Tactical decision-making for autonomous driving: A reinforcement learning approach
Licentiate thesis, 2019

The tactical decision-making task of an autonomous vehicle is challenging, due to the diversity of the environments the vehicle operates in, the uncertainty in the sensor information, and the complex interaction with other road users. This thesis introduces and compares three general approaches, based on reinforcement learning, to creating a tactical decision-making agent. The first method uses a genetic algorithm to automatically generate a rule based decision-making agent, whereas the second method is based on a Deep Q-Network agent. The third method combines the concepts of planning and learning, in the form of Monte Carlo tree search and deep reinforcement learning. The three approaches are applied to several highway driving cases in a simulated environment and outperform a commonly used baseline model by taking decisions that allow the vehicle to navigate 5% to 10% faster through dense traffic. However, the main advantage of the methods is their generality, which is indicated by applying them to conceptually different driving cases. Furthermore, this thesis introduces a novel way of applying a convolutional neural network architecture to a high level state description of interchangeable objects, which speeds up the learning process and eliminates all collisions in the test cases.

genetic algorithm

Monte Carlo tree search

deep reinforcement learning

tactical decision-making

autonomous driving

neural network


Carl-Johan E Hoel

Chalmers, Mechanics and Maritime Sciences, Vehicle Engineering and Autonomous Systems

An Evolutionary Approach to General-Purpose Automated Speed and Lane Change Behavior

Proceedings of 16th IEEE International Conference On Machine Learning And Applications (ICMLA),; (2017)

Paper in proceedings

Automated Speed and Lane Change Decision Making using Deep Reinforcement Learning

IEEE Conference on Intelligent Transportation Systems, Proceedings, ITSC,; (2018)p. 2148-2155

Paper in proceedings

C. J. Hoel, K. Driggs-Campbell, K. Wolff, L. Laine, and M. J. Kochenderfer, Combining planning and deep reinforcement learning in tactical decision making for autonomous driving

Areas of Advance

Information and Communication Technology


Subject Categories

Information Science


Computer Science

Thesis for the degree of Licentiate – Department of Mechanics and Maritime Sciences: 2019:07


Chalmers University of Technology

More information

Latest update