In this paper, we present a new approach for detecting trucks transporting illicit radioactive materials using radiation data. The approach is motivated by the high number of false alarms that typically results when using radiation portal monitors. Our approach is a three-stage anomaly detection process that consists of transforming the radiation sensor data into wavelet coefficients, representing the transformed data in binary form, and detecting anomalies among data sets using a proximity-based method. The approach is evaluated using simulated radiation data, and the results are encouraging. From a transportation security perspective, our results indicate that the concomitant use of gross count and spectroscopy radiation data improves identification of trucks transporting illicit radioactive materials. The results also suggest that the use of additional heterogeneous data with radiation data may enhance the reliability of the detection process. Further testing with real radiation data and mixture of cargo is needed to fully validate the results.

Location-awareness is an important requirement for many mobile wireless applications today. When GPS is not applicable because of the required precision and/or the resource constraints on the hardware platform, radio interferometric ranging may offer an alternative. In this paper, we present a technique that enables the precise tracking of multiple wireless nodes simultaneously. It relies on multiple infrastructure nodes deployed at known locations measuring the position of tracked mobile nodes using radio interferometry. In addition to location information, the approach also provides node velocity estimates by measuring the Doppler shift of the interference signal. The performance of the technique is evaluated using a prototype implementation on mote-class wireless sensor nodes. Finally, a possible application scenario of dirty bomb detection in a football stadium is briefly described.

One of the most important features of artificial neural networks in emerging, brain-inspired, nanoarchitectural design is their inherent ability to perform massively parallel, nonlinear signal processing. When operating in a system-wide asynchronous regime, such networks may exhibit a phenomenon referred to as "computational chaos", which impedes the efficient retrieval of information usually stored in the system's attractors. We illustrate the emergence of computational chaos from fixed point and limit cycle attractors for node communication delays in a widely used neural network model. In particular, the complete Lyapunov spectrum associated with the network dynamics is computed, and conditions that prevent the emergence of chaos are derived.

A feedforward multilayer neural net is trained to learn the correspondence between seismic data and well logs. The introduction of a virtual input layer, connected to the nominal input layer through a special nonlinear transfer function, enables ultrafast (single iteration), near-optimal training of the net using numerical algebraic techniques. A unique computer code, named DeepNet, has been developed, that has achieved, in actual field demonstrations, results unattainable to date with industry standard tools.

We address the model acquisition problem for an unknown planar polygonal terrain by a team of two or three point robots. The robots are equipped with visual sensors which acquire all visible parts of the terrain by scan operations executed from their locations. The robots communicate with each other via wireless connection. The performance is measured by the number of the sensor (scan) operations. We employ the restricted visibility graph methods an a hierarchical setup. For terrains with convex obstacles and for teams of n (where n=2, 3) robots, we prove that the sensing time is reduced by a factor of 1/n. For terrains with concave corners, the performance of the algorithm for the n(2, 3) robot team is expressed in terms of the sizes of n-connected components, and the sizes and diameters of (n-1) or less connected components.

We consider the problem of learning functions based on finite samples by using feedforward sigmoidal networks. The unknown function f is chosen from a family that has either bounded modulus of smoothness and/or bounded capacity. The sample is given by (X/sub 1/, f(X/sub 1/)), (X/sub 2/, f(X/sub 2/)), ...(X/sub n/, f(X/sub n/)). Where X/sub 1/, X/sub 2/, ..., X/sub n/, are independently and identically distributed according to an unknown distribution P/sub X/. General results guarantee the existence of a neural network, f/sub w/*, that best approximates f in terms of expected error. However, since both f and P/sub X/ are unknown, computing f/sub w/* is impossible in general. We propose to compute probability and approximately correct (PAC) approximations to f/sub w/*, based on alternative estimators, namely: 1) the nearest neighbor rule, 2) local averaging, and 3) Nadaraya-Watson estimators, all computed using the Haar system. We show that given a sufficiently large sample, each of these estimators guarantees a performance as close as desired to that of f/sub w/*. The practical importance of this result sterns from the fact that, unlike neural networks, the three estimators above are linear-time computable in terms of the sample size.

We present two classes of convergent algorithms for learning continuous functions and regressions that are approximated by feedforward networks. The first class of algorithms, applicable to networks with unknown weights located only in the output layer, is obtained by utilizing the potential function methods of Aizerman et al. (1970). The second class, applicable to general feedforward networks, is obtained by utilizing the classical Robbins-Monro style stochastic approximation methods (1951). Conditions relating the sample sizes to the error bounds are derived for both classes of algorithms using martingale-type inequalities. For concreteness, the discussion is presented in terms of neural networks, but the results are applicable to general feedforward networks, in particular to wavelet networks. The algorithms can be directly adapted to concept learning problems.