In this paper a control strategy for a redundant PUMA robot is presented. This robot is obtained from the conventional PUMA robot by addition of a joint parallel to the elbow joint. For redundancy resolution the following approach is chosen. From the position and pointing direction of the end effector of the PUMA robot the position and orientation of the fourth link of a 4R-manipulato? is calculated and the redundancy is resolved for this manipulator. This is done by adding an equation to the relationship between the joint angles and position and orientation of the fourth link. By this approach a control strategy is derived that allows motions of the end effector of the PUMA robot in a large part of its work space and shows repeatable behaviour. Furthermore, the redundancy is utilized so that the fiexture of the wrist is kept small.

Planning appropriate trajectories can significantly increase the productivity of robot systems. To plan realistic time-optimal trajectories, the robot dynamics have to be described precisely. In this paper, a neural network based algorithm for tim e-optimal trajectory planning is introduced. This method utilises neural networks for representing the inverse dynamics of the robot. As the proposed neural networks can be trained with data obtained from exciting the robot with given torque inputs, they will capture the complete dynamics of the robot system. Threfore, the trajectories generated will be mo re realistic than those obtained by using nominal dynamic equations based on nominal parameters. Time-optimal trajectories are generated for a PUMA robot to demonstrate the proposed method.