中文摘要：

为解决双足跳跃机器人在突发状况下的适应性控制问题，依据生物中枢模式发生器（CPG）运动控制原理，建立了可变结构的 CPG 运动控制机制；模拟生物神经网络的自重构特性，在运动控制机制中添加了时间延迟响应环节，用于突发状况的判断与调节。在机器人肢体突发故障或运动受到限制时，此机制可通过对自身状态的感知自发地调整神经网络结构，以继续维持稳定的运动。最后，在仿真平台上验证了此控制机制在正常行走与肢体自由度突然受到限制2种情况下的控制效果。

英文摘要：

For solving the adapt control problem of biped hopping robot in unexpected situations,we build a CPG control mechanism which is able to change its control structure according to robot motion status simulating the CPG control mechanism in animals.Imi-tating the self reconfigurable feature of animal control mechanism,a time delayed response link is added to this control mechanism,for judging and adjusting the robot motion under unexpected situations.When the limbs of robot suddenly fails or the motion of robot limb is restricted,this control mechanism can sponta-neously adjust the structure of neural network accord-ing to perception of the robot motion status,and main-tain stable hopping motion.Finally,this control mecha-nism is studied on a simulation platform under two situations：normal hopping and hopping when one robotlimb suddenly fails.