Instant prediction for reactive motions with planning

Reactive control and planning are complementary methods in robot motion control. The advantage of planning is the ability to find difficult solutions, optimize trajectories globally and not getting stuck in local minima but at higher computational cost. On the other hand, reactive control can handle dynamic or uncertain environments at low computational cost, but may get stuck in local minima. In this paper, we propose a new approach to integrate both reactive control and planning using a short time prediction. The system is mainly composed of a predictor, a planner and a reactive controller. The system uses the planner to modify the target for the reactive controller. The predictor simulates the robot future states and evaluates the reactive motion to trigger the planner in advance. The latency due to the high computational costs for planning is compensated since the motion is already simulated, therefore, the resulting motion is smoother than without the predictor. When the system's environment becomes more uncertain and dynamic, the system works reactively and iterates faster so that the system adapts to the environment automatically. We tested the scheme in a simulator and realized it on our humanoid robot ASIMO.