基于气动柔性驱动器,设计了气动柔性五自由度手指,并介绍了其结构原理。分析建立了五自由度手指的运动学方程,并采用遗传算法解决了该五自由度手指逆运动学中的冗余解问题。基于神经网络设计了手指控制器,并进行了手指指尖轨迹跟踪实验。实验结果表明,手指指尖的运动实验曲线与理论曲线比较接近,说明采用遗传算法及神经网络控制能够实现对五自由度手指的较精确实时控制。
Based on flexible pneumatic actuator, a new kind of flexible pneumatic 5--DOF finger was proposed. The structure and operating principle of the 5- DOF finger was presented. The kinematics equation of the 5- DOF finger was built and the redundancy problem of the inverse kinematics solution was resolved by genetic algorithm. The BP neural network controller of the finger was designed and the experiments of the fingertip trajectory tracking were carried out. Experimental results show that the experimental curve matches well with the theoretical curve, and genetic algorithm and neural network controller can control the 5--DOF finger exactly in real-time.