祝贺LASIS实验室硕士生赵珂的论文《Lightweight pavement-adaptive TimesNet for road preview information correction in preview model predictive control of active suspensions》被机械领域TOP期刊《Mechanical Systems and Signal Processing》(MSSP) 录用。

Abstract: Accurate road-preview information is essential for active suspension systems to achieve effective and high-performance control. However, in real-world driving scenarios, sensor noise and environmental disturbances can cause severe perceptual errors, degrading controller performance. To address this issue, this paper proposes a pavement-adaptive TimesNet, termed P-TimesNet, for road-preview correction in preview model predictive control of active suspensions. The proposed network reconstructs road preview information by fusing biased preview inputs with vehicle dynamic responses. The reconstructed preview sequence is then incorporated into the prediction and optimization process of the model predictive control (MPC) controller, forming an integrated P-TimesNet-MPC framework for active suspension control under perceptual errors. Two typical perceptual error scenarios are constructed to validate the effectiveness of the proposed method: when the road grade perceived by the visual preview is lower than the actual grade, the root mean square values of vehicle body acceleration, suspension dynamic deflection, tire dynamic load, and actuator force are reduced by 4.30%, 6.36%, 2.36% and 2.48% compared with MPC without information correction; when the perceived road grade is higher than the actual grade, the corresponding reductions are 45.86%, 62.50%, 27.55% and 65.35%. These results demonstrate that the proposed P-TimesNet-MPC framework effectively improves active suspension performance under road-preview perceptual errors. In addition, experimental results on a quarter-car active suspension platform further verify the effectiveness of the proposed P-TimesNet-MPC under two typical perceptual error scenarios.

Keywords: Active suspension; Road preview error correction; Pavement-adaptive TimesNet;Model predictive control; Preview error correction; Time series prediction.