The Use of a Simple Neural Network in the Motion Control of Unmanned Undersea Vehicles
M.J. Norton · 2005
The kinematic motion control of multiple simulated autonomous vehicles is described. This control is accomplished via the use of a simple neural network which accepts range, bearing, and heading information to produce a speed and rudder angle. The controller is explained and used to show several kinds of behavior in two autonomous vehicles (UUY and Intelligent Torpedo). The following material describes work in progress and should not be taken to be complete or finished. The neural network techniques are being used in a simplistic manner which do not currently adapt or learn. This capability will be added as research continues.