Adaptive PID path tracking control for three-wheeled differential drive unmanned vehicles

Small differential-drive unmanned vehicles are vital carriers for intelligent warehousing and logistics. Path tracking control directly governs the autonomous driving accuracy and operational reliability of vehicles, possessing prominent engineering application value. Conventional fixed-parameter PID path tracking controllers hold invariable gains, which fail to cope with complex operating conditions including speed fluctuations and abrupt curvature changes. Such controllers suffer from slow dynamic response at low speeds and severe overshoot and oscillation at high speeds, restricting the further improvement of tracking performance for differential-drive unmanned vehicles. To address the poor adaptability to varying working conditions, this paper establishes a double closed-loop adaptive PID path tracking control system for three-wheeled differential-drive unmanned vehicles. The inner loop adopts incremental PID to stabilize the speeds of left and right driving wheels. For the outer loop, an adaptive PID controller is developed to online tune control gains based on real-time lateral path deviation and vehicle velocity, eliminating the weak adaptability of traditional fixed-parameter PID. Physical vehicle experiments demonstrate that at a target speed of 5 m/s, the steady-state lateral error on straight paths is within ±2.1 cm, and the maximum lateral error on 120° obtuse broken-line curves is limited to ±5.0 cm. Compared with the traditional fixed-parameter PID algorithm, the proposed method reduces overshoot by 40.7% and thoroughly suppresses straight-path oscillation, remarkably enhancing the path tracking accuracy and working-condition robustness of unmanned vehicles.