Multi-objective KCF-improved CSO for permanent magnet motor rotor detection and drive

Traditional approaches to detecting permanent magnet motor rotors predominantly center on sensor analysis and the refinement of control strategies. However, these methods inevitably encounter challenges, including elevated costs and vulnerability to interference. Consequently, to enhance the detection accuracy of Permanent Magnet (PM) motor rotors, this study introduces an innovative "combined visual-control optimization" approach, which integrates a Multi-objective kernel correlation filtering (MOKCF)tracking algorithm with an improved Cat Swarm optimization (CSO) algorithm. Firstly, it proposes monocular visual tracking of the rotor attitude and realizes real-time tracking of the target feature points of the rotor by using the Kalman filter (KF) to correct the tracking algorithm. At the same time, the CSO improved by quantum chaos and mapping relationship is used to realize electric drive. The results show that the improved multi-objective algorithm can effectively track the target, and the root mean square and mean absolute values of the lateral deviation of the deflection angle and the spin angle are not more than 0.042° and 0.029° respectively. The maximum average convergence time of the improved cat swarm algorithm does not exceed 0.1s, and the torque and speed of the motor control system are more stable, and the motion error of the rotor feature points is smaller. The method proposed by this study has significant advantages in improving detection accuracy and optimizing driving performance, providing a new technical reference for intelligent detection and fault diagnosis of motor rotors.