国产欧美精品一区二区,中文字幕专区在线亚洲,国产精品美女网站在线观看,艾秋果冻传媒2021精品,在线免费一区二区,久久久久久青草大香综合精品,日韩美aaa特级毛片,欧美成人精品午夜免费影视

Aiming at the problem that the quadruped robot can not continue to operate effectively and independently, the adaptive control model based on hierarchical learning is proposed. The model structure consists of an upper state policy controller (SDC) and a lower base motion controller (BDC). The SDC estimates the expected motion sub-strategy of the robot's legs and posture, and the BDC sub-motion strategy is activated to control the robot to express the athletic behavior. Damage to the robot is manifested in the complete loss of athletic ability in any leg. The adaptive control of the model is reflected in the robot's self-adjusting strategy after the leg fails. In Unity3D, a four-legged robot with anti-joint multi-degree of freedom is built. The SDC monitors the state of the robot and adjusts the strategy. The BDC output gives the joint PD controller speed control. Simulation and experimental results show that the model shows a fast and stable effect on the robot's self-adjusting motion strategy.