基于神经-肌肉架构的仿生并联躯干柔顺控制

模仿动物躯干的构造搭建了六自由度仿生并联躯干。提出一种神经肌肉控制方法，通过控制虚拟肌肉的伸缩使仿生并联躯干在外力冲击下进行柔顺运动。建立了神经运动控制网络，根据非线性振荡器生成的节律信号规划了仿生并联躯干的运动轨迹。运用逆运动学求解了关节转角，按照转角信息控制电机带动仿生并联躯干运动。构建的虚拟肌肉模型在外力作用下驱动躯干关节产生顺应运动。仿真模型和并联躯干实体测试结果表明该方法可有效提高并联躯干的柔顺缓冲性和快速响应性。

Neuromuscular Architecture Based Compliance Control of Bionic Parallel Torsos

 A 6-DOF bionic parallel torso was constructed by mimicking the structure of animal torsos. A neuromuscular control method was proposed to make the bionic parallel torso move flexibly under the impacts of external forces by controlling the stretching of virtual muscles. A neuromuscular control network was established to plan the motion trajectory of the bionic parallel torsos based on rhythm signals generated by nonlinear oscillators. The inverse kinematics was applied to solve the joint rotation angles， and the motor was controlled to drive the bionic parallel torso movement according to the rotation angle information. A virtual muscle model was constructed to drive the torso joints to produce compliant motion under the action of external forces. The simulation model and the physical test results of the parallel torsos show that the method may effectively improve the flexibility of cushioning and fast response of the parallel torsos.