Robotic task level programming using neural networks

Robot programming is a difficult, complex and time consuming operation. It consists of three main stages, the definition of points/locations, program coding (including error planning) and finally program proving. Due to problems associated with each of these stages, alternative techniques are sought to reduce the programming duration, to simplify the programming complexity and to improve the success of the application. Task level programming (TLP), which has been a goal for robotics researchers for many years, aims to reduce the complexity of the robot program and to ease the implementation of the operation by embedding significant knowledge into the robot and its control system. The aim of this paper is to introduce a novel technique which enables the implementation of a TLP system. The paper identifies the generation of the task description which uses artificial neural networks (ANN) to recognise both the significant aspects of the task and to operate and continuously monitor the robot during the complex movements required of a mechanical assembly task.