Abstract: This paper investigates modeling, on line parameter identification and adaptive control for the milling process. A second order discrete time transfer function is derived for the milling process. Based on the ordinary recursive least squares, a modified identification algorithm is suggested so as to eliminate the parameter distortion caused by the degeneration or expansion of the covariance matrix. The adaptive control law for the milling process is developed using pole assignment principle according to the desired objective transfer function of the closed loop system including the adaptive controller and the process. A PC/DSP based adaptive control system is built up to verify the algorithms of the adaptive controller. The modified recursive least squares and the pole assignment adaptive control law are plugged into a real time software platform, Intelligent Machining Modules(IMM), as plug in modules(PIMs) written in ANSI C codes. The simulation and experiment have demonstrated that the algorithms for identification and adaptive control are stable, reliable and robust. The milling process controlled by the adaptive controller behaves with the desired performance.

Key words: milling process, CNC, a daptive contr ol, on-line identification, modified RLS, pole assignment