A computer simulation environment for the trajectory planning of complex robotic systems

This paper presents a trajectory-planning scheme and a simulation environment, for redundant manipulators with a treelike topology. The method is based on Liegeois' local trajectory-planning algorithm and combines obstacle avoidance and joint limit constraints. An underdetermined linear system of equations is first obtained from the velocity equations. Then, the singular value decomposition is used and a weighted minimum norm solution of the system is computed A correction to this solution is then generated as two joint velocity vectors lying in the nullspace of the Jacobian matrix and which are added to the minimum norm solution in order to avoid the obstacles and the joint limits. A simulator of mt complex 24-dof system is finally introduced and described The simulation environment can he used for the development of automatic trajectory planning or shared trajectory-planning strategies as well as for the training of operators. Examples of results obtained with the simulator are given.