Learning and approximation in database for feedforward control of flexible manipulator

Abstract

The author considers a new method for feedforward programmed control of a robotic system with complex dynamics. The method uses a learning procedure and a database for previously learned controls rather than a mathematical model of the system. For point-to-point control of a two-link flexible-joint manipulator, the method results in a fairly little amount of learning and database storage memory.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>