数据驱动动力学与控制研究若干进展

Dynamics and control is a discipline that studies the dynamic mechanisms of systems and their control strategies, and plays an important role in modern engineering and scientific research. The complexity caused by geometric nonlinearity, the non-smoothness of contact forces, and the uncertainty of environmental interferences and multi-physics problems poses significant challenges to dynamic modeling, prediction and intelligent control. The rapid development of data-driven methods has provided new ideas and new research paradigms for addressing these challenges. Recent researches have shown that data-driven methods can not only solve some problems that traditional dynamics methods cannot address but also significantly enhance the ability to predict dynamical behavior and design advanced structures. These methods lay the foundation for intelligent research in dynamics and control and demonstrate great potential and scientific value in the modeling, analysis, and regulation and control of complex systems. This paper briefly reviews the research progress of data-driven methods in areas such as robot motion control, transonic aeroelastic modeling and analysis, dynamics design, stochastic dynamics, neurodynamics, fault diagnosis and remaining useful life prediction of machinery. It also discusses the challenges and trends in these fields.