Abstract: A force-position decoupling pneumatic servo control system for polishing force was set up to solve force-position coupling problem in traditional polishing process. The real-time control results of pneumatic control system for polishing force were researched by adopting traditional PID control， feed-forward PID control and single neuron PID control， using data acquisition card and RTW real-time toolbox in the Matlab/simulink. The parameters in feed-forward PID control can be identified by online recursive least squares algorithm. The weighting coefficients in single neuron PID control can be modified by the supervised Hebb learning rules. The real-time control experimental results of polishing force indicate that the better control effect can be realized by feed-forward PID control and single neuron PID control than by PID control. The duration for stabilizing process of the polishing force is short by feed-forward PID control and the oscillation is little during the stabilizing process of the polishing force by single neuron PID control.

Key words: polishing force control, force-position decoupling, feed-forward PID, single neuron PID, on-line identification