TCP position tolerances in joint torque minimizations

A new algorithm for link and path optimizations of manipulator arms is formulated. The algorithm considers precision in tool center point (TCP) positioning. A penalizing function is introduced to allow divergence from the desired TCP locations. The ever increasing operational demands of robots presuppose high power drives which tends to adversely affect the size of gearing and magnitudes of stress levels in the moving parts. In order to strike a balance between the conflicting demands of rapid task execution and the desire to limit available power, the algorithm was formulated using Newton-Euler dynamics to acquire the joint torques as functions of the parameters, which define the motion. By varying the parameters that define the motion (position, velocity, acceleration and link kinematics) while keeping the cycle time constant, it is possible to minimize the maximal value of the joint torques required for a sequential picking motion, transportation, placing motion and transportation back to the next picking location. Cycle optimizations are made for the Linkoping Flywheel Robot with only revolute joints (LFR), the new version of the Linkoping Flywheel Robot (LFRp) with one prismatic joint, the SCARA concept and a cylindrical robot concept. The simulations show that by tolerating a position error of 1 mm of the TCP, the required torque could be reduced significantly compared to that required for a position error of 0.1 mm. The conceptual robots exist in mathematical models in Matlab and SIMULINK