APPLICATION OF DYNAMIC PROGRAMMING TO ROBOT KINEMATICS.

The authors present a dynamic programming (DP) algorithm to calculate robot manipulator kinematics and plan optimized trajectories. The algorithm optimizes desirable cost functions such as absolute accuracy or time of travel. They show that when necessary, suboptimal (higher cost) alternative can be advantageously accepted. For example, if an optimal trajectory is blocked by an obstacle, the robot can reach the destination by following a suboptimal path. The authors also present computer simulations of many diverse applications of DP to robot kinematic problems. The DP algorithm requires more computations than conventional techniques, but permits optimization of objective criteria and flexibility in path planning.