To address the challenges to driving stability caused by large-curvature steering of wire-controlled mining vehicles in narrow tunnels, a fused four-wheel steering (4WS) control strategy based on real-time estimation of vehicle state parameters is proposed. A comprehensive longitudinal–lateral–yaw dynamics model for 4WS is established, and a comparative study is conducted on three control methods: proportional feedforward control, yaw rate feedback control, and fused control. Expressions for steady-state yaw rate gain under different control modes are derived, and the stability differences in 4WS characteristics among these strategies are thoroughly analyzed. To overcome the difficulty in directly acquiring state information for chassis steering control, a vehicle state parameter estimator based on the unscented Kalman filter (UKF) is designed. To enhance the robustness to noise and computational real-time performance of vehicle state estimation in complex environments, a method for real-time estimation of noise covariance matrices using innovative sequences is adopted, improving the estimation accuracy of the algorithm. To validate the effectiveness of the control strategies, a co-simulation platform integrating Carsim and Matlab/Simulink is developed to simulate the performance of the three 4WS control methods under step steering and sinusoidal steering input conditions. The results show that, under low-speed conditions, 4WS strategies increase the yaw rate by approximately 50% and reduce the turning radius by over 45%, significantly enhancing steering maneuverability. Under medium-high speed conditions, 4WS strategies decrease the yaw rate by up to 68% and increase the turning radius by 17–29%, effectively suppressing oversteering tendencies to comprehensively improve stability, with the integrated control strategy demonstrating the best performance. Under both test conditions, the fused feedforward and feedback control strategy reduces the steady-state yaw rate by approximately 12.7% and 48.7%, respectively, compared to other control strategies, demonstrating superior stability.

In response to the problem that single sensor data is easily affected and interfered with by the environment, making it difficult to collect accurate gearbox fault signals, resulting in low fault recognition accuracy, this paper proposes a multi-sensor self-learning weighted fusion method. This method not only effectively improves the accuracy of fault diagnosis but also has good interpretability. Firstly, the signals of different sensors are transformed into time–frequency images through sliding window and continuous wavelet transform to improve the signal to noise ratio; In order to fully extract fault features from the time–frequency images, a feature extraction module is constructed using model transfer method for feature extraction; Then, the self-learning weighted fusion module amplifies the fault features in the feature matrix while suppressing interference features to improve the accuracy of gearbox fault diagnosis; Finally, the classification is carried out through the classification module. The average accuracy of this method under different operating conditions on the gearbox dataset of Southeast University and the WT planetary gearbox dataset of Beijing University of Technology is higher than 99.9%.