3 METHODS OF TRAINING MULTILAYER PERCEPTRONS TO MODEL A ROBOT SENSOR

This paper discusses three methods of training multi-layer perceptrons (MLPs) to model a six-degrees-of-freedom inertial sensor. Such a sens or is designed for use with a robot to determine the location of objec ts it has to pick up. The sensor operates by measuring parameters rela ted to the inertia of an object and computing its location from those parameters. MLP models are employed for part of the computation. They are trained to output the orientation of the object in response to an input pattern that includes the period of natural vibration of the sen sor on which the object rests. After reviewing the working principle o f the sensor, the paper describes the three MLP training methods (back propagation, optimisation using the Levenberg-Marquardt algorithm, evo lution based on the genetic algorithm) and presents the experimental r esults obtained.