We fuse a self-organizing hyperellipsoid clustering (HEC) Kohonen neural network with the Julier-Uhlmann-Kalman filter (JUKF) to perform map building and low-level position estimation. The HEC Kohonen uses the Mahalanobis distance to learn elongated shapes (typical of sonar data) and obtain a stochastic measurement of data-node association. The number of nodes is regulated by measuring how well a node model matches its associated data. The HEC Kohonen can handle high-dimensional problems and can be generalized to other pattern recognition problems. The JUKF compliments the HEC Kohonen in that it performs low-level (nonlinear) tracking more efficiently and more accurately than the extended Kalman filter. By estimating and propagating error covariances through system transformations, the JUKF eliminates the need to derive Jacobian matrices. The inclusion of stochastic information inherent to the HEC map renders the JUKF an excellent tool for our HEC-based map building, position estimation, motion planning and low-level tracking.

This paper documents how two mobile robots can share knowledge about their environment. The two mobile robots we use have different sensor configurations, drive systems and other physical attributes including weight and size. "Knowledge" is generated by the region-feature neural network (RFNN), and can be transferred on two general levels: (1) a complete transfer of a matured neural network; and (2) a transfer of matured features. This transferred knowledge can also be "tuned" with and without locking the feature level synaptic weights. We examine the impact and feasibility of sharing (on both levels, with and without locking features) neural network modules trained on actual sonar data in the global self-localization (GSL) problem. Significant reductions in training time are realized and presented. We also describe the neural network architecture and our general approach to solving the GSL problem in a time-, translation- and rotation-invariant way.

In this paper we show how a self-organizing Kohonen neural network can use hyperellipsoid clustering (HEC) to build maps from actual sonar data. Since the HEC algorithm uses the Mahalanobis distance, the elongated shapes (typical of sonar data) can be learned. The Mahalanobis distance metric also gives a stochastic measurement of a data point's association with a node. Hence, the HEC Kohonen can be used to build topographical maps and to recognize its own topographical cites for self-localization. The number of nodes can also be regulated in a self-organizing manner by using the Kolmogorov-Smirnov (KS) test for cluster compactness. The KS test determines whether a node should be divided (mitosis) or pruned completely. By incorporating principal component analysis, the HEC Kohonen can handle problems with several dimensions (3D, time-series, etc.). The HEC Kohonen is also multifunctional in that it accommodates compact geometric motion planning and self-referencing algorithms. It can also be used to solve a host of other pattern recognition problems.

Autonomous vehicles require that all processes be efficient in time, complexity and data storage. In fact, an ideal system employs multifunctional models where ever possible. This paper presents the region- and feature-based neural network (RFNN) as a viable pattern analysis process engine for solving a variety of problems with a single math model. The RFNN employs receptive fields and weight sharing which compensate for noise, minor phase shifts and occlusions. The RFNN also utilizes greedy adaptive learning rates and mature feature preservation to expedite the overall training process. A novel ad hoc approach called "shocking" is used to solve the instability problem inherent to greedy adaptive learning rates. The basic RFNN "feature" is grounded in computer vision morphology in that the neural network autonomously learns subpatterns unique to various problems. This paper comprehensively describes the flexible RFNN architecture and training process and presents two problems that can be solved by the RFNN: sensor pattern-recognition and traffic sign recognition.

An approach to global self-localization for autonomous mobile robots has been developed using self-organizing Kohonen neural networks. This approach categorizes discrete regions of space using mapped sonar data corrupted by noise of varied sources and ranges. Our approach is similar to optical character recognition (OCR) in that the mapped sonar data can, over time, assume the form of a character unique to that room. Hence, it is believed that an autonomous vehicle can be capable of determining which room it is in based on mapped sensory data ascertained by wandering through and exploring that room. With some pre-processing and a robust explore routine, the solution becomes time-, translation- and rotation-invariant.