This PDF file contains the front matter associated with SPIE Proceedings Volume 12549, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Visual homing is a lightweight approach to visual navigation to direct a single robot to a visual landmark without the use of a map or any form of GNSS. However, applications that require a team of robots to operate robustly with respect to map and GNSS requirements, may also require groups of robots to move together and/or in a coordinated fashion. Examples include agricultural robotics where several robots must coordinate to deliver material to a visually identified rallying point, or a reconnaissance and surveillance mission where robots must move towards a (potentially moving) visual target in a formation. We present and evaluate several visual homing algorithms extended to handle team homing. We present evaluation results for stationary rallying points and moving target examples using a ROS/Gazebo simulated urban environment and real Turtlebot3 robots. We show that our algorithms produce accuracies within a 95% confidence interval in both simulation and physical experiments.

This paper proposes an approach to design a quadruped robot with a dynamic spine using Reservoir Computing. Four bending sensors are integrated into the continuous dynamics spine to adjust spine force for balancing and enable closed-loop control of the Center of Mass (CoMs) mapping onto the dynamic surface. Passively modeling the system enabled a quadrupedal robot with a flexible spine to navigate rough terrain with improved efficiency, agility, and minimized impact on explosive power. Additionally, paper introduce to develop a novel, intelligent control mechanism for the quadrupedal robot to operate effectively on dynamic, flexible surfaces. Our approach focuses on a hierarchical distributed control mechanism developed for leg to cooperatively change centralized frames for a better functional reflection.

We present a novel approach to underwater fleet communication using single-photon avalanche diode (SPAD) arrays on the receiver. Using a large array and a focusing lens, the array can be spatially allocated to different sectors within the lens’s field of view. This eliminates the need for precise pointing of the receiver towards the transmitter and enables simultaneous reception of multiple transmitters across the entire array. We demonstrate the feasibility of our approach through a prototype receiver that utilizes a 7x7 SPAD array, a wide field-of-view lens and FPGA processing. We were able to achieve a raw data rate of 6 Mbps across a folded 25m clear water channel. Through the use of artificial attenuation with ND filters, we estimate that the achievable distance for a raw data rate of 6 Mbps in clear water with a BER of 10^-3 is approximately 400m. In over-the-air testing, we were able to achieve simultaneous communication over the entire 7x7 array at 6 Mbps.

In this paper, Chameleon Swarm Algorithm, which is a new metaheuristic algorithm, is used to design a speed reducer gearbox. The gearbox is used in an autonomous vehicle, and it is supposed to take in consideration the total weight of the gear sets along with other contains that include the stresses; including the bending and surface stresses for the gears and the stresses of the shaft, and he deflections of the shafts, beside other contains. The algorithm must find the optimal solution while satisfying the eleven constraints, which makes it a multi-objective optimization problem. The algorithm minimizes the energy dissipation in the gears as the design is optimal. The algorithm will be tested in Monte Carlo simulation to show its stability and robustness in such an application. The computer resources are examined, and the results are compared to the particle swarm optimization. The proposed method shows better results and better performance compared to the PSO.

PLX's Orthogonal Laser Metrology Module (O-LAMM) integrates core Monolithic Optical Structure Technology™ (M.O.S.T) [1] and precision Beam Steering Technology [2], creating a compact and robust 3D laser-scanning system for in situ metrology. The O-LAMM generates a laser line directed into a scanning mirror system, scanning multiple targets at high speeds of 100 Hz using retro-reflective cooperative markers on robotic end-effectors. The system captures X and Y locations upon laser-retroreflector interaction, and through triangulation with a second O-LAMM unit at a known distance, it enables Z location determination. By networking multiple O-LAMMs, blind spots within a target volume are minimized, providing a full specification 6-DOF high-precision position measuring device. This innovative solution is an essential component for adaptive manufacturing and Industry 4.0 applications, maintaining arc second accuracy through the integration of multiple systems while offering speed and flexibility across a wide range of measurement-related tasks in production lines. The O-LAMM system is well-suited for testing, inspection, positioning, deformation analysis, and tracking applications, delivering unmatched high -speed geometric data collection and Intelligent Manufacturing System (IMS) applications.

Robotic manipulators involve the motion in 3D space. The motion of the manipulator is usually done through rotational movements using rotational joints, and/or through translational movements using prismatic joints. In this work, a prismatic-prismatic robotic arm is modeled in a matrix form and then it is simulated through MATLAB©. The simulation solves the dynamic of the system numerically. The motion of the manipulator’s end-effector of the double prismatic joints is observed.

The use of passive sensors to minimize signature in night-time autonomous driving of robotic ground vehicles has been a goal for roughly three decades. Demonstrations have been done in the past at low to moderate speeds with vehicles using stereo pairs of thermal cameras for 3-D perception; however, there has never been an end-to-end model of the probability of mission success in this application, defined here as the probability of colliding with an obstacle and the expected rate of false obstacle detections as a function of distance traveled and other relevant parameters. We integrate and extend prior work on modeling the performance of 3-D perception for obstacle detection with thermal stereo vision to provide the first such model, We include experimental results with a stereo vision algorithm based on a deep neural network (“deep stereo”) on LWIR stereo images and on synthetically generated LWIR and visible stereo images to characterize key elements of sensor performance.

As one of the classic fields within the area of computer vision, image classification and segmentation solutions as topics have expanded exponentially in terms of accuracy and ease of use. On Mars, the atmospheric and surface conditions can lead to the sudden onset of a dust storm, or a more common dust devil, causing a multitude of issues for both equipment and crew. The ability to identify and locate area which should be avoided due to these storms is necessary for mission safety. Many current techniques are not practical due to being hefty and computationally expensive for specific tasks that require the ability for swift deployability onto systems with more stringent constraints. This paper proposes a novel approach to the problem of segmentation by marrying an efficient yet powerful Vision Transformer based model with traditional signal processing techniques to ensure peak performance. With the National Aeronautics and Space Administration (NASA) looking to land a team on Mars, this paper takes on the real time hurdle of classifying and segmenting dust storms within remote satellite equatorial photos, using a model designed to be integrated on any and all future systems, increasing overall mission success.

The Air Force Civil Engineer Center’s C-17 Load Cart is a large, 150-ton machine based on a modified Caterpillar 621G scraper for testing experimental pavements used in airfield surface construction and repair. Long lasting, durable, preparein- place, minimally resourced pavements represent a critical technology for airfield damage repair, especially in expeditionary settings, and formulations must be tested using realistic loads but without the expense and logistical challenges of using real aircraft. The Load Cart is an articulated vehicle consisting of the 621G tractor and a custom trailer carrying a weighted set of landing gear to simulate the loads exerted during aircraft landing and taxiing. During the test a human driver repetitively traffics the vehicle hundreds of times over an experimental patch of pavement, following an intricate trafficking pattern, to evaluate wear and mechanical properties of the pavement formulation. The job of driving the Load Cart is dull, repetitive, and prone to errors and systematic variation depending on the individual driver. This paper describes the full-stack development of an autonomy kit for the Load Cart, to enable repeatable testing without a driver. Open-source code (Robot Operating System), commercial-off-the-shelf sensors, and a modular design based on open standards are exploited to achieve autonomous operation without the use of GNSS (which is challenged by operation inside a metal test building). The Vehicle Control Unit is a custom interface in PC-104 form factor allowing actuation of the Load Cart via CAN J1939. Operational modes include manual, tele-operation, and autonomous.

Path planning and obstacle avoidance are crucial tasks in the robotics and autonomous industry. Path planning seeks to determine the most efficient path between a start and an end point, whereas obstacle avoidance seeks to avoid collisions with static or dynamic obstacles in the environment. On this work, we utilize the Chameleon Swarm Algorithm (CSA), which is a metaheuristic approach, for path planning and obstacle avoidance on a predetermined map with static obstacles. This CSA extracted the optimal path from several possible different paths, and the results showed that it has slightly superior performance compared to PSO.

Autonomous vehicles (AVs) employ a wide range of sensing modalities including LiDAR, radar, RGB cameras, and more recently infrared (IR) sensors. IR sensors are becoming an increasingly common component of AVs’ sensor packages to provide redundancy and enhanced capabilities in conditions that are adverse for other types of sensors. For example, while RGB cameras are sensitive to lighting conditions and LiDAR performance is degraded in inclement weather such as rain, IR sensors are unaffected by lighting conditions and can contribute additional meaningful information in inclement weather. The US Army Corps of Engineers, Engineer Research and Development Center (ERDC) has developed the ERDC Computational Test Bed (CTB) to provide a suite of tools that can be used to support virtual development and testing of AVs. The CTB includes physics-based vehicle-terrain interaction, sensor and environment modeling, geo-environmental thermal modeling, software-inthe- loop capabilities, and virtual environment generation. Thermal modeling capabilities within the CTB utilize decades of near-surface phenomenology and autonomy research. Recent additions have been made to support large-domains commonly required for autonomous vehicle operations. These additions provide high-fidelity, physics-based thermal transfer and IR sensor models for creating high-quality synthetic imagery simulating IR sensors mounted on AVs. Highly parallelized thermal and IR sensor models for large-domain AV operations will be presented in this paper.

The proliferation of small or micro unmanned aerial vehicles (UAV) gives rise to a potential threat for both public and military security. The small footprint and unpredictable dynamics of drones make detection and tracking difficult. Traditional methods of defence and protection may be ineffective against this new danger. This paper presents the work on developing DroneSwatter, a counter unmanned aerial system developed to track, follow, and take down a drone threat (Target Drone) using an agile, low-cost drone interceptor (Hunter Drone). The DroneSwatter project aims to apply machine learning techniques for counter-drone scenarios. Detection tasks are performed using deep learning detection algorithms. Simulation is used to build a tracking control model via proportional-derivative (PD) and machine learning algorithms. Optical pursuit based on images collected from the onboard camera of a Hunter Drone is implemented to track a Target Drone. Field experiments were conducted to test the feasibility and functionality of the current software and hardware methods for the DroneSwatter system. A benchmark was established by flying a target drone in designed patterns and the performance of the DroneSwatter tracking system was evaluated based on what speeds the Hunter Drone could follow the Target Drone in the field testing.

Quadruped locomotion and gait patterns as trotting, galloping, trot running has been investigated and applied to a variety of existing quadruped robots such as Big Dog from Boston Dynamics, A1 Robot dog from Unitree. Most of these studies are based either on biology inspired gaits or the best possible locomotion that can be performed by the individual robot with its pre-set mechanics and its availability of the degree of freedoms. While these are already available as their basic modes, a wide number of researchers are investigating locomotion via deep neural nets. These are making headlines in the research community for efficiency of use, and yet the explainability is lacking in most cases. Just like a Large Language Model might give spurious results here and there for basic common sense questions, these deep neural nets also make errors with unknown interpretability to the inputs. Regarding training, they require careful tuning of hyperparameters and training with a number of parameters unknown to user predictions. For example, on the field we might have a terrain which is flat for a certain length, in addition to a rocky climb, followed by a slippery slope. The combinations are as many as possible and the existing state of the art is heavily depending on human intervention and training predictions to handle the change of modes of the gait patterns that can fit into the terrain underneath. In this paper, we develop a novel embodied explainable machine learning algorithm which can help minimize the training as well as human intervention when autonomous operations are required. Specifically, we utilize the Markov Decision Process (MDP) along with rules set forth by DARPA in the Explainable AI (XAI) research. The XAI research enables us to generate textual explanations of the behavior by utilizing the MDP and reinforcement learning to generate mission oriented and situation aware cost functions along with the ones which are already pre-programmed. We validate our hypothesis in real hardware across different conditions.

Recent years have seen the emergence of novel UAV swarm methodologies being developed for numerous applications within the Department of Defense. Such applications include, but are not limited to, search and rescue missions, intelligence, surveillance, and reconnaissance activities, and rapid disaster relief assessment. Herein, this article investigates an initial implementation of learning UAV swarm behaviors using reinforcement learning (RL). Specifically, we present a study implementing a leader-follower UAV swarm using RL-learned behaviors in a search-and-rescue task. Experiments are performed through simulations on synthetic data, specifically using a cross-platform flight simulator with Unreal Engine virtual environment. Performance is assessed by measuring key objective metrics, such as time to complete the mission, redundant actions, stagnation time, and goal success. This article seeks to provide an increased understanding and assessment of current reinforcement learning strategies being developed for controlling (or at a minimum suggesting) UAV swarm behaviors.

In this paper, the design of vehicle door weight is minimized using a newly developed machine learning algorithm that is referred to as Grey Wolf optimizer (GWO). GWO is a metaheuristic technique that show good and robust performance in solving optimization applications. It is a nature-inspired algorithm that is based on the hunting grey wolf while hunting and catching the prey. The algorithm is known to have simple, yet efficient, structure. On the other hand, the design of the car’s door during an impact is known to be a multi-objective optimization problem that is based the European Enhanced Vehicle-Safety Committee. The algorithm needs to minimize the door weight will be satisfying several constrains. The design depends on 11 parameters including the B-pillar inner, B-pillar reinforcement, floor side inner, cross members, door beam, door beltline reinforcement, roof rail, materials of B-pillar inner floor side inner barrier height and hitting position. Monte Carlo simulation is used to test the method and its robustness and stability.

Low-level helicopter operations are typical military missions, for example in forward air medical evacuation missions. Up to now, these types of missions are either carried out in manual flight or with a rather conservative automation on rare platforms. For the latter, the only possible intervention method for the pilot is to manually take over the helicopter. It can be envisaged that future platforms will provide more dynamic low-level automation capability. At the same time, it is very likely that the crew will have to fulfill other tasks like managing unmanned systems. This will fully decouple the pilot from the flight control task for periods of time and reduce the ability to quickly take over the helicopter under threat conditions. Therefore, automation functions need to be available to avoid threats and alter the planned path on short notice which further reduces a comprehensive system understanding and the ability for the pilot to intervene. This paper presents a multimodal cueing concept for human-machine shared control under automatic trajectory following low-level operation, which is being developed within the project “US-German Advanced Technologies for Rotorcraft Project Agreement”. The system enables the helicopter to follow planned low-level paths and provides the pilot with tactile, auditive and visual cues. The trajectory following function is complemented with a collision avoidance method to create a “carefree” automation. Intermediate results for the multimodal cueing and interaction concept are presented, which were gathered from validation sessions and workshops with expert pilots at DLR’s Air Vehicle Simulator (AVES).

The concept of enabling drone-swarm engagement simulations using particle-dynamics models and near-neighbors tracking algorithms, motivated by SDI battle management, is examined. The general approach of using particle-dynamics models and near-neighbors tracking algorithms for modeling drone-swarm engagements is similar to nonequilibrium molecular-dynamics modeling of mixing dissimilar particulate materials. With respect to particle-dynamics representation of swarm-engagements, fundamental quantities that can represent characteristics of drone interactions, are interparticle potential functions, which are a function of drone-drone separation, the types of drones interacting, and the nature of the interaction. These potential functions provide formal representation of both deterministic and non-deterministic dronedrone interaction scenarios. The complexity of drone-swarm engagements, similar to that of SDI scenarios, characterized by small time-periods of engagement, multiple types of blue-red force interactions, and the requirement of near-neighbor target tracking, suggest that such a tool be necessary. The utility of the tool in creating potential-theory based control algorithms for swarm-on-swarm engagements is demonstrated using particle-dynamics simulations.

One crucial capability of unmanned systems is their ability to make decisions and inferences like humans. In this paper, we develop a novel logical system that imitates the way humans engage in reasoning with statements possessing varying degrees of ambiguity and unpredictability. Our proposed logical system is constructed using an axiomatic approach with self-evident rules, which allows us to define statemental operations and logical equivalence without the need for a concept of truth valuation. Our logical system includes both statemental algebra and truth calculus, which are designed to manipulate statements and assess their credibility. We believe that our proposed logical system has the potential to enhance the intelligence of unmanned systems.

In today’s world, intelligent systems are expected to handle diverse types of events that can be represented as dichotomy, probabilistic, fuzzy, or mixed statements. This paper proposes an analytic framework called “statemental analysis,” which deals with the operations and truth values of finite and infinite sequences of statements of various categories. The framework treats vagueness and randomness in a unified manner using the concepts of statemental algebra and truth calculus. It allows for generalization of many results in probability theory to statements with significant implications. We demonstrate the utility of the proposed framework with an example of reliability analysis of uncertain systems.

The limitations of current logical systems in addressing vagueness and randomness restrict their application in the field of artificial intelligence (AI). To address this issue, we introduce an axiomatic mathematical system called statemental credibility logic (SCL). First, we jointly define logical operations and equivalence. Then, we introduce the concept of a truth measure. By extending the law of excluded middle and the law of non-contradiction, and using them in conjunction with other self-evident rules in the axiomization of SCL, we can incorporate classical propositional calculus and probability theory. Furthermore, SCL can handle statements that exhibit varying degrees of vagueness and randomness. We also extend SCL to deal with predicates that involve uncertainty. We believe that SCL and its extension have the potential to improve the reasoning capabilities of intelligent systems.

The average performance of uncertain dynamic discrete-event systems remains a persistent concern in the field of control engineering. In this paper, we propose to use a Monte Carlo method to analyze uncertain systems by determining whether their average performance exceeds an acceptable level. Specifically, we formulate the performance analysis as a problem of statistical hypothesis testing of mean values. Using a mean-preserving transform, we convert this problem into one of statistical hypothesis testing of probabilities, which can be solved using our adaptive Monte Carlo test. This test is based on Wald’s sequential probability ratio (SPRT). We demonstrate the applicability of our method by investigating the average performance of a control system with parametric uncertainty.

A conflict of interest exists between alerting the public over cellular communications and ensuring drivers operate vehicles hands-free. Active shooters, criminal pursuits, and kidnappings on our roadways can create dangerous conditions that may lead to fatalities and serious injuries. Thus, sharing suspected criminal images within future vehicular ad-hoc networks may improve transportation safety and help authorities crowdsource feedback for apprehending criminals. We present a novel system allowing broadcasting images of suspected criminals and related vehicle information into the vehicular adhoc network as a public emergency notification system for the population on the road. Image-based public safety messages broadcast from the infrastructure-to-vehicle (I2V) link into the vehicular ad-hoc network on the roadways. Similar studies show issues with bandwidth limitations of the I2V link and ensuring a high packet reception rate (PPR) for vehicle density in multimedia messages. In this work, we transmit images instead of video to mitigate network challenges and increase packet reception probability. We successfully transmit images within the Dedicated Short-Range Communication (DSRC) vehicular ad-hoc network. The results show that the vehicular ad-hoc network can receive traveler information messages with large image payloads even at high vehicle densities. In addition, the proposed system does not negatively affect the PPR of the existing vehicular ad-hoc network.

Drones can be a candidate technology to support on-demand connectivity in scenarios where networking infrastructure can be difficult to deploy or can be expensive. In a context where many terrestrial nodes such as IoT devices need to send data about environment, drones can move in the considered area to collect data sensed by IoT devices. In this case, Delay Tolerant Network is an architecture that can support the intermittent communication between drones and IoT devices and huge amount of data can be collected and forwarded to remote servers to be analyzed. At this purpose, social paradigm and opportunistic communication between IoT devices and Drones can improve the data collection process. Social rank in the IoT networks is considered as a metric to drive the Drone path planning for providing an efficient connectivity and coverage.

Recent interest in unmanned aerial vehicles (UAVs) has grown due to the wide range of possible civilian uses for these aircraft. However, present robot navigation technologies still need to be improved in various situations. Researchers are particularly interested in the 'Sense and Avoid' capacity as a critical issue. UAVs operating in civilian areas must have this functionality to do so safely. Numerous path planning and navigation algorithms have been developed for autonomous decision-making and control of UAVs. These path-planning algorithms are divided into either heuristic and non-heuristic or accurate methods. Both existing UAV route planning algorithms for the first and second techniques will be thoroughly compared in this work. Each algorithm is put through its paces in three diverse obstacle scenarios. Each method has been evaluated under various global and local obstacle information availability conditions while comparing the computational time and solution optimality.

Unintended Radiated Emissions (URE) are emitted by all electrical devices and can be analyzed to determine changes in state on the emitting device. This paper aims to analyze UREs from a target device and classify if the device has changed operating states given a new measurement of the UREs. The UREs for a Raspberry Pi in different operating states are collected and analyzed. We detect if the target device changes operating states using a recurrent neural network to predict the URE spectrum power given multiple previous URE measurements. The predicted emissions are then compared to the measured emissions and a deep neural network classifies the measured emissions as a state change. Multiple other model types are compared including statistical classifiers and more complex machine learning models and our proposed model is found to perform the best in our dataset. We achieved an accuracy and F1-score of 90% in our real-world dataset.