Scholarlink

<p>Safe trajectory tracking under a steering actuator failure remains a critical challenge for intelligent vehicles, particularly for four-wheel independently driven (4WID) intelligent vehicles. Existing model predictive control (MPC)-based fault-tolerant control (FTC) schemes typically assume known fault characteristics, depend on observers with uncertain convergence, or disregard feasibility loss that arises when high-curvature reference trajectories exceed the residual maneuvering capability following complete steering failure. This paper proposes a capability-aware trajectory-tracking framework that integrates a fault-tolerant feasible region (FTFR) into MPC and performs online reshape the reference targets (RRT) while exploiting differential wheel torques for steering control. The FTFR is hierarchically constructed from tire-, vehicle-, and road-level constraints to bound the achievable longitudinal and lateral forces as well as the yaw moment. The RRT module adjusts the vehicle speed and smooths the reference curvature to keep the target trajectory within the FTFR. The MPC optimizes the increments of the control inputs subject to the FTFR and actuator constraints, while the four-wheel torque allocation is formulated as a quadratic programming (QP) problem. The effectiveness of the proposed approach is demonstrated through Simulink–CarSim co-simulations of a high-speed overtaking maneuver at 80&#xa0;km/h on a road surface with friction coefficient μ = 0.8, under a complete steering cut-off fault. Compared with conventional differential-steering FTC schemes, the proposed FTFR + RRT controller maintains accurate path following after fault onset, with smaller lateral tracking errors and minimal overshoot near curvature peaks, resulting in a reduction of the mean trajectory deviation from 0.2789 to 0.1521&#xa0;m, and a decrease of the standard deviation of speed fluctuations by 27.9%. Furthermore, it preserves yaw-rate continuity with reduced post-fault oscillations, maintains the vehicle speed close to the desired value with smaller fluctuations, and prevents tire-torque saturation while retaining comfortable safety margins. By unifying feasible-region constraints with online reference reshaping within the MPC framework, the proposed method provides a practical safety envelope that preserves both trajectory-tracking performance and control feasibility for 4WID vehicles, even under complete steering loss.</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

Trajectory Tracking Control for 4-WID Vehicles Under Steering Failure Based on a Fault-Tolerant Feasible Region and Reference Trajectory Reshaping

  • 期刊论文

  • 发表日期：

    2026年3月23日

查看全文
  • Yiheng Shi,
  • Qiangqiang Yao,
  • Zhendong Zhu

摘要

Safe trajectory tracking under a steering actuator failure remains a critical challenge for intelligent vehicles, particularly for four-wheel independently driven (4WID) intelligent vehicles. Existing model predictive control (MPC)-based fault-tolerant control (FTC) schemes typically assume known fault characteristics, depend on observers with uncertain convergence, or disregard feasibility loss that arises when high-curvature reference trajectories exceed the residual maneuvering capability following complete steering failure. This paper proposes a capability-aware trajectory-tracking framework that integrates a fault-tolerant feasible region (FTFR) into MPC and performs online reshape the reference targets (RRT) while exploiting differential wheel torques for steering control. The FTFR is hierarchically constructed from tire-, vehicle-, and road-level constraints to bound the achievable longitudinal and lateral forces as well as the yaw moment. The RRT module adjusts the vehicle speed and smooths the reference curvature to keep the target trajectory within the FTFR. The MPC optimizes the increments of the control inputs subject to the FTFR and actuator constraints, while the four-wheel torque allocation is formulated as a quadratic programming (QP) problem. The effectiveness of the proposed approach is demonstrated through Simulink–CarSim co-simulations of a high-speed overtaking maneuver at 80 km/h on a road surface with friction coefficient μ = 0.8, under a complete steering cut-off fault. Compared with conventional differential-steering FTC schemes, the proposed FTFR + RRT controller maintains accurate path following after fault onset, with smaller lateral tracking errors and minimal overshoot near curvature peaks, resulting in a reduction of the mean trajectory deviation from 0.2789 to 0.1521 m, and a decrease of the standard deviation of speed fluctuations by 27.9%. Furthermore, it preserves yaw-rate continuity with reduced post-fault oscillations, maintains the vehicle speed close to the desired value with smaller fluctuations, and prevents tire-torque saturation while retaining comfortable safety margins. By unifying feasible-region constraints with online reference reshaping within the MPC framework, the proposed method provides a practical safety envelope that preserves both trajectory-tracking performance and control feasibility for 4WID vehicles, even under complete steering loss.