Real-Time Motion Planning for Humanoid Robot Based on Embedded Vision System

A control and decision system based on embedded system are designed with the features of humanoid robot. The video which is captured by image processing system is recognized through improved look-up table method. A smooth motion planning algorithm for humanoid robot is proposed and next several motions for humanoid robot are achieved by the local motion planner based on the minimum consuming energy. For the competition rules of sprint held in FIRA (federation of international robot-soccer association), decision is made for humanoid robot to plan the motions with the method of gait generation off-line and adjustment on-line. Simulations for image processing and sprint experiments on real humanoid robot HIT-2 are achieved to verify the algorithm and decision present.