遥控作业器基于主动时延神经网络的感知和控制

遥控作业器是实现未知或危险环境中作业的有力手段．由于遥控机械手大多工作在不确定的环境中，操作者难以预先知道环境目标的动力学特性，在机械手和环境发生力的交互过程中，需要考虑机械手和环境碰撞时的强非线性对控制回路的影响，因此对机械手和环境碰撞的实时感知和识别是非常重要的．本文利用接近觉传感器测量机械手顶端与环境物体的距离，提出了基于主动时延神经网络的遥控作业器的感知和控制方法．仿真实验表明了该方法的有效性和对接近觉传感器量程的要求

SENSING AND CONTROL OF TELEOPERATOR BASED ON ACTIVE TIME DELAY NEURAL NETWORK

Teleoperator plays an important role in operation in unknown or hazardous environment.Because the environment is unknown or partly unknown,the human operator is not able to recongnize the dynamics of the environment before he controls the slave manipulator to interact with the environment.When there are interactions between slave manipulator and environment,it is necessary to consider the nonlinearity of collision in the control loop,therefore,the sensing and recongnization of collision between slave manipulator and environment on time is important.In this paper,proximity sensor is used to measure the distance between slave manipulator and environment,and a controller based on active time delay neural network is proposed to control the nonlinear contact,the effectiveness of which is showed by the result of simulation experiment.