Abstract: A force-position decoupling penumatic servo control system for polishing force is set up to solve the problem of force-position coupling in the traditional polishing process. The real-time control results of penumatic servo control system for polishing force are investigated by adopting traditional PID control, feedward PID control and single neuron PID control strategies with data acquisition card and RTW toolbox in Matlab/simulink. The parameters in feedward PID control can be identified by a recursive least squares algorithm. The weighting coefficients in single neuron PID control can be modified by the supervision Hebb learning regulations. The real-time control experimental results of polishing force indicate that the better control can be achieved by feedward PID control and single neuron PID control strategies than by traditional PID control. The duration for stablizing the polishing force is short by feedward PID control and the oscillation is little during the stablizing process of polishing force by single neuron PID control.

Key words: Polishing force control, Force-position decoupling, Feedward PID, Single neuron PID, On-line identification