Study on Electric Sunroof Pinch Detection of Cars Based on Model Reference Adaptive Cholesky Decomposition Filter

This paper proposes a cars’ electric sunroof pinch detection algorithm based on a model reference adaptive Cholesky decomposition filter (MRACDF). Because of the inaccuracy of noise variance and the roundoff errors caused by the propagation of traditional filters, the pinch detection for electric sunroofs is susceptible to errors and serious accidents. To avoid accidents on passengers, especially pinched children, it is necessary to improve the accuracy and reliability of the pinch detection. To meet this necessity, this paper first analyzes the characteristics of the electric sunroof control system and establishes the system model. Second, it proposes a novel model reference method to calculate plant noise, and proposes a sliding–fading grey prediction algorithm to estimate observation noise. The robust Cholesky decomposition theory is used to overcome the defects of a traditional filter. Third, a robust pinch detection algorithm based on MRACDF is designed for cars’ electric sunroofs. Finally, a new state variable for pinch detection is proposed and compared with traditional methods in detail. To verify the electric sunroof pinch detection algorithm, experiments are conducted on the electric sunroof platform of laboratory cars. Experimental results show that the pinch detection algorithm has excellent performances in different conditions.