Master thesis presentation

​Oskar Larsson: The Oskillator Artificial Force Field Highway Chauffeur. An Algorithm for Autonomous Driving on Highways

Examiner: Alexei Heintz

Abstract: The subject of this thesis has been to develop a module for combined behavior and trajectory planning based on artificial force fields for autonomous driving on highways. Given a desired velocity and headway time as driver input, a range of acceptable lanes as route input, and enough information about the host, the road and other road-users, it determines an appropriate acceleration vector in the world coordinate system for each point in time, to drive the car safely and comfortably on highways.

The algorithm consists of calculating a longitudinal acceleration, \textit{along the road}, by considering the desired velocity, properties of the vehicles ahead, such as distance, velocity and acceleration, and the curvature of the road ahead. The lateral acceleration, \textit{across the road}, is then calculated by considering the road curvature, the acceptable lanes and the other cars on the road. The thesis has been focused on finding a set of \textit{force components} to achieve the different behaviors, e.g. the Lane Component to center in lane, the Trail Component to slow down to a vehicle in front, etc., and a way of composing the components, with an analytically determined damping, to achieve an acceleration that is feasible, safe and comfortable. The output acceleration is continuous with respect to all inputs.

Category Student project presentation
Location: MV:L15, Chalmers tvärgata 3
Starts: 24 February, 2020, 10:00
Ends: 24 February, 2020, 11:00

Published: Mon 17 Feb 2020.