给出了一种三维环境下双足行走的参数化步态规划方法,建立了仿人机器人13质量块约束动力学模型.考虑单腿支撑和双腿支撑无冲击连续切换的六点边界约束条件、可行步态物理约束条件以及ZMP稳定性约束条件,以关节输出力矩函数的二次型积分值最小作为优化指标,采用参数化步态优化方法,将复杂关节轨迹的规划问题转化为分段多项式系数组成的有限参数向量的优化问题,得到了快速和慢速两组光滑无振动的优化步态.仿真和样机实验验证了该方法的有效性.
A parametric gait planning of 3D biped walking is proposed and a 13-rigid-body constrained dynamic model of humanoid robot is also established. With the considerations of six-point boundary constraints of successive and impact-less steps including SSP (single-support-phase) and DSP (double-support-phase), physical constraints of feasible gait and ZMP (zero moment point) stability constraints of locomotion, tow groups of slow and fast walking speed with smooth and nonjerky optimal joint trajectory curves to minimize integral quadratic amount of joint driving torques are obtained by adopting the parametric gait optimization approach, which makes the complicated joint trajectory planning problem transform into the optimization problem of limited parametric vectors composed of piecewise polynomial coefficients. The effectiveness of this method is confirmed by dynamic simulations and walking experiments on an actual humanoid robot.