基于变参数的远程康复训练机器人神经网络控制

为使患者的远程康复训练达到更为理想的效果,针对患者在训练时肌肉痉挛对双向遥操作系统的稳定性以及对从机械手的速度平滑性的影响,提出了一种新的基于BP神经网络变参数控制方法.通过检测患者训练时力、加速度、速度、位置等的变化,采用BP神经网络自动调整从端控制参数,从而消除了系统的不稳定性和减少了对系统平滑性的影响,并具有很强的鲁棒性.分析和仿真试验结果表明,此方法与传统的控制方法相比,可有效地克服患者因肌肉痉挛带来的干扰并具有较好的稳定性和平滑性.

Neural network control of tele-rehabilitation robot based on variable parameter

During the tele-rehabilitation traing,the rehabilitant's muscle spasm may result in the bilateral telerobot system instability and cause its slave's speed unsmoothness,which make the rehabilitation training inefficient.In order to guarantee the stability and reduce the fluctuation of the speed,a new method based on the variable parameter with BP(back propagation) neural network was brought forward.By measuring the changes of the limb's force,acceleration,velocity and position,BP neural network automatically adjusted the control parameter of the slave side.Not only the stability can be guaranteed,but also the slave's speed unsmoothness can be decreased.The system has strong robustness.Analysis and simulation results show that this method is much more stable and smooth than the traditional control and can effectively restrain the interference resulting from rehabilitant's spasm.