Real-Time Path-Tracking Control for Commercial Trucks Based on Constraint-Handling Trajectory Prediction

Commercial truck path tracking is strongly affected by large mass and yaw moment of inertia, steering-rate constraints, and signal delays, whereas optimization-based predictive controllers can impose high online-computational costs. This study proposes a real-time path-tracking method based on Constraint-Handling Trajectory Prediction (CHTP). A dynamic model predicts the vehicle’s future pose, and a Stanley-based tracking law computes the desired steering angle from the predicted state. A constraint-handling module then explicitly limits the steering angle and its rate of change. The proposed CHTP method requires no online optimization. The method was evaluated through MATLAB/Simulink-TruckSim co-simulations and hardware-in-the-loop (HIL) tests. Under the low-speed unladen condition, CHTP reduced the maximum absolute-displacement error by 74.60% compared with the conventional Stanley controller. Under the high-speed unladen, low-speed heavy-load, and high-speed heavy-load conditions, CHTP completed the lane-change maneuver with bounded tracking errors, whereas the conventional Stanley controller failed to maintain convergent tracking. Across the four basic path-tracking conditions, the maximum absolute displacement and heading errors of CHTP did not exceed 0.3930 m and 0.1133 rad, respectively. Although nonlinear model predictive control (NMPC) generally achieved higher tracking accuracy, CHTP reduced the mean solution time by 94.17–95.96% relative to NMPC, with a maximum solution time of 1.1416 ms. Additional robustness tests showed that CHTP maintained bounded tracking errors and stable lateral dynamic responses under positioning errors, low road adhesion, and random response delays, while preserving its real-time computational performance. In the HIL test with a total loop delay of approximately 0.22 s, extending the prediction time from 0.20 s to 0.42 s limited the maximum displacement and heading errors to 0.2290 m and 0.1046 rad, respectively, with a maximum solution time of only 0.9895 ms. Additional prediction-time tests showed a trend consistent with the simulation results, further supporting the proposed delay-compensation mechanism. These results demonstrate that CHTP provides a favorable balance among tracking accuracy, robustness, and real-time performance.