Robert S. Woodley

A testbed system for nonlinear or intelligent control

There has been a large amount of publications concerning nonlinear and intelligent control in recent years. It has become an important field of research. As each new algorithm becomes available, the need for realistic and accurate test bed systems is apparent. Some previous simulation models have not been real world systems, or a simplification of a much larger system. It is the goal of this document to present a much more complete description of one real world model so that others may use this model to test their control algorithms. The system analyzed in this document is a trailer truck system. The model, based on the physical system, extends the models of previous publications. A full dynamical representation is given. Tests show the models accuracy against the actual system.

Neural network based control for a backward maneuvering trailer truck

The control of non-linear systems using neural networks has gained increasing interest in recent years. The non-linear capabilities of neural networks can be utilized for a large number of non-linear systems that are not controllable by linear techniques. However, controllers based on neural networks are quite difficult to design due to the lack of good training data. The physical plant developed and designed in this work is a scale model of a trailer truck. A neural network based controller is to be developed to drive the truck backwards from any initial condition to a loading dock. One approach to solve this control problem is to design a valid path, and then track the path. Since the valid path depends on initial conditions, a neural network is trained to generate selective points on the path for any initial condition. The path data used to train the neural network are obtained from software based on a previously published work on similar systems. Based on generated path data, a control is designed and is able to successfully maneuver the truck to the loading dock from any initial condition that was within a particular region. The region represented a group of non-trivial trajectories that the truck is to follow.

An adaptive beamforming by a generalized unstructured neural network

In this paper, an adaptive array beamforming by an unstructured neural network based on the mathematics of holographic storage is presented. This work is inspired by similarities between brain waves and the wave propagation and subsequent interference patterns seen in holograms. Then the mathematics to produce a general mathematical description of the holographic process is analyzed. From this analysis it is shown that how the holographic process can be used as an associative memory network. Additionally, the process may also be used a regular feed-forward network. The most striking aspect of these network is that, using the holographic process, the apriori knowledge of the system may be better utilized to tailor the neural network for an adaptive beamforming problem. This aspect, makes this neural network formation process particularly useful for the beamforming.

A generalized unstructured artificial neural network architecture: a first study

We present an unstructured neural network based on the mathematics of holographic storage. This work was inspired when we discovered there are similarities between brain waves and the wave propagation and subsequent interference patterns seen in holograms. We then analyzed the mathematics to produce a general mathematical description of the holographic process. From this analysis we are able to show how the holographic process can be used as an associative memory network. Additionally, the process may also be used as a regular feedforward network. The most striking aspect of these networks is that, using the holographic process, the a priori knowledge of the system may be better utilized to tailor the neural network for a particular problem. This aspect, makes this neural network formation process particularly useful for control.