We investigate the assignment of assets to tasks where each asset can potentially execute any of the tasks,
but assets execute tasks with a probabilistic outcome of success. There is a cost associated with each possible
assignment of an asset to a task, and if a task is not executed there is also a cost associated with the nonexecution
of the task. Thus any assignment of assets to tasks will result in an expected overall cost which we
wish to minimise. We propose an approach based on the Random Neural Network (RNN) which is fast and of
low polynomial complexity. The evaluation indicates that the proposed RNN approach comes at most within
10% of the cost obtained by the optimal solution in all cases.
The world that we live in is filled with large scale agent systems, from diverse fields such as biology, ecology or
finance. Inspired by the desire to better understand and make the best out of these systems, we propose an approach
which builds stochastic mathematical models, in particular G-networks models, that allow the efficient representation of
systems of agents and offer the possibility to analyze their behavior using mathematics. This work complements our
previous results on the discrete event simulation of adversarial tactical scenarios. We aim to provide insights into
systems in terms of their performance and behavior, to identify the parameters which strongly influence them, and to
evaluate how well individual goals can be achieved. With our approach, one can compare the effects of alternatives and
chose the best one available. We model routine activities as well as situations such as: changing plans (e.g. destination
or target), splitting forces to carry out alternative plans, or even changing on adversary group. Behaviors such as
competition and collaboration are included. We demonstrate our approach with some urban military planning scenarios
and analyze the results. This work can be used to model the system at different abstraction levels, in terms of the
number of agents and the size of the geographical location. In doing so, we greatly reduce computational complexity
and save time and resources. We conclude the paper with potential extensions of the model, for example the arrival of
reinforcements, the impact of released chemicals and so on.
The changing face of contemporary military conflicts has forced a major shift of focus in tactical planning and evaluation from the classical Cold War battlefield to an asymmetric guerrilla-type warfare in densely populated urban areas. The new arena of conflict presents unique operational difficulties due to factors like complex mobility restrictions and the necessity to preserve civilian lives and infrastructure. In this paper we present a novel method for autonomous agent control in an urban environment. Our approach is based on fusing terrain information and agent goals for the purpose of transforming the problem of navigation in a complex environment with many obstacles into the easier problem of navigation in a virtual obstacle-free space. The main advantage of our approach is its ability to act as an adapter layer for a number of efficient agent control techniques which normally show poor performance when applied to an environment with many complex obstacles. Because of the very low computational and space complexity at runtime, our method is also particularly well suited for simulation or control of a huge number of agents (military as well as civilian) in a complex urban environment where traditional path-planning may be too expensive or where a just-in-time decision with hard real-time constraints is required.
Denial of service attacks, viruses and worms are common tools for
malicious adversarial behavior in networks. Experience shows that
over the last few years several of these techniques have probably
been used by governments to impair the Internet communications of
various entities, and we can expect that these and other
information warfare tools will be used increasingly as part of
hostile behavior either independently, or in conjunction with
other forms of attack in conventional or asymmetric warfare, as
well as in other forms of malicious behavior. In this paper we
concentrate on Distributed Denial of Service Attacks (DDoS) where
one or more attackers generate flooding traffic and direct it from
multiple sources towards a set of selected nodes or IP addresses
in the Internet. We first briefly survey the literature on the
subject, and discuss some examples of DDoS incidents. We then
present a technique that can be used for DDoS protection based on
creating islands of protection around a critical information
infrastructure. This technique, that we call the CPN-DoS-DT
(Cognitive Packet Networks DoS Defence Technique), creates a
self-monitoring sub-network surrounding each critical
infrastructure node. CPN-DoS-DT is triggered by a DDoS detection
scheme, and generates control traffic from the objects of the DDoS
attack to the islands of protection where DDOS packet flows are
destroyed before they reach the critical infrastructure. We use
mathematical modelling, simulation and experiments on our test-bed
to show the positive and negative outcomes that may result from
both the attack, and the CPN-DoS-DT protection mechanism, due to
imperfect detection and false alarms.
In many critical applications such as airport operations (for capacity planning), military simulations (for tactical training and planning), and medical simulations (for the planning of medical treatment and surgical operations), it is very useful to conduct simulations within physically accurate and visually realistic settings that are represented by real video imaging sequences. Furthermore, it is important that the simulated entities conduct autonomous actions which are realistic and which follow plans of action or intelligent behavior in reaction to current situations. We describe the research we have conducted to incorporate synthetic objects in a visually realistic manner in video sequences representing a real scene. We also discuss how the synthetic objects can be designed to conduct intelligent behavior within an augmented reality setting. The paper discusses both the computer vision aspects that we have addressed and solved, and the issues related to the insertion of intelligent autonomous objects within an augmented reality simulation.
In this paper we apply neural network techniques and physically based models to determine the surface shape of chips from scanning electronic microscopy images. Deducing some specific feature's vertical cross-section within an integrated circuit from 2D top down scanning electron microscope images of the feature surface is a difficult `inverse problem' which arises in semiconductor fabrication. This paper refines our previous work on the reconstruction of semiconductor wafer surface shapes from top down electron microscopy images. One of the approaches we have developed directly maps from the CD-SEM intensity waveforms to line profiles. The other novel method we describe is based on an approximate physical model, where we assume a simplified mathematical representation of the physical process that produces the SEM image from the electron beam's interaction with the feature surface. Our results are illustrated with a variety of real data sets.
Modern video encoding techniques generate variable bit rates, because they take advantage of different rates of motion in scenes, in addition to using lossy compression within individual frames. We have introduced a novel method for video compression based on temporal subsampling of video frames, and for video frame reconstruction using neural network based function approximations. In this paper we describe another method using wavelets for still image compression of frames, and function approximations for the reconstruction of subsampled frames. We evaluated the performance of the method in terms of observed traffic characteristics for the resulting compressed and subsampled frames, and in terms of quality versus compression ratio curves with real video image sequences. Comparisons are presented with other standard methods.
We propose a method for learning and generating image textures based on learning the weights of a recurrent Multiple Class Random Neural Network (MCRNN) from the color texture image. The network we use has a neuron which corresponds to each image pixel, and the local connectivity of the neurons reflects the adjacent structure of neighboring neurons. The same trained recurrent network is then used to generate a synthetic texture that imitates the original one. The proposed texture learning technique is efficient and its computation time is small. Texture generation is also fast. This work is a refinement and extension of our earlier work where we considered learning of grey-level textures and the generation of grey level or color textures. We have tested our method with different synthetic and natural textures. The experimental results show that the MCRNN can efficiently model a large category of color homogeneous microtextures. Statistical feature extracted from the co-occurrence matrix of the original and the MCRNN based texture are used to confirm the quality of fit of our approach.
This paper presents a novel technique for texture modeling and synthesis using the random neural network (RNN). This technique is based on learning the weights of a recurrent network directly from the texture image. The same trained recurrent network is then used to generate a synthetic texture that imitates the original one. The proposed texture learning technique is very efficient and its computation time is much smaller than that of approaches using Markov Random Fields. Texture generation is also very fast. We have tested our method with different synthetic and natural textures. The experimental results show that the RNN can efficiently model a large category of homogeneous microtextures. Statistical features extracted from the co- occurrence matrix of the original and the RNN based texture are used to evaluate the quality of fit of the RNN based approach.
Linear Combination of Order Statistics (LOS) filters are a special case of the Choquet integral filters. LOS are a class of nonlinear filters parameterized by a set of n weights. Different values of the weights lead to different filters. Examples include the median and other order statistic filters, local averaging filters, and trimmed average filters. Differences of LOS filters have been used in the past as target detection filters by nonlinearly comparing a small, targets size region with the surrounding region. The delta operator, proposed by Gelenbe et. al. for land mine detection, can be represented as a special case of a difference of LOS operators. Weights of LOS operators can be determined by solving an optimization problem, represented as a quadratic program. In this paper, experiments are conducted in determining optimal differences of LOS operators using the DARPA backgrounds data. The results are that the delta-operator is the solution of the optimization problem for this data set.
In pattern recognition, it is crucial to be able to represent objects with feature that contain as much of the information as possible in compact form. A typical 8-bit grayscale digitized image can be sorted using M by N values that represent the intensity levels of individual pixels where M and N are image dimensions. Pattern recognition algorithms use various methods for feature extraction, like chain codes, Fourier descriptors, and invariant moments. We will propose features that will characterize objects much more efficiently. Our feature scan be viewed as basis functions that lead to a set of images within an equivalence class. In order to illustrate the method with an application, these features are then used to train a set of learning RNNs which can be used to detect targets within clutter with high accuracy, and to classify the targets or man-made objects from natural clutter. Experimental data from SAR imagery is used to illustrate the performance of the proposed algorithm. Currently, we are investigating the applicability of this approach to a set of GPR mine data.
Many multicast ATM switch architectures have been proposed which differ greatly in the method in which replication of cells is handled. Depending on the switch architecture, the processing overhead incurred due to the cell copy function may be non-negligible. We develop a queuing model for a multicast switching node which accounts for this overhead. In the model, the source sends some number of duplicate cells, each of which is replicated at the switch to yield the total number of required copies. We use constrained optimization to determine the optimal amount of source duplication which minimizes the mean response time of the system. It is found that higher resource duplication is favored when the copy function overhead is comparable to the service time of a single cell. The model is then enhanced to account for errors and retransmission for reliable multicast.
Detection and recognition of target signatures in sensory data obtained by synthetic aperture radar (SAR), forward- looking infrared, or laser radar, have received considerable attention in the literature. In this paper, we propose a feature based target classification methodology to detect and classify targets in cluttered SAR images, that makes use of selective signature data from sensory data, together with a neural network technique which uses a set of trained networks based on the Random Neural Network (RNN) model (Gelenbe 89, 90, 91, 93) which is trained to act as a matched filter. We propose and investigate radial features of target shapes that are invariant to rotation, translation, and scale, to characterize target and clutter signatures. These features are then used to train a set of learning RNNs which can be used to detect targets within clutter with high accuracy, and to classify the targets or man-made objects from natural clutter. Experimental data from SAR imagery is used to illustrate and validate the proposed method, and to calculate Receiver Operating Characteristics which illustrate the performance of the proposed algorithm.
KEYWORDS: Sensors, Land mines, Electromagnetic coupling, Electromagnetism, Mining, Robotics, Computer simulations, System on a chip, Explosives, Detection and tracking algorithms
Typically, a human agent or a robotic device may sweep a suspected minefield in a systematic up and down pattern to search for explosive mines with the help of an appropriate sensor or sensor system, such as an EMI (Electromagnetic Induction) sensor. In this paper we consider alternative search patterns which take advantage of a priori knowledge of the minefield. In previous work, a gradient based search algorithm has been designed and shown to be an effective search strategy using simulations on hypothetical minefield data. This paper considers a suite of fast search heuristics based on a hierarchical two level approach, and evaluates these algorithms with the realistic sensory data, specifically the Electromagnetic Sensory Data from DARPA. Heuristics considered include a hierarchical version of our gradient based algorithm, a nearest neighbor type greedy heuristic, and a heuristic which is inspired from an approximate solution of the traveling salesman problem.
The cost, and the closely related length of time, spent in searching for mines or unexploded ordnance (UXO) may well be largely determined by the number of false alarms. False alarms result in time consuming digging of soil or in additional multisensory tests in the minefield. In this paper we consider two area based methods for reducing false alarms. These are (1) the previously known 'declaration' technique, and (2) the new (delta) -Technique which we introduce. We first derive expressions and lower bounds for false alarm probabilities as a function of declaration area, and discuss their impact on receiver operation characteristic (ROC) curves. Secondly we exploit characteristics of the statistical distribution of sensory energy in the immediate neighborhood of targets and of false alarms from available calibrated data, to propose the (delta) -Technique which significantly improves discrimination between targets and false alarms. The results are abundantly illustrated with statistical data and ROC curves using Electromagnetic Induction Sensor data made available through DARPA from measurements at various calibrated sites.
Detecting objects in images containing strong clutter is an important issue in a variety of applications such as medical imaging and automatic target recognition. Artificial neural networks are used as non-parametric pattern recognizers to cope with different problems due to their inherent ability to learn from training data. In this paper we propose a neural approach based on the Random Neural Network model (Gelenbe 1989, 1990, 1991, 1993), to detect shaped targets with the help of multiple neural networks whose outputs are combined for making decisions.
Enhancing image quality and combining observations into a coherent description are essential tools in various image processing applications such as multimedia publishing, target recognition, and medical imaging. In this paper we propose two novel approaches for image enlargement and image fusion using the Random Neural Network (RNN) model, which has already been successfully applied to the problems such as still and moving image compression, and image segmentation. The advantage of the RNN model is that it is closer to biophysical reality and mathematically more tractable than standard neural methods, especially when used as a recurrent structure.
KEYWORDS: Probability theory, Computer simulations, Mathematical modeling, Asynchronous transfer mode, Switches, Radon, Switching, Local area networks, Statistical analysis, Computer engineering
The call establishment process in ATM networks can constitute a source of overload in the switches of the network. We simulate the call set-up phase of an ATM connection, including path selection, bandwidth reservation and call rejection, and simulate the flow of call establishment messages to estimate by simulation and analytically, the queue lengths of call establishment processing at the input, output and intermediate switches. A simplified analytical model is derived to estimate the queue lengths of call establishment jobs at each node of the network when advanced reservation with perfect information is implemented. The analytical model provides a lower bound to queue lengths, and remains within order of magnitude accuracy when the call request traffic is high.
Research on demining includes many different aspects, and in particular the design of efficient and intelligent strategies for (1) determining regions of interest using a variety of sensors, (2) detecting and classifying mines, and (3) searching for mines by autonomous agents. This paper discusses strategies for directing autonomous search based on spatio-temporal distributions. We discuss a model for search assuming that the environment is static, except for the effect of identifying mine locations. Algorithms are designed and compared for autonomously directing a robot, in the case where a single search engine carrying a single sensor.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.