A*-based trajectory planning in dynamic environments for autonomous vehicles

  • A*-basierte Trajektorienplanung in dynamischen Umgebungen für autonome Fahrzeuge

Nantabut, Chinnawut; Abel, Dirk (Thesis advisor); Kowalewski, Stefan (Thesis advisor)

Aachen : RWTH Aachen University (2023)
Dissertation / PhD Thesis

Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2023

Abstract

A collision-free trajectory of autonomous cars is of utmost importance. The method for calculating this has to be developed in order to deal with the complex environments and the time available in the guidance level of automated driving. This is normally fixed in 1 second. An efficient trajectory algorithm that is based on A* has shown its potential since the beginning of the era of autonomous cars and is still applicable and developed recently. Since it does not consider the nature of the vehicle dynamic, a lot of methods based on this algorithm have been introduced. Furthermore, environmental modeling plays a major role as well since the autonomous car has to understand the scenario as best as possible. In this dissertation, a series of trajectory-planning algorithms based on A* are introduced in dependence on the types of environments filled with static and dynamic obstacles. A perception of the environment is also modeled and understood. The interaction of all traffic participants is simulated for a more naturalistic environmental model. For all cases, a typical trajectory tracking algorithm such as Stanley controller is utilized. Furthermore, one of the trajectory planning algorithms based on A* is also implemented and evaluated on a real model vehicle.

Institutions

  • Chair and Institute of Automatic Control [416610]

Identifier

Downloads