuous beam test results show that, the good sensing performance is demonstrated by grating array strain cable which can accurately measure the strain distribution of continuous beam, and the load point can be located by the strain cable, and its positioning accuracy is 0.5m, and the sensitivity of the cable is about 1.25με/pm. The test results of real bridge show that strain distribution of the whole bridge under dynamic and static load can be measured, and the maximum deviation of dynamic and static measurements is 6.8% and 5.6%. The temperature compensation method is proposed for the temperature influence in the test process, and the effect is good. In conclusion, the feasibility of the distributed strain measurement method of bridge based on weak FBG sensing has been verified.

In view of the low level of intelligence in monitoring and controlling the granary environment by manual and semi-automatic methods in the process of grain storage in China, an intelligent granary monitoring system based on OneNET cloud platform is designed for real-time remote monitoring and intelligent control of granary. The system takes STM32 chip as the intelligent node, and monitors the temperature and humidity, light intensity, CO₂ and other environments of the granary in real time. After the collected information is fused by weighted average method, it intelligently controls the opening and closing of insulation doors and windows. Through the wireless WiFi module ESP8266, the collected granary data is transmitted to OneNET cloud platform in real time, so that the granary administrator can remotely monitor the granary. The system improves the intelligent level of granary environmental information monitoring, ensures the quality of granary storage, and has certain popularization and application prospects.

In order to alleviate the increasingly serious urban traffic congestion, a signal control scheme based on the characteristics of congestion development was proposed. The traffic flow prediction model is established based on the Cellular Automata (CA) model. The micro change law of the road network traffic flow was described, and the development characteristics of traffic congestion were predicted in real time. Signal control strategies under different congestion characteristics were obtained after research to prevent traffic congestion. The road network formed by the intersection of Xiyuan Road and Chang'an Road in Luoyang and the surrounding associated intersections was taken as a case, and the Simulink platform was used to build a simulation environment. The signal control scheme in this paper was simulated together with fixed signal control and self-organizing signal control. The results show that: after adopting this scheme, the traffic operation indicators of the road network are significantly improved; The development of congestion is effectively controlled; The ability to balance traffic load is significantly improved.

Conventional SLAM only requires the construction of sparse maps for localization, while in order to meet the safe driving needs of unmanned vehicles, they need to understand the edges of the road, i.e., with a semantic level of understanding. In addition, unmanned vehicles are more sensitive to lateral errors than longitudinal errors, which requires SLAM algorithms with higher accuracy for lateral errors. We investigate the ORB-SLAM3 algorithm by introducing satellite maps as a priori knowledge, using the corners in satellite maps to initialize the odometer, remove the accumulated errors, and correct the previous positions; using the results of particle filtering and lane line identification to further optimize the localization results of ORB-SLAM3 and to draw maps with lane semantic information. Our experiments show that our algorithm significantly reduces the cumulative error without loopback, improves the localization accuracy, and yields lane line maps with large engineering applications.

Because the existing methods ignore the influence of the boundary characteristics of the disturbance stability region, the early warning effect is not good. For this reason, an intelligent early warning system for power theft in distribution network based on Internet of Things technology is proposed. The hardware part of the system is composed of monitoring module, human-machine monitoring module and early warning display module, which eliminates interference data and noise data and improves the accuracy of finding abnormal grids. In the software part of the system, the boundary characteristics of the stability region of the electric theft risk are extracted, and the big data parameter fusion model of the risk assessment is established to achieve the quantitative assessment and decision-making of the theft risk. Finally, the improved equivalent resistance method is used to calculate the theoretical line loss of the distribution line characteristics, analyze the fluctuation of the parallel line loss, and then compare the time coincidence of the line loss and the change point of the current difference curve with Pearson correlation coefficient to realize the identification of the power theft risk, further improving the accuracy of the early warning. The experimental results show that when the method is applied to the early warning of electricity theft, the MAPE value of the system is less than 0.1%, which has a better early warning accuracy.

In order to help busy office workers take better care of pets, this paper designs a new intelligent pet feeder system. The system is mainly composed of embedded microcontroller circuit, drive control module, WiFi communication module, human-computer interaction module and video stream compression transmission module. The core chip of STM32f407ZGT6 is used to control the pet feeder for timing and quantitative feeding, and real-time audio and video and pet remote interaction are used to effectively understand the pet's emotion, so as to realize the pet's emotional comfort and emotional communication; The feeder is connected with the router, and the mobile phone, mobile terminal and computer are all communicated and controlled through the router and the feeder; It supports the local control of the feeding device by keys and touch screen, and also supports the remote control of the pet feeding device by mobile APP and web page, which makes the user's operation more convenient.

With the implementation of the "double carbon" plan, a large number of wind turbines, photovoltaic generators and energy storage devices are connected to the distribution network (DN), posing new challenges to the stability and economic operation of the DN. The uncertainty of wind power generation and photovoltaic power generation puts forward higher requirements for dispatching control of DN. The addition of energy storage devices can restrain the fluctuation of wind and solar output, but considering its high cost, it is necessary to reasonably use energy storage resources. In order to reasonably dispatch the resources in the DN and reduce the economic cost and network loss, a DN optimal dispatch method based on improved Harris Hawk algorithm (IHHO) is proposed. Firstly, Sobol sequence is used to form uniformly distributed initial population (IP) position; Secondly, the nonlinear energy factor is used to make the algorithm still have the ability to jump out of the local optimum in the later stage; Finally, the Newton iteration method is used to enhance the local search energy in the development stage. The example analysis shows that the optimal control algorithm can effectively reduce the network loss and operation cost of the DN on the basis of the safe operation of the DN.

In order to improve the intelligent and secure management of community access control, a design scheme of face access control system based on SSD and OneNET cloud platform is proposed. The hardware system is mainly composed of the main control module of Raspberry Pi, human infrared detection sensor, RFID card reader, Web server and OneNET cloud platform. Support multiple door opening methods: face door opening, RFID card door opening and web remote door opening. Provide a Web background management system and APP mobile terminal user system for property management and residents. Face detection adopts a simple and convenient SSD framework for training. Select MobileNet-V3 as the backbone network and train a lightweight model suitable for running on mobile devices. All access control devices are connected to the OneNET cloud platform through the Raspberry wireless WiFi module, which is used for device access, device control and unified management of devices.

The orderly control strategy of the optical storage charging station can significantly reduce the impact of the disordered charging of many vehicles on the power grid. However, users' response to orderly charging will cause inconvenience to some extent. Therefore, this paper proposes a coordinated control strategy of optical storage charging station based on peak-valley period, taking the minimum total load of charging station and the minimum charging cost of users within the user charging period as the objective function, and adopts the bee colony algorithm to comprehensively analyze the impact of different peak-valley period division and different user responsiveness on the power grid impact. The effectiveness of this strategy is verified by an example analysis. This method significantly reduces the impact of the optical storage charging station on the power grid.

Accurate interaction strength prediction plays an important role in urban planning and development. Based on the existing national interaction intensity data between cities, the research uses GCN model, ChebNet model and GAT model based on graph convolutional neural network to realize the prediction of urban interaction intensity, and determines the optimal model through accuracy evaluation. Further we introduce geographical features, optimize the optimal GAT model, and finally the graph neural network model considering geographical characteristics is constructed to realize real-time, efficient, and long-term prediction of urban interaction intensity. It is found that there is a strong interaction betweend cities.

This design can not only meet the users 'demand for parking location, but also facilitate the administrator's management of the parking lot. The design purpose is to build the whole parking management system. At the user side, take the MCU as the core control unit, and the parking space detection module, alarm module, communication module and indicator light display module are combined. The management side adopts programming to design a parking lot administrator operating system to meet the related functions of parking management, charging timing, parking query, and finally realizes the intelligent parking lot management system that meets the users and management at the same time. The design effectively avoids the traffic congestion in the parking lot, greatly promotes the parking operation of the car owners, and meets the needs of the managers for the efficiency, performance and management of the parking lot. This design is compared with the relevant scheme, has the unique characteristics of simple operation, intelligence and convenience, and the design is comprehensive, not limited to a single level of users or management, can provide reference for researchers engaged in intelligent parking management in the future.

While the electric vehicle industry is developing rapidly in China, the problem of "difficult charging" is becoming more and more prominent and has gradually become a "bottleneck" restricting its development. In view of the current situation of "difficult charging", this project researches an electric vehicle charging scheduling method based on road conditions, based on which an intelligent charging algorithm based on the charging vehicles under different road conditions is researched, in view of the long charging time of electric vehicles and the small number and uneven spatial distribution of charging piles in urban areas. grid-based dynamic scheduling algorithm for electric vehicles in order to improve the scheduling problem of electric vehicles.

With the rapid development of China's road system, road maintenance has become an important issue facing road development, and road surface defect detection is the primary link in road maintenance. In order to address the most challenging crack detection in road surface defect detection, larger convolution kernels were used in YOLOv5, which have larger receptive fields and can obtain more crack feature information. A large convolution kernel structure was also used, in which structural re-parameterization was applied to improve the detection accuracy of the model without increasing the detection speed. Furthermore, in order to further improve the detection speed of the model, depthwise separable convolution was applied to the large convolution kernel structure at the expense of sacrificing some accuracy.

In the face of the fierce market competition in the current transportation industry, high-speed railway stations, as an important part of the high-speed rail system, should continuously improve to meet the growing demand for high-quality services from passengers. From the perspective of passenger satisfaction evaluation of high-speed railway stations, a three-level evaluation index system is constructed. Based on this, a high-speed railway station passenger satisfaction evaluation method based on Bayesian network is proposed. The Bayesian network structure is learned based on survey data to construct a Bayesian network corresponding to the evaluation index system, and the passenger satisfaction comprehensive score of the station is obtained through Bayesian inference using the inference calculation rules between different level indicators. Taking Beijingnan Railway Station as an example, empirical research is conducted, and the score of Beijingnan Railway Station is 7.11, indicating a good level of passenger satisfaction.

Making use of the OpenGL graphics standards and technology of Virtual Reality, facing the demand of the forest management users, this paper described the principle and method for forest scene synthesis technology and put forward the design mentality of a virtual forest scene system. The virtual forest scene synthesis-editing system was developed with open source skeleton animation engine Cal3D, parametric tree modeling, quadtree technology, which was an integration of two and three dimension forest landscape simulation. The system provided many functions such as forest asset statistics, query and analysis, three dimention scene immersion roaming, entity model editing, garden planning and design and so on. At last, this synthesis-editing system was applied to construct complex sika deer habitat forest landscape, combined with multi-elements (digital terrain, animal behavior, tree, dwarf grass, etc.) modeling and rendering.

As the overwhelming choice of Chinese families, home-based care has the advantages of convenience, economy and high efficiency. However, there are also some shortcomings, such as low professional level, poor emergency ability and limited time and space environment. In this regard, the paper aims at improving the effectiveness of home-based care service, and builds a home-based care system based on the Internet of Things technology, which provides a reference prototype for the service model and realization path of smart care. The whole system is B/S architecture, and the front end is an interactive page, which supports community management or medical staff to manage the aged care service. The back-end is the system server, which is developed by the framework of SpringBoot2 under Javaweb technology, aiming at completing the call and control of various data information. In addition, for all kinds of intelligent sensors and monitoring devices involved in home-based care services, real-time data collection and uploading will be completed by WIFI communication technology under MQTT protocol, and the data storage and management will be completed by relying on cloud platform technology. After simulation test, the functions of the system run smoothly and can meet the actual needs of smart home-based care services.

Data acquisition is a prerequisite for performing big data analytics. However, as the diversity and timeliness of data increase, the complexity of data collection also increases. In this paper, we take enterprise data on a big data investment platform as the research object, and design two data collection models, static data collection based on incremental crawlers and dynamic data collection based on query topic crawlers, for the static and dynamic characteristics of this data. In the experiments, this paper tests the effectiveness of these two web crawler methods and proves that they can collect static and dynamic investment data comprehensively and accurately. Thus, this study provides an effective data collection scheme that helps improve the accuracy and reliability of big data analysis.

Full-view Synchronized Measurement System (SYMS) realizes the dynamic monitoring of electronic power system. At present, the main station of SYMS can receive real-time phasor data from all over the country with SMD for loads(SML), which is independently developed. In this paper, after the master station of SYMS system is confronted with massive data access, openPDC data center cannot be compatible with SML data, openHistorian database cannot classify and recognize massive phase-quantity data, and it is difficult to read and call historical data. An adaptive method of openPDC and SML based on GSF framework is proposed, as well as a real-time phase classification algorithm for openHistorian and SML data, which realizes real-time data transmission from SML to openPDC, and real-time sorting and reading of SML data by openHistorian. A historical data storage algorithm based on.NET framework has been proposed to solve the problem of accessing and calling phasor historical data in openHistorian. The test results show that the proposed algorithm improves the data processing capability of SYMS system.

With the rapid development of unmanned aerial vehicles (UAVs) in urban logistics distribution in recent years, it demonstrates that the use of UAVs in this field will realize network operation in the future. However, identifying sites with high importance in the urban logistics UAV distribution networks is a crucial problem in developing this field. To address this, we construct a directional weighted network model of urban logistics drone distribution using basic knowledge of complex networks, with cargo flow between sites as the side weights. Classical statistical characteristic quantities of urban logistics UAV distribution networks are extracted as the evaluation indexes for key node identification in the network. The TOPSIS comprehensive evaluation method is applied to construct the evaluation system of network node importance. Using Chengdu Gaoxin South District as an example, the importance ranking of sites in this zone's urban logistics UAV distribution network is obtained, and suggestions are given on the order of site opening. The identification of key areas is of great reference value for the future development of urban logistics UAV distribution networks.

This design mainly uses Internet of Things technology to upgrade and transform the environmental monitoring system of the power transformation and distribution station. The sensing layer uses temperature and humidity sensors, water level sensors, and smoke sensors. The application layer consists of control modules such as fans, water pumps, temperature and humidity, and alarms. The actions of the executing components are controlled by the STM32 chip. The network layer packages the collected information into data packets based on the MQTT communication protocol of TCP/IP. The ESP8266 wireless Wi Fi module transmits the data packets to the Alibaba Cloud ECS server, which publishes the data packets to WeChat applets. Users can use a WeChat applet to monitor the ambient temperature and humidity values, smoke concentration, water level information, and the status of alarms, fans, and pumps in the power distribution room at any time. When the alarm conditions are met, the terminal will automatically turn on the corresponding control equipment, and can also independently and remotely control the operation of the equipment through WeChat applets, which better solves the problem of manual inspection and low efficiency in traditional power transformation and distribution stations.

Aiming at the problem of untimely fire prevention and control due to missed and false alarms in the traditional fire early warning system of substations, an intelligent fire classification and early warning algorithm based on multi-sensor information fusion is proposed in this paper. Different from the fire warning with a single sensor, firstly, the algorithm proposed in this paper combines the temperature, CO concentration and smoke sensors to build a multi-sensing fusion layer of the fire detection model, which improves the detection sensitivity to a certain extent. Then, the algorithm uses support vector machine (SVM) to classify and warn fires based on the feature information collected by the multi-sensor fusion layer. Finally, the experimental verification is carried out based on the national standard test fire dataset. The experimental results show that the proposed model can effectively and accurately classify and predict the occurrence of fire, and improve the accuracy of fire early warning decision-making to a certain extent.

This paper studies the localization problem of wireless sensor network in indoor environment. In order to meet the mobile users' demand for location service in indoor environment, an indoor distributed localization method based on coordinate projection technology and barycentric coordinate representation is proposed, which can realize large-scale localization of wireless sensor networks in 3D indoor environment only through a few anchor nodes. Meanwhile, this method has no too many restrictions on the deployment of wireless sensor networks, and can flexibly adjust sensor nodes, and has strong scalability. Finally, the effectiveness of the proposed localization method is verified by numerical simulation.

In new era, the development of cities’ travel space is emphasized by governments work report and 15-minute community living circles is increasingly concerned by people to meet the children's daily activities. Therefore, to instruct development of travel space in the child-friendly 15-minute community living circle, the paper built a model to evaluate the levels of child-friendliness in a community of Shanghai and provide some advises for future. Key indicators from questionnaires analysis and literature review are quantified and used in the machine learning model. Through evaluation and visualization in ArcGIS, the research has found that children's satisfaction level with travel is influenced by a combination of safety, comfort and fun. In the case community, the levels of safety in different streets are close and good. But, the levels of fun and comfort are unsatisfactory in whole and there is a big gap between different streets. Some measures are suggested to be implemented to meet the needs of children’s travel space, enhancing the quality of child-friendliness of the living circle.

Aiming at the deicing work of high voltage transmission lines, the existing deicing robots for high voltage transmission lines are analyzed using TRIZ theory,and the technical contradictions that exist in the deicing robots that are unable to surmount obstacles or difficult to control and dangerous in the process of surmounting obstacles are summarized. Applying 39 engineering parameters to establish the corresponding contradiction matrix to find the appropriate invention principle,using the invention principle to carry out innovative design and establish the corresponding dwarf model for analysis, a new type of intelligent deicing robot with strong obstacle surmounting ability for high-voltage transmission lines is designed, which is used to solve the low efficiency, long time and high risk of artificial deicing method; The heating and ice melting method has high energy consumption and harsh service conditions. The three-dimensional modeling software Soildworks is used to model and assemble the deicing robot, and the finite element analysis method is used to verify the feasibility of the structure of the key components, which provides a feasible idea for the further design of the intelligent deicing robot for high-voltage transmission lines.

With the rapid proliferation of Internet of Things (IoT) devices, the security risks associated with IoT devices are increasing rapidly. It is necessary to identify various types of intelligent IoT devices. This paper proposes an algorithm based on time series features to group and merge data packets and calculate statistical features to identify IoT devices. The protocol features in the response message header of IoT devices are taken into account. Furthermore, with the feature selection technique of maximum information coefficient and information gain, important feature subsets are selected by considering both the network properties and security properties of the features. Finally, based on machine learning algorithms, traffic features of smart devices are modeled to achieve identification of different IoT devices. The experimental results show that the proposed method has good adaptability and achieved higher identification accuracy in two different public datasets.

Network redundancy is a guarantee strategy for industrial networks. The paper uses single-ring redundancy technology, designs and builds a network architecture, connects network redundant equipment, and communicates with the host computer data through S7-1200 PLC, which can realize the application of rapid response backup system when the network is disconnected.

To study drivers’ driving behaviour in dilemma zones at signalized intersections, dilemma zones at signalized intersections are taken as the research objects. Based on analyzing the driving characteristics of motor vehicles. And collect the driving behavior parameters of video observations and road parameters of field observations, take these parameters as the factors which influence the drivers’ decision, conduct the significance analysis with SPSS(Statistical Product Service Solutions) for significant influencing factors, and then take these factors as the input value of BP(Back Propagation) neural network model, BP neural network model is set up and tested based on TensorFlow in Python. The prediction model of drivers’ decision-making in dilemma zone of signalized intersections under different speed limits is obtained, the research shows that the lower the speed limit, the better the accuracy of the intersection prediction. Further verified by comparison with the prediction model of drivers’ decision-making based on binary logistic, the accuracy of different prediction models is analyzed based on actual driver decisions.

This research uses video detection equipment to obtain traffic flow data, calculate the probability of secondary accidents after an accident, automatically adjust the speed limit value of expressway according to the real-time changes of road traffic flow after an accident, release the speed limit value through variable information information boards, and actively intervene in the operation of traffic flow on the expressway, so as to reduce the probability of secondary accidents and accident risk, It aims to improve traffic flow operation, reduce speed dispersion, ease traffic congestion, and improve driving safety.

In this study, the subject was the highway road network of Fuzhou City, the scale was the county-level administrative units of the city, GIS was used as the network analysis instrument, and average minimum travel time, weighted average travel time, and isochronous circle map were adopted to analyze and demonstrate the status of the traffic accessibility of Fuzhou City and the spatial layout of the accessibility. The results of the study showed that the accessibility of Fuzhou City is in circle-layer distribution, in which the city’s central area has high accessibility, with a short time to access the peripheral county areas, but the peripheral county areas have low accessibility, with poor outward connectivity. Thus, the traffic accessibility of Fuzhou City is of significant spatial differences. Therefore, Fuzhou City shall attach importance to the construction of transport infrastructure in the peripheral county areas, and rationally arrange and improve its traffic network, hence increasing its overall traffic accessibility. The city also shall enhance the connections among its different counties (or county-level city) and districts, and utilize the radiation of the economic and traffic advantages of the central urban area to drive the development of the peripheral county areas.

In order to clarify the evolution path of sudden traffic accidents under the special state of "construction while opening to traffic" in expressway reconstruction and expansion projects, and improve the scientificity of emergency command and decision-making and the pertinence of disposal measures, this paper first analyzes 12 accident scenarios of 4 typical risk events under the interactive influence of reconstruction and expansion construction operations and traffic operations. After that, the key scenario elements are selected based on the risk matrix, and the scenario representation method is determined by using knowledge element theory for reference. The scenario model of reconstruction and expansion of sudden traffic accidents based on Bayesian network is constructed to realize scenario deduction. Finally, taking the traffic accident of a province reconstruction and expansion project as an example, the accident development path is simulated. The results show that after the reconstruction and expansion of the construction lane is narrowed, the construction traffic organization is unreasonable, and the influence of rainy days is superimposed, the probability of vehicle scraping anti-collision facilities is 86.2%, the probability of local traffic congestion is 72.8%, and the probability of major traffic accidents with casualties and vehicle damage is 63.1%. Its evolution path is basically consistent with the actual development of accidents, and it is clear that on-site control, traffic diversion and remote diversion are the key measures to control the deterioration of accidents.

With the continuous development of civil aviation, the airline's flight network and fleet size are increasing. In order to ensure that all flights of airlines can successfully perform flight missions and improve the overall operation efficiency of airlines, airlines need to invest more human and material resources to carry out the scheduling of flight schedules. At the same time, in some special cases, when some airports or flights are delayed due to irresistible factors, how to formulate a reasonable flight recovery schedule quickly and effectively is a key that affects the revenue and operation efficiency of airlines. Aiming at the flight delay scenarios under different circumstances, this paper proposes a flight schedule recovery method based on the existing operations of airlines, aiming at maximizing the marginal benefit, that is, the comparison between the marginal revenue of a flight and the marginal cost paid by the airline. This paper intends to improve the efficiency of flight recovery in the case of airline flight delays, and provide some reference for relevant enterprises.

In order to solve the problem of “parking difficulty”in cities, the distribution of parking lots in the central urban area should be fully analyzed. Taking Tianjin as an example, the city POI data is obtained through the Gaode open platform, and the powerful spatial analysis and data processing functions of GIS are used to analyze the spatial distribution characteristics of urban parking facilities from the macro levels, meso levels, and micro levels, respectively, for the spatial distribution characteristics of parking facilities, the correlation analysis between parking facilities and other land types, and the service level of parking facilities. The current supply of parking facilities can provide technical support for special planning of parking, and this method can also be used as a reference for other cities.

Confronted with fast-paced technological progress and surging travel needs, some large-scale railway passenger stations now have fallen short of passengers' demands for comfortable trips due to their complicated passenger inbound process, circuitous inbound routes, and serious passenger traffic congestion. Taking Zhengzhou East Railway Station, a large-scale passenger transport hub in China, as the case study, this paper first elaborates on the construction of the simulation model based on Anylogic for reproducing inbound passenger flow lines thereof. Next, problems existing in the current passenger flow organization are analyzed using the average inbound time of passengers from each entrance and the average queue length for each kind of passenger service equipment as evaluation indexes. And then the optimized joint measures for different operational issues including four universal flow line adjustment methods are proposed, namely, reducing flow line density, increasing flow velocity, eliminating flow line bottleneck and relieving flow line crossing. Flow line simulation experiments herein have verified the rationality and effectiveness of the foregoing optimization measures for flow line organization of large-scale railway passenger stations.

With the urban rail network tends to be more complex, the contradiction between capacity and passenger demand becomes more prominent. To address this problem, this paper constructs a comprehensive evaluation method for train capacity delivery scheme, which can quantitatively analyze the matching relationship between capacity and passenger demand. It can provide guidance for the development of operation plans. The evaluation indices are selected from three aspects: operation, service, and enterprise cost, among which the operation indices are selected considering the characteristics of section, route, and line network respectively. Combined with the ordering relation analysis method (G1) and the anti-entropy weight method (anti-EWM), a comprehensive evaluation model from the perspective of subjective and objective is developed, the model can reasonably allocate capacity under limited conditions and alleviate the contradictory relationship between passenger demand and capacity. In addition, the model can improve the operation level and the service quality. Furtherly, the comprehensive scoring of Nanjing Metro Network is conducted with better evaluation effect, which can provide reference for the selection of capacity delivery.

The heat fade of fully loaded trucks while braking will readily occur on long downgrade route sections of mountainous highways due to speed-control and continuous braking. If the brake fails, it will lead to serious traffic accidents.In this article, the brake temperature variation process of fully loaded medium trucks was simulated through the bench tests. First, two kinds of single forecasting models were established based on experimental data and regression analysis, and the combination forecasting model of the braking risk of fully loaded medium trucks on the long downgrade route sections was established based on the entropy weight method. At the same time, the regularities of drum brake temperature variations under different initial speeds were analyzed, and the critical temperature value and the maximum critical braking times of the brake failure corresponding to different initial braking speed were obtained. The results show that the higher the initial braking speed is, the higher the failure temperature of the brake; however, the lower the maximum critical braking times of the brake failure are, which shows that the higher the initial braking speed is, the easier it is to reach the failure state. The risk combination forecasting model is more reliable and scientific.

With the continuous improvement of the city brain, lacking application and exploration in depth of implement scenarios for collection, aggregation and dynamic operation of total factor traffic data in the field of traffic improvement planning has been realized. Taking the key urban scenes as the research object, the paper deeply explores the traffic operation mechanism, establishes an integrated systematic treatment process of "operation monitoring - problem diagnosis - supply and demand analysis - scheme design - effect evaluation", provides all-round perception, analysis and judgment support for urban managers. Taking the Peking University People’s Hospital as an example. A system is established in the paper to perceive total factors of traffic behavior participants, to optimize scheme design by combining traditional improvement measures and information means and to evaluate the improvement scheme through integrated microscopic traffic simulation model The application demonstration rudimental system of "panoramic perception - monographic study - effect evaluation" is established to realize the closed loop of decision-making process systematization, with a view to providing reference of the transformation of traffic improvement planning and design technology system under the information background.

Double carbon target has become the national development strategy. Transportation is a key area of energy consumption and greenhouse gas emissions in China, and the task of emission reduction is arduous. China is still in the stage of rapid development of motorization, car ownership continues to increase, exacerbating the urban traffic congestion. Traffic congestion will lead to an increase in energy consumption, thereby increasing vehicle emissions. Therefore, in order to quantify the impact of carbon emissions on urban traffic emission reduction under traffic congestion, the process dynamic detection data of CO2 and CO, HC, and NO from the annual inspection data of gasoline vehicles in Jinan (simple transient working condition method) are used as the basis, and the carbon emissions under traffic congestion are modeled by using the multinomial regression analysis method in combination with different working conditions of traffic congestion. The results show that the number of stops (acceleration and deceleration times) and idling time of vehicles in traffic congestion have a large impact on the emission of gaseous pollutants. Under the idling condition, the emission of various gaseous pollutants presents a constant value; under the deceleration condition, the emission rate of gaseous pollutants is negatively correlated with the vehicle speed; under the acceleration condition, the emission rate of gaseous pollutants is positively correlated with the vehicle speed. when the vehicle is in moderate congestion, its carbon emission is 4.02 times higher than that in smooth traffic, and when the vehicle is in severe congestion, its carbon emission is 7.2 times higher than that in smooth traffic. Reducing traffic congestion can effectively reduce vehicle carbon emissions.

In recent years, with the large-scale deployment of 5G network, some business relying on 5G technology or 5G network equipment have gained new vitality. Intelligent bus is one of the typical application scenarios. In this paper, we analyze the current demand for intelligent bus, sort out and dismantle the key business scenarios of intelligent bus that can be reformed based on 5G network. Through the top-level architecture design, 5G WAN-priority customized network service mode is introduced in smart bus business, also we sort out the intelligent bus business and clarify the technical requirements, making quantitative analysis of key business scenarios for video surveillance. Besides, solutions are built for the smart traffic monitoring/operation system, smart buses, smart bus stations and other business scenarios. Through this paper, we aim to conduct a cutting-edge discussion on the ecosystem of intelligent bus, so as to promote the operational implementation of 5G mobile communication network on intelligent bus business under the background of high-speed development of smart city.

Building footprint information is crucial for urban-related applications. Due to the rapid growth of the number and types of community-level buildings, the completeness and coverage of building footprint data at the community level are lagging. Efficiently and accurately generating community-level building footprint maps for the update of community electronic maps, post-disaster assessment, navigation services, etc. is an urgent and challenging task. To fill the data gaps of community-level building footprint with high accuracy, we design a building map translator named BuildingMap- GAN that is configured with a novel generator: MACU-Net. It has multi-scale connections with channel attention and asymmetric convolution blocks (ACB), which are used for fully capturing the spatial structure information of buildings and roads. An experiment is conducted on real-world community dataset in Wuhan, China, in which the street network data of 700 communities were used for training and the street network data of 42 communities were used for verification. Compared with the two baseline models, F1 scores are improved by 6.7% and 4.5%, respectively, and IoU values increase by 5.0% and 3.2%, respectively. Our model can generate refined building footprint data with higher precision and is closer to the real building distribution.

As part of the Intelligent Transportation System (ITS), traffic flow modeling and optimization have become widely employed because of their significant economic efficiency, safety, repeatability, usability, and controllability properties. This study justifies the selection of Shenzhen and Qingdao as simulation sites and delves into the simulation approach, encompassing static road network modeling, vehicle movement modeling, output analysis, and control system research. The primary objective is to enhance traffic control strategies and boost public transportation efficiency. To achieve this, the research examines, refines, and tests an array of traffic control strategy models, including cycle optimization, green time optimization, and early green initiation. Building on these data adjustments, the control flow for single-phase active public transit priority and multi-phase operational public transit priority is optimized and assessed. Simulation and optimization of the two control strategies reveal that public transportation prioritization offers a more pronounced advantage on roads during peak hours, making it better suited for cities grappling with congestion during rush hours.

Given the imperfect and non-comprehensive situation of the domestic evaluation system on the transportation service quality level, this article will take urban public transportation as the research object, from the perceived services level of passengers, aim to construct an evaluation system regarding urban public transportation service level. To explain the applicability of the result, the survey data obtained from the investigation are thoroughly analyzed in terms of their reliability and validity. A further evaluation is also conducted on the level of public transportation service in urban using the analytic hierarchy process.

With the advent of the era of big data, the application of big data technology is deepening in the construction of smart cities. Due to its many advantages, big data technology plays an important and irreplaceable role in the construction of smart cities, such as revitalizing industries, assisting the government and facilitating people's livelihood, and can exert a positive impact on the development of many industries. Therefore, many industries are researching and developing the application of big data technology. In the hope of improving the efficiency of development. The effective application of big data technology in smart city planning can speed up the construction of smart city and make it more perfect. Combined with the present situation of smart city construction in our country, this paper will put forward some suggestions on the application of big data technology in smart city construction.

Accurate prediction of passenger flow can provide an important basis for the operation and management of urban rail transit. Previous models are based on the learning of static graph structures and cannot achieve the distinction of network structures by attention mechanism. In order to realize the learning of dynamic spatio-temporal characteristics of passenger flow and improve the accuracy of passenger flow prediction, a neural network model based on attention mechanism is proposed in this paper. It consists of a spatial attention module and a temporal attention module. The model uses three different coding strategies to enhance the learning capability of the attention mechanism for spatial location and structural features. In the temporal attention module, bi-directional GRU and attention are combined to extract dynamic changes in the temporal dimension of passenger flow data. Experiments on the Hangzhou Metro dataset demonstrate that this model outperforms the classical model.

The road freight industry is an important part of China's comprehensive transportation system. Promoting the high-quality development of road freight requires improving the industry's dynamic monitoring and analysis capabilities. The origin destination (OD) distribution of road freight demand is one of the key monitoring and analysis contents in the road freight industry and there are many problems in the traditional data acquisition methods. Because it is difficult to obtain the freight OD data, this paper relies on the vehicle geographic information data in the existing road freight vehicle dynamic monitoring platform, uses the data processing technology based on the big data of freight vehicle tracks, focuses on the key technologies such as long-term parking node screening and achieves efficient and accurate identification of freight vehicle parking points. This paper builds a foundation for subsequent OD analysis, improves the dynamic analysis and monitoring system of China's road freight industry. It strongly supports the transformation of industry management from "traditional experience judgment" to "data decision-making", promotes the improvement of industry management level, and promotes the modernization of industry governance system.

In order to improve the efficiency of handling operations at a certain bay on a ship in a container terminal, the operation sequence optimization of quay crane and yard crane is studied under the condition of considering the storage in the yard. Using linear programming method of operational research, a mixed integer programming model is established to minimize the operation time of quay crane and give the operation sequence of the quay crane. A mathematical example is designed to compare with the dual-cycle operation strategy that usually does not consider the container relocation operation in the yard. In addition, the sensitivity of the model is analysed to verify the results of the model under different scenarios. The experimental results show that the dual-cycle operation strategy considering relocation in yard some-times has less dual cycle times but shorter total working time than the dual-cycle operation strategy without considering relocation in yard, which verifies its necessity. And while optimizing the operation sequence, it can give suggestions on the operation of the yard, indicating that the constructed model can better handle the optimization problem of the quay crane operation considering container relocation in yard.

At present, the urban logistics distribution of electric vehicles in the sharing economy has a good development opportunity. Under the premise of vehicle sharing, the bi-level programming method is used to study the location-routing problem of different vehicles in multi-depot considering dynamic energy consumption and profit distribution. This paper considers the influence of the weight of the cargo, the driving distance and the dynamic road network information on the energy consumption of electric vehicles. In order to balance the interests of companies and vehicle owners, a location routing model for electric vehicle distribution is constructed with the objective of minimizing the cost of dispatching vehicles by the company and the minimum energy consumption of the vehicle owners during the whole journey including the return trip, constrained by the hard time window, the maximum vehicle load and the maximum return no-load distance. Design a heuristic algorithm that combines two-layer variable neighborhood search and genetic algorithm to solve the model. The results of case analysis show that the model can obtain satisfactory solutions for companies and vehicle owners. The distribution activities of shared vehicles among multiple distribution centers reduce the total cost of the company, and the vehicle owner completes the distribution task with minimal energy consumption. This paper provides an effective method for solving the location-routing problem of electric logistics distribution network and profit distribution between transportation companies and vehicle owners under the sharing economy.

To create more vitality and sustainable cities, enhance quality of life for residents is the key issue of urban renewal. Multi-source data is main data source with ability of integration and analysis of data from multiple sources including POI, mobile phone signaling data, GPS track data, to better understand and improve the overall health and wellbeing of urban areas. In this paper, we proposed a framework of urban vitality estimation using mobile phone signaling data and POI data by combing Moran’s I to determine the suitable analysis grid. As a result, the area of triangle made up of Old city, Baishahe, and Yangjiapu of Weifang city has a high degree of vitality, as it has both a diverse mix of land uses, and some concentration of land uses in certain areas.

Nowadays, urban rail transit network (URTN) has become an indispensable and important transportation infrastructure in rapidly developing cities, then how to effectively guarantee the normal operation of urban rail transit network and effectively prevent the impact of risky incidents on the performance of URTN to be increasingly prominent. This paper identifies critical URTN stations and links through a quantitative study of urban metro network vulnerability modelling analysis, taking into account urban bus connections. The paper also uses big data acquisition and analysis processing techniques to obtain more realistic and accurate traffic analysis data for the passenger flow characteristics of the urban metro network. Afterwards, the vulnerability of the URTN network is analysed through the accessibility measurement of the change in travel cost when different stations (links) are deliberately attacked under alternative transportation modes, so as to identify the critical stations and links in the URTN network. This paper presents an analysis based on the Shenzhen metro network as a case study. The results show that the consequences of disrupting different stations on URTN accessibility clearly differ when stations with different passenger flow characteristics are processed in the context of big data technology, and that some stations that are not disrupted are found to become more vulnerable under the failure of surrounding stations. The proposed approach provides a reliable metric and methodology for the identification of critical stations and links in urban rail transport networks, and their impact on decision-making under disruption.

The study of low-carbon multimodal transport path optimization problems considered in a fuzzy demand environment has important theoretical and practical significance in the situation of high-quality development. By analysing the demand uncertainty problem in the transport process, an improved simulated annealing genetic algorithm is designed to solve the model. The impact of various carbon policies on multimodal transport solutions, costs and carbon emissions is analysed through arithmetic examples. The results show that: 1) the improved simulated annealing genetic algorithm is better than the traditional genetic algorithm in terms of time finding and effect finding to achieve the lowest cost and lowest carbon emission; 2) the carbon tax policy is studied through the example and it is found that the carbon tax constraint is relatively lenient and the improper setting of carbon tax will lead to the increase of total cost; the model and algorithm proposed in this paper can provide theoretical support to the policy making departments and multimodal transport enterprises to optimize transport solutions. The model and algorithm proposed in this paper can provide a theoretical basis for policy making authorities and multimodal transport enterprises to optimize transport solutions.

The current automated container terminal frequently completely isolates the external and internal container trucks in space for the safety of the vehicles. Physical isolation needs more space for roads and will reduce yard space. Furthermore, when a traditional container terminal is upgraded to an automated container terminal with physical isolation, the plan layout needs to be changed dramatically with substantial investment. With the development of technology, unmanned container trucks are gradually enabled in automated container terminals. It is necessary to ensure traffic safety in the absence of physical isolation in the port. To address path conflicts that arise when internal unmanned and external container trucks are combined in traffic, this article focuses on the impact of internal unmanned and external container truck avoidance strategies on traffic scheduling in automated container terminals. A mathematical model is built in this paper and solved by a genetic algorithm to study the influence of two avoidance methods: Internal Unmanned Container Truck First and External Container Truck First. The study's findings indicate a correlation between avoidance strategy benefits and drawbacks and the proportion of internal unmanned trucks to external container trucks.

This paper studies the risk assessment of traffic flow on major highways under high-speed driving conditions and the safety warning after traffic incidents. Firstly, by analyzing the operation characteristics of the main line traffic flow and the evolution characteristics of the upstream traffic flow after the event, a collaborative control strategy for the dynamic assessment and accurate early warning of the main line traffic flow is proposed. Secondly, based on the operating speed, traffic flow and head-time distance state parameters of the main traffic flow, the risk assessment model of the main traffic flow is established, which provides an algorithm basis for the safe operation status evaluation and active control of the main traffic flow. Third, based on the driver’s reaction process after the traffic incident and the evolution process of main traffic flow, a traffic incident safety warning distance function is proposed, which provides a quantitative function for the accurate early warning and prevention of traffic accidents in the upstream traffic flow after traffic incidents. Through the dynamic assessment and control of operation risks of expressway main line traffic flow, the accurate risk warning of expressway main line traffic flow under the traffic event conditions will significantly improve the dynamic safety control level and risk warning ability of the expressway main line.

In order to increase the accuracy and real-time performance of airport vehicles path tracking, a new route re-planning strategy to avoid obstacles based on the model predictive control (MPC) algorithm is suggested for vehicle online motion planning. The use of on-board sensors to obtain information on obstacles and the reference path, path the desired paths of obstacle avoidance through path re-planning controller, and the re-planning of the desired path information into the MPC controller for controlling the output of the front wheel angle, steering wheel through the active steering to simultaneously achieve obstacle avoidance, in order to track the reference path. Ultimately, Carsim and Simulink construct a closed-loop control system simulation platform for route re-planning and trajectory tracking, using a double lane change scenario as an example. The findings indicate that the suggested technique may accomplish reliable obstacle avoidance at varying vehicle speeds and steady reference route tracking.

Traffic flow prediction is a crucial component of establishing the intelligent transportation system. Accurate and real-time traffic flow prediction is of great significance for urban traffic management. With the recent development of artificial intelligence, deep learning-based methods have been effective tools for traffic flow prediction. However, locality unawareness and data scarcity are still two open issues to be considered in traffic flow prediction tasks. To address these problems, we propose a novel locality-aware spatial-temporal attention neural network, named LASTANN, in this paper. Specifically, we propose two elaborate attention modules to perceive local information in spatial and temporal dimensions. We also propose an auxiliary module based on contrast learning to strengthen the representation ability of model. We verify the effectiveness of the proposed model on two real-world traffic flow datasets. The experimental results demonstrate that the proposed LASTANN outperforms other baselines consistently and each component enhances the prediction performances significantly.

Congestion control has been widely investigated and become an essential issue in transport network especially for peak period. However, existing methods are concentrated on congestion detection and utilize the transmission dynamics to analyze the characteristics of network congestion, which ignores the development of avoiding congestion occurrences and providing the possible solution for congestion situation. Therefore, we utilize the convolutional neural network to arrange the vehicles with limited routes to avoid the transport congestion. Initially, we collect the transport network and number of existing vehicles information about Chaozhou, which is a city in China. Subsequently, a convolutional neural network is established to dispose the congestion issue when the transport network in peak period. The neural network contains the convolution layer, data pro-processing layer, softmax to optimize the output and a full-connected layer as obtaining the final results. At last, two compared congestion control mathematical models are also simulated in same situation to evaluate our trained neural network. From our extensive experimental results, we can conclude that our proposed method can effectively avoid the occurrences of transport congestion and contain reasonable computation costs compared with existing congestion control algorithms.

In recent years, with the rapid development of road traffic in China, the problem of road traffic safety has become increasingly prominent, especially the illegal behavior of vehicles not using high beam lights according to regulations when driving at night, which has gradually become the main cause of traffic accidents at night. According to statistics, about 43% of nighttime traffic accidents are related to the abuse of high beam lights. The main reasons are that the driver is not clear about the provisions of the high beam lamp, not familiar with the operation, and deliberately turns on the light for a long time. The problem has affected the normal and stable social order. Therefore, it is of great significance to effectively prevent the abuse of high beam lights by vehicles at night for promoting traffic safety and social order at night. This paper takes the abuse of high beam lights by vehicles at night as the research object, analyzes the problems of night traffic in detail in combination with various conditions of the night road, and studies and proposes a new intelligent warning system optimization scheme to prevent the abuse of high beam lights by road vehicles at night. Through research and creation of high beam data collection algorithm processing, the intelligent traffic system for driving at night is established to reduce the incidence of traffic accidents and ensure traffic safety.

In order to cope with the large-scale demand for medical supplies from medical institutions in the context of public health emergencies, improve the efficiency of emergency treatment and reduce casualties and economic losses. We analyze the factors of medical demand urgency, propose the concept of demand urgency, and thus introduce the reward and punishment mechanism. The emergency distribution model is constructed with the sum of vehicle dispatch cost, transportation cost and reward and punishment cost as the objective function considering the demand urgency. The model and algorithm are validated by using ant colony algorithm to solve the case on a large scale.

This paper considers how best strategies should be chosen to effectively mitigate light pollution situations in cities with different light pollution scenarios. Firstly, a mathematical programming model is used to analyze the light pollution characteristics of four typical areas: protected areas, rural communities, suburban communities and urban communities, and four states are abstracted. Subsequently, using reinforcement learning models, the four abstracted states are used as the state space of the intelligences, while promoting green building and eco-city design, rationalizing the layout and height of road lighting, and improving the performance of lighting equipment and lighting solutions as the three governance strategies, constitute the action space. Through continuous training it was concluded that areas with strong road and residential lighting, frequent night-time camping and other activities adopt the strategy of rationalizing the layout and height of road lighting; areas with dense night-time light sources, high light intensity and long duration adopt the strategy of promoting green architecture and eco-urban design. And areas with a high demand for night lighting, high air pollution index or large open areas adopt the conclusion of increasing lighting equipment and lighting solutions.

Coastal cities are rich in natural resources and geographical advantages. The ecological quality of coastal areas is one of the important factors affecting the sustainable development of coastal cities. Dynamic monitoring and quantitative evaluation of ecological environmental quality can provide a scientific basis for ecological environmental protection and sustainable development. This paper took Ningbo Hangzhou Bay New Area as the research object, based on four issues of remote sensing image data in 2013, 2015, 2017, and 2020 and constructed the remote sensing based ecological index model of Ningbo Hangzhou Bay New Area by extracting index information such as greenness, humidity, dryness, and heat. It is concluded that during the study period, the improvement area of eco-environmental quality is larger than the degradation area, indicating that the eco-environmental quality of Hangzhou Bay New area is developing in a good direction from 2013 to 2020.

This paper focuses on light pollution and explores the management of urban light pollution. Firstly, after investigation, we know that atmospheric influence factor, night light optical index and regional planning factor are important indicators to measure the level of urban light pollution. Thus, we established a TOPSIS model based on entropy weight method, which can arrive at the conclusion that regional planning factors have the greatest influence on light pollution. Based on this, we then proposed a management strategy for the serious light pollution situation in urban commercial areas. Finally, we use Cellular Automata to simulate the process of the effect of this treatment strategy on urban light pollution.

In view of the problems of long construction period and low precision in traditional three-dimensional modeling, it has been unable to meet the development needs of smart city construction. Based on UAV tilt photogrammetry technology, this paper carries out three-dimensional real scene modeling of Zhongdu ancient town. The results show that the three-dimensional model of Zhongdu ancient town constructed by UAV oblique photogrammetry technology is realistic, rich in texture and has high accuracy, which greatly reduces the cost of three-dimensional modeling and effectively improves the efficiency of model production and provides important technical support for the construction of ' smart city '.

Urban rail transit has received widespread attention due to its advantages such as fast speed, high punctuality, green and low-carbon. To explore the characteristics of passengers in urban rail transit, a station classification method based on data mining is proposed. Using urban rail transit inbound and outbound swiping data, the station is divided into 5 categories using K-means clustering algorithm, with the inbound and outbound passenger flows of each station as variables during the three periods of the day, morning peak, and evening peak. The results show that the inbound and outbound passenger flow data can better reflect the spatiotemporal characteristics of different types of rail stations. Finally, the passenger travel characteristics of different types of stations are analysed, and the identification study of different types of stations can provide reference for the planning, design, and operation management of rail transit stations.

In order to build an information security evaluation index system for smart cities, this paper collects 19 risk factors that affect the information security of smart cities from the four dimensions of environment, data, users and management based on literature analysis. Then, the study uses Delphi and DEMATEL to discriminate and improve the above risk factors, to build an information security evaluation index system for the current situation of smart cities in China. This study can provide certain reference suggestions for the development of smart city construction.

In order to meet the requirements of intelligent and information development of urban traffic information control operation and maintenance, an intelligent operation and maintenance system of the urban traffic signal is put forward. First, the overall architecture of the system is constructed. According to the management requirements of urban traffic signals, the architecture based on the interface layer, platform layer, and application layer is proposed. Secondly, a calculation method of the health degree of traffic signals based on multi-level comprehensive evaluation is proposed. For the signal equipment from the operation, maintenance to replacement of the scientific closed-loop management, to complete the optimization of its running status monitoring and fault management. It is helpful for the traffic control center to find and solve the fault of the signal the first time, improve the efficiency of operation and maintenance, and ensure traffic safety.

The existing traditional neural network reconstruction models have some questions, including the high training epochs, low recognition rate, and complex structure. The restricted Boltzmann machine (RBM) is an excellent generative learning model for feature extraction, which is a simple model compared to other deep neural networks. The Discriminative Fuzzy Restricted Boltzmann Machine (DFRBM) is proposed by extending its parameters from natural numbers to fuzzy ones. Then we introduced the random permutation (RP) algorithm, with the hidden units random permutation, Discriminative Infinite Fuzzy Restricted Boltzmann Machine (Dis-iFRBM) is proposed. Dis-iFRBM is a better RBM model than DFRBM and Classical RBMs.We further investigate and compare the generative ability of the Dis-iFRBM on image reconstruction. First, we transform the MSTAR SAR piece to HRRPs images. Then the Dis-iFRBM, DFRBM, and Classical RBMs are compared in detail under optimal conditions on the HRRPs images that transformed from MSTAR data sets. Specifically, they can be trained by relatively limited datasets into excellent stand alone classifiers and retain satisfactory generative capability simultaneously. The comparison of experimental images shows that the Dis-iFRBM possesses better generative capability than the Discriminative Fuzzy Restricted Boltzmann Machine (DFRBM). Meanwhile, surveillance images in the city, license plate number recognition, and other scenarios need damaged image restoration. Dis-iFRBM as a model can save computational resources of terminal image devices that deploy in the Urban Internet of Things. The experiment results indicate that the Dis-iFRBM outperforms image restoration. It can achieve smaller average reconstruction errors (AREs) while given a small number of hidden units. Finally, experimentation over several classical RBMs revealed the proposed approach’s preferable reconstruction capability.

Predicting the motion and behavior of surrounding vehicles is an essential task for motion planning and decision-making of autonomous vehicles in complex traffic conditions. In this paper, we propose a short-term vehicle trajectory prediction framework using attention mechanism integrated GRU network. We use an encoder-decoder model as the main architecture. A gate recurrent unit (GRU) coupled with temporal attention and graph attention is used to extract and fuse more important information which could be used for trajectory prediction. The temporal attention could extract temporal information and graph attention could consider interactions between surrounding vehicles within sensing range. The extracted information is fed into fully connected layers to obtain predicted trajectory. The publicly next generation simulation (NGSIM) I-80 and US-101 datasets are used to evaluate proposed model. Compared to other prediction models, our model shows improvement on final displacement error (FDE) and average displacement error (ADE). The results show that model with attention mechanism improves prediction accuracy by 1% ~5% in 5 second prediction horizon.