Gait Planning Algorithms Of The Humanoid Robot And Its Realization

Gait planning is an important topic which influences the technology development in the field of humanoid robots research. It's the theoretical basis and key technology to realize stable biped walking. In this thesis, the author performed deep researches on gait planning algorithms of the humanoid robots, based upon the mechanical characteristics of the humanoid robot. He contrived several somewhat generalized planning algorithms. The gait planning algorithms were utilized in the bipedal locomotion experiments of a real humanoid robot and several kinds of stable walking were successfully realized.The main achievements of this thesis include:The kinetic and dynamic characteristics of the humanoid robot were carefully studied, and a computation engine that could be applied to other humanoid robots with the same DOF configuration was built up. The stability of the humanoid robot during its walking process could be precisely quantitatively described. And the driving torques of the joints of the robot could be computed.According to former researches by others, a generalized static walking gait planning algorithm based on interpolation optimizations of walking postures was put forward. According to different gait parameters, several gaits were generated, simulated and verified.Utilizing the joints' influences against the robot stability, a gait generation algorithm for fast and stable dynamic walking of the humanoid robot was contrived, which is based on optimization of the angular accelerations of the main support leg's joints. Several dynamic gaits were generated, simulated and verified.In response to the control characteristics of the real humanoid robot, an off-line gait compensation algorithm was put forth based on the information of the sensors in the robot and the gait generation algorithm. With walking experiments of the THBIP-I humanoid robot and the compensation of the theoretic gait data, the robot realized stable bipedal locomotion such as walking and going upstairs, proving the effectiveness of the gait planning algorithm and its realization method.