基于FPA的三关节气动手指的神经网络控制

Neural Control of Three-joint Pneumatic Finger Based on FPA

介绍了一种基于FPA的气动柔性手指,该手指具有体积轻、能直接驱动和柔性好等特点。根据该手指的运动学模型,提出了神经网络控制方法,建立位置控制模块网络(PCNET)和驱动器控制模块网络(ACNET),通过ACNET控制伺服阀输入电压,解决系统延迟问题,通过PCNET解决手指的位置控制。仿真结果表明,该控制算法能很好地实现手指控制的精确度和快速性。

A new pneumatic finger based on FPA was presented, the finger is flexible and can be driven directly. Based on the dynamic model of the finger, a neural network controller was designed, position control net （PCNET） and actuator control net （ACNET） were established. ACNET controls the input voltage of the sever valve and resolves the delay of the system. PCNET controls the position of the finger. The simulating result shows the neural net controller has excellent precision, it is proved a new way to control the flexible pneumatic finger.