Hailong Mu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hailong Mu has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Lai Peng
China University of Mining and Technology
2 shared publications - Xinhua Liu
School of Mechatronic Engineering, China University of Mining and Technology, Xuzhou, China
2 shared publications - Dezheng Hua
China University of Mining and Technology
1 shared publication - Jun Wu
University of Oxford 1
1 shared publication - Shuo Xin
China University of Mining and Technology
1 shared publication - Ting Zhang
School of Mechanical and Electrical Engineering Soochow University Suzhou China
1 shared publication
Latest Publications
- 3 June 2026
- Published by IOP Publishing in Smart Materials and Structures
- Vol. 35 (6)
Research on LQR control strategy of magnetorheological semi-active suspension based on improved sand cat swarm optimization algorithm
To address the problems of nonlinear damper dynamics, actuator saturation, and empirical parameter tuning in magnetorheological (MR) semi-active suspension control, this paper proposes an LQR-clipped control strategy optimized by Improved Sand Cat Swarm Optimization (ISOSO) algorithm. A quarter-car two-degree-of-freedom semi-active suspension model was first established, and the structural characteristics and dynamic properties of the MR damper were investigated through vibration tests on an MTS810 electro-hydraulic servo fatigue testing system. Based on the measured time, displacement, velocity, current, and damping force data, a feedforward neural network (FFNN) was developed to model the nonlinear dynamics of the MR damper and predict its damping force accurately. On this basis, an LQR-clipped controller was designed to map the ideal LQR control force into the realizable damping force range of the MR damper, thereby satisfying semi-active actuation constraints. Meanwhile, ISOSO integrating Tent chaotic mapping, quasi-opposition-based learning, elite pool guidance, sine-cosine adjustment, and Cauchy mutation was employed to optimize the LQR weighting matrices. Simulation results under Class C random road and speed-bump excitations demonstrate that the proposed ISOSO-LQR strategy outperforms passive suspension and SCSO-LQR in vibration reduction and suspension travel control. Road tests under random road and speed-bump conditions further verify that the proposed strategy can effectively suppress vehicle body vibration while maintaining reasonable suspension working space, providing an effective solution for constrained optimal control of MR semi-active suspension systems.
- 3 June 2026
- Published by IOP Publishing in Smart Materials and Structures
- Vol. 35 (6)
Research on LQR control strategy of magnetorheological semi-active suspension based on improved sand cat swarm optimization algorithm
To address the problems of nonlinear damper dynamics, actuator saturation, and empirical parameter tuning in magnetorheological (MR) semi-active suspension control, this paper proposes an LQR-clipped control strategy optimized by Improved Sand Cat Swarm Optimization (ISOSO) algorithm. A quarter-car two-degree-of-freedom semi-active suspension model was first established, and the structural characteristics and dynamic properties of the MR damper were investigated through vibration tests on an MTS810 electro-hydraulic servo fatigue testing system. Based on the measured time, displacement, velocity, current, and damping force data, a feedforward neural network (FFNN) was developed to model the nonlinear dynamics of the MR damper and predict its damping force accurately. On this basis, an LQR-clipped controller was designed to map the ideal LQR control force into the realizable damping force range of the MR damper, thereby satisfying semi-active actuation constraints. Meanwhile, ISOSO integrating Tent chaotic mapping, quasi-opposition-based learning, elite pool guidance, sine-cosine adjustment, and Cauchy mutation was employed to optimize the LQR weighting matrices. Simulation results under Class C random road and speed-bump excitations demonstrate that the proposed ISOSO-LQR strategy outperforms passive suspension and SCSO-LQR in vibration reduction and suspension travel control. Road tests under random road and speed-bump conditions further verify that the proposed strategy can effectively suppress vehicle body vibration while maintaining reasonable suspension working space, providing an effective solution for constrained optimal control of MR semi-active suspension systems.
- 1 June 2026
- Published by IOP Publishing in Smart Materials and Structures
- Vol. 35 (6) , 065048
A novel dynamic modeling method for magnetorheological damper based on physics-informed residual learning
Magnetorheological dampers (MRD) exhibit strongly nonlinear hysteretic behavior, which poses significant challenges for accurate dynamic modeling and limits the full exploitation of their controllable performance. To address this issue, a physics-informed residual learning framework is proposed for modeling the nonlinear dynamics of MRD. A physics-informed B-spline curve hysteresis (PIBSH) model is first developed to establish a morphology-based white-box parametric formulation, enabling a physically interpretable phenomenological representation of the damping force. The proposed model decomposes the damping force into linear components and a nonlinear hysteretic component described by B-spline functions. By leveraging B-spline basis functions and control points, the model effectively captures the mapping between motion states and local hysteretic characteristics. To further enhance predictive accuracy, a Kolmogorov–Arnold transformer (KAT) is constructed, integrating nonlinear feature extraction capability of the Kolmogorov–Arnold network with the attention-based temporal modeling advantage of the transformer. Based on this structure, a hybrid model termed PIBSH–KAT (PIKAT) is established, where KAT learns and compensates for the residuals between the PIBSH predictions and experimental measurements. Comparative experiments demonstrate that the PIBSH model achieves superior accuracy in representing local hysteretic behavior, while the residual-enhanced PIKAT model further improves prediction precision and generalization performance, achieving a maximum accuracy of 0.9956 under multiple working conditions.
- 1 June 2026
- Published by IOP Publishing in Smart Materials and Structures
- Vol. 35 (6) , 065048
A novel dynamic modeling method for magnetorheological damper based on physics-informed residual learning
Magnetorheological dampers (MRD) exhibit strongly nonlinear hysteretic behavior, which poses significant challenges for accurate dynamic modeling and limits the full exploitation of their controllable performance. To address this issue, a physics-informed residual learning framework is proposed for modeling the nonlinear dynamics of MRD. A physics-informed B-spline curve hysteresis (PIBSH) model is first developed to establish a morphology-based white-box parametric formulation, enabling a physically interpretable phenomenological representation of the damping force. The proposed model decomposes the damping force into linear components and a nonlinear hysteretic component described by B-spline functions. By leveraging B-spline basis functions and control points, the model effectively captures the mapping between motion states and local hysteretic characteristics. To further enhance predictive accuracy, a Kolmogorov–Arnold transformer (KAT) is constructed, integrating nonlinear feature extraction capability of the Kolmogorov–Arnold network with the attention-based temporal modeling advantage of the transformer. Based on this structure, a hybrid model termed PIBSH–KAT (PIKAT) is established, where KAT learns and compensates for the residuals between the PIBSH predictions and experimental measurements. Comparative experiments demonstrate that the PIBSH model achieves superior accuracy in representing local hysteretic behavior, while the residual-enhanced PIKAT model further improves prediction precision and generalization performance, achieving a maximum accuracy of 0.9956 under multiple working conditions.