Autonomous Docking of a Small Electric Ferry

Abstract

Docking is one of the most challenging and error-prone tasks in marine transportation. This project presents a controller that autonomously docks a small, dual-motor catamaran ferry using exclusively Global Navigation Satellite System (GNSS) and Inertial Measurement Unit (IMU) data. The controller consists of four subsystems: an extended Kalman filter for state estimation; a basic path planner; a Model Predictive Controller (MPC) assisted by a Linear Quadratic Regulator (LQR) for path following; and a thrust allocator that converts desired forces into angle and thrust commands for the motors. Running at 100 Hz, the LQR assists the 5 Hz MPC by attenuating high-frequency disturbances and model errors. Through system identification and physics modeling, a hydrodynamic maneuvering model is developed to predict the motion of the ferry. The model accounts for water damping, Coriolis and centripetal forces in a rotating body frame, rigid-body acceleration forces, and added mass effects. All development, system identification experiments, and performance evaluation tests are carried out in the Virtual RobotX (VRX) simulator. The system is evaluated under zero, medium (4 m/s wind, 0.5m waves), high (7 m/s wind, 1.5m waves), and high-wind (11 m/s wind, no waves) disturbances, with two runs each, to compare the combined MPC-LQR controller with a standalone MPC. Both architectures successfully dock the ferry under zero and medium disturbances. Under high disturbances, both succeed to dock in one of the runs, while both fail in the other. Under high-wind, the MPC-LQR controller outperforms the standalone MPC by successfully docking in both runs while the standalone MPC only succeeds in one. While both controllers reach the correct position across all disturbance levels, tracking heading remains the primary challenge; both respond slowly to heading changes, though the MPC-LQR provides slightly better tracking. These results demonstrate the feasibility of autonomous docking for small catamaran ferries using an MPC-LQR controller and exclusively onboard sensing under environmental disturbances.