Sensorimotor coordination and sensor fusion by neural networks

Abstract

A neural network method to solve the motion determination problem for robot end-effector motion without any calibration is presented. Two kinds of sensory data, i.e., camera images and laser ranges, are used as the input to a multilayer forward network in order to associate the direct transformation from the sensory data to the required motions. This provides a practical sensor fusion method. Using a recursive motion strategy, the requirement for the exactness of the learned transformation is relaxed. Experimental results verify the method.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>