行走能耗大是困扰四足机器人的难题，续航能力的不足极大地限制了机器人的任务执行能力。本项目以机器人能耗最优为目标，联合优化分布柔顺驱动系统刚度匹配与步态融合运动控制，获得能耗最优构型及驱动参数。深入研究基于足底力最优分配的柔顺控制策略，探索刚度、阻尼与运动性能的定量关系，提出分布柔性驱动器刚度辨识的方法。重点突破运动步态与支撑足端接触力的传递关系，构建基于足底力最优分配、腿部动力学及关节角度补偿的阻抗控制策略。利用随机线性互补问题的理论方法，通过定义步态运动中足底力与速度的互补关系，将腿足触地的离散动力学转换成连续可求解问题，实现相位混合变换的高速自由步态规划，提高机器人的行走效率。本项目从能耗最优角度出发，不仅为四足机器人的控制策略提出新的理论方法，而且提高其续航能力和行走效率，对四足机器人的应用发展具有重要意义。

The large energy consumption of walking is a problem that plagues the four-legged robot. The lack of endurance greatly limits the working ability of the robot. This project sets the optimal energy consumption of the robot as the goal, and jointly optimizes the stiffness matching and gait fusion motion control of the distributed compliant drive system to obtain the optimal configuration and driving parameters of the robot. In-depth study of the compliant control strategy based on the optimal distribution of the foot force is conducted, in order to explore the quantitative relationship between stiffness, damping and motion performance, and to propose a method for distributing the stiffness of the compliant actuation. The key breakthrough is the transmission relationship between the locomotion gait and the contact force of the foot, to construct the impedance control strategy based on the optimal distribution of the foot force, leg dynamics and joint angle compensation. Using the theoretical method of the stochastic linear complementarity problem, to define the complementary relationship between the foot force and the velocity in the locomotion gait. The discrete dynamics of the leg and foot touch is converted into a smooth continuous problem, and the high-speed free gait planning of phase hybrid transformation is realized, in order to improve the walking efficiency of the robot. From the perspective of optimal energy consumption, this project not only proposes a new theoretical method for the control strategy of the quadruped robot, but also improves its endurance and walking efficiency, which is of great significance to the application development of the quadruped robots.