Optimization of the Parking Manoeuvre for a 1-Trailer Truck

Abstract: Motion planning for autonomous heavy duty vehicles has been intensively studied during the last years due to its promising benefits for industry in terms of money savings and safety. The system treated in this thesis is a 1-trailer truck. The goal was to develop and evaluate different methodologies in order to optimize the reverse parking manoeuvre for this specific system, while considering its kinematic constraints and collision avoidance. So far most of the methodologies developed for the system at hand focus on finding a feasible path connecting a start configuration to a goal configuration, but very little attention has been paid to the optimality of the solution given. Along the report the author assumes that a feasible solution is available and the four methodologies are used to optimize the initial path. The algorithms were tested in different environments in order to assess their strengths and weaknesses. Results show that a good description of the obstacles in the map is required to obtain a good solution in all class of environments, and if that description is scarce the optimization for narrow and spatially constrained environments does not provide the desired results. Furthermore, it was discovered that using a good feasible solution as a starting point reduces the convergence time of the algorithms presented.