Configuration Space Motion Planning for a RoboticManipulator : Using Optimization Methods for Multipoint Trajectories and a Virtual Vechicle Approach for Work Space Constraints

Sammanfattning: A common method to design trajectories is by cubic splines to interpolate a set of via-points for a robot. The trajectories are optimized with respect to an objective function including the total time with a set of kinematic constraints. The execution time for this problem is longer then the sampling time of an embedded computer and restricts the use of this method in on-line applications. To modify this procedure to be usable on-line a point-to-point motion planner is developed using the output from the optimization programming at each iteration. This point-to-point motion planner runs on a separate process and contains a virtual vehicle to restrict the end-effector speed in its operational space. This results in an adaptive motion planner capable of handling kinematic actuator constraints and end-effector speed constraints in the operational space. The thesis ends with results from simple simulations demonstrating the effects.