Abstract
A pressure servosystem, with compressed air as its primary power source, plays a pivotal role in automotive braking. The substitution of expensive proportional valves with high-speed switching valves (HSVs) for chamber pressure control remains a prominent challenge for researchers. In addressing the single-chamber dual-valve pressure tracking system, a novel approach is proposed using an adaptive neuro-fuzzy inference system (ANFIS) that enhances fuzzy control through neural network refinement. Integration with mode switching is employed to ameliorate chamber pressure tracking performance. This strategy amalgamates the learning capability of neural networks with the inferential capacity of fuzzy logic, effectively handling the intricate nonlinear characteristics of pneumatic systems. Experimental results demonstrate that for step signals in the range of 0.3–0.6 MPa, the maximum overshoot is reduced to 0.0041 MPa, and the random step error ranges between −0.01287 and 0.01275 MPa. The relative root-mean-square error for a 0.5 Hz harmonic signal is diminished by 26.91%.