Signal processing and force control for precision machining with grinding system integrating an aerostatic suspension compact pneumatic actuator

Signal processing is the core component of precision machining, where reliable contact force signals underpin monitoring, control, and quality assurance. Pneumatic actuators are highly favored in grinding applications due to their compliance, but friction and stick–slip phenomena distort force signals and degrade closed-loop performance. This study developed an aerostatic suspension compact pneumatic actuator (ASCPA) that suppresses friction at its source and compares it with traditional pneumatic actuators (TPA) and commercial low-friction pneumatic actuators (LFPA). Through piston-level fluid dynamics simulation, a balance between radial load capacity and air consumption was achieved, enabling prototype manufacturing and friction validation testing. From a manufacturing-oriented signal perspective, the adaptive state-space force signal recognition method EM-KF-RTS was proposed. This method integrates a Kalman filter with a Rauch-Tung-Striebel smoother and employs an expectation maximization algorithm for online parameter adaptation. Using only standard force values and air supply pressure measurements from the test bench, this estimator significantly enhances resistance to vibration, electromagnetic interference, and sensor noise, enabling precise force reconstruction. The actuator-algorithm stack was evaluated under constant, sinusoidal, and variable-amplitude frequency-modulated conditions, as well as during sudden external contact and inter-chamber pressure changes on the dynamic grinding test bench. Results demonstrate that the EM-KF-RTS method achieves stable force recognition under constant, sinusoidal, and variable amplitude-frequency reference signals, as well as during sudden external contact and inter-chamber pressure changes. It exhibits superior dynamic response performance, lower root mean square error, and stronger interference suppression capabilities.