Scholarlink

<p>To address the high-precision and robust steering angle tracking requirements of steer-by-wire systems, this paper proposes a ‘fuzzy-radial basis-particle swarm sliding mode control’ strategy for the first time. Its core innovation lies in designing a radial basis neural network with only one adaptive weight to estimate the unknown dynamics and disturbances of the system in real time using a minimal parameter learning method, and constructing a dual-loop adaptive sliding mode architecture. The inner loop uses fuzzy logic to adjust the switching gain online to suppress chattering, while the outer loop employs particle swarm optimisation to update the sliding surface coefficients in real time, achieving fast response and energy optimisation. Additionally, based on Lyapunov theory, a rigorous stability proof is provided for the dual adaptive loop, ensuring global consistent bounded convergence. CarSim-Matlab joint simulation shows that under three typical operating conditions (Sine reference signal, Single lane change, Double lane change), the ‘fuzzy-radial basis-particle swarm sliding mode control’ strategy reduces the maximum angular tracking error by 23.6–92.5% and the average absolute angular tracking error by 17.6–93.9% compared to traditional sliding mode control and fault-tolerant sliding mode predictive control, the maximum displacement tracking error is reduced by 41.1–90.7%, and the average absolute displacement tracking error is reduced by 30.9–95.1%. Additionally, the algorithm significantly reduces vibration, validating its practical value.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Intelligent adaptive control for precision steering in steer-by-wire vehicles

  • 期刊论文

  • 发表日期：

    2026年2月11日

查看全文
  • Yang Sun,
  • Sen Liu,
  • Yinuo Ma,
  • Junxing Zhang,
  • Xuhuai Liu,
  • Wenqin Duan,
  • Zhicheng Zhang

摘要

To address the high-precision and robust steering angle tracking requirements of steer-by-wire systems, this paper proposes a ‘fuzzy-radial basis-particle swarm sliding mode control’ strategy for the first time. Its core innovation lies in designing a radial basis neural network with only one adaptive weight to estimate the unknown dynamics and disturbances of the system in real time using a minimal parameter learning method, and constructing a dual-loop adaptive sliding mode architecture. The inner loop uses fuzzy logic to adjust the switching gain online to suppress chattering, while the outer loop employs particle swarm optimisation to update the sliding surface coefficients in real time, achieving fast response and energy optimisation. Additionally, based on Lyapunov theory, a rigorous stability proof is provided for the dual adaptive loop, ensuring global consistent bounded convergence. CarSim-Matlab joint simulation shows that under three typical operating conditions (Sine reference signal, Single lane change, Double lane change), the ‘fuzzy-radial basis-particle swarm sliding mode control’ strategy reduces the maximum angular tracking error by 23.6–92.5% and the average absolute angular tracking error by 17.6–93.9% compared to traditional sliding mode control and fault-tolerant sliding mode predictive control, the maximum displacement tracking error is reduced by 41.1–90.7%, and the average absolute displacement tracking error is reduced by 30.9–95.1%. Additionally, the algorithm significantly reduces vibration, validating its practical value.