Trajectory optimization of a shuttle mounted robot
Markus Glocker, Manuel Vogel, Oskar von Stryk · TUbilio (Technical University of Darmstadt) · 1999
Optimal path planning for a shuttle mounted remote manipulator system is in-vestigated. Shuttle, robot and load are modelled as a multi-body system of rigid bodies whereby computer-aided modelling is used to dervive the equations of motion. Minimum time and minimum energy trajectories of the robot are inves-tigated which re-establish the shuttle's attitude at the beginning of the maneuver also at the end. The resulting constrained optimal control problems are solved by a recent version of the optimal control software DIRCOL [12] which is based on discretization of state and control variables, collocation and nonlinear pro-gramming (NLP). An improvement of internal gradient approximations, the full utilization of the discretized problem structure as well as of the sparsity of the NLP Jacobian are important factors improving the efficiency and robustness of the method. The resulting NLPs are solved by a recent version of the large-scale SQP method SNOPT [5]. A three-dimensional animation of the space robot trajectories is obtained using the graphics library SIGMA [1].