基于改进型神经网络的焊接机器人焊枪控制算法研究

Control algorithm for robot welding torch based on improved neural network

焊接是一个复杂、参数多变、时变非线性的过程,对于焊缝跟踪控制问题,传统的PID控制方法很难有效的对焊缝进行跟踪控制。在传统PID控制中引入神经网络,提出了一种双权值神经网络控制器的焊缝跟踪控制算法。通过控制伺服电机驱动的十字滑块运动,从而实现焊枪的精确定位。通过MATLAB仿真试验验证了双权值神经网络控制器的焊缝跟踪控制算法能够较好的实现焊缝的跟踪控制。

Welding is a complicated time-varying nonlinear process with variable parameters. The traditional PID control is very difficult to track and control the weld seam. A new double weights neural network controller for seam tracking control algorithm was proposed by combining the traditional PID control into neural network. Through controlling the cross slider movement driven by servo motor,the accurate positioning of the welding torch was achieved. Simulation by MATLAB indicates that the developed double weights neural network controller can be used to track and control the weld seam.