Dynamic footstep planner algorithms for humanoid soccer robots

We present a footstep planner for biped robots using the method of sequence approximation .The error caused by executing the previous step is able to accumulate to the input data of the next step, which achieves the self-correction of the planner. Meanwhile the input data of every step is directly gained by formulas, so the computational complexity is low. And using this method robots can track dynamic targets easily. At last, experimental results show the validity of the algorithm on the RoboCup3D simulation platform.