Trajectory Planning and Motion Control of Autonomous Electric Boats

For cities with ample access to rivers and canals it is possible to use them as an alternative to reduce the load on land transportation infrastructure such as bridges and roads. Traditional water-based transportation relies on the pilot to plan the trajectory and control the vessel, both of which depend on the experience of the pilot and are prone to human error. This thesis aims to address the generation of feasible trajectories and their tracking applied to autonomous electric boats for the transportation of people and goods on canals. The trajectory tracking is divided in two stages: path planning and velocity generation. The first stage generates a feasible path by means of a Dubins curve based RRT*, a sampling-based path planning algorithm that is able to take into account constraints on the curvature of the watercraft. The second stage generates a smoothed trapezoidal velocity profile considering bounds on surge velocity and acceleration of the vessel and assigns it to the reference path acting as the time law that defines the trajectory. Subsequently, this trajectory is given as a reference to a nonlinear model predictive controller (NMPC) local motion planner that solves an optimization problem for the best control input able to track the reference while avoiding collisions with walls and obstacles.