Increasing the Efficiency of Policy Learning for Autonomous Vehicles by Multi-Task Representation Learning

Research output: Contribution to journalArticleScientificpeer-review

5 Downloads (Pure)


Driving in a dynamic, multi-agent, and complex urban environment is a difficult task requiring a complex decision-making policy. The learning of such a policy requires a state representation that can encode the entire environment. Mid-level representations that encode a vehicle's environment as images have become a popular choice. Still, they are quite high-dimensional, limiting their use in data-hungry approaches such as reinforcement learning. In this article, we propose to learn a low-dimensional and rich latent representation of the environment by leveraging the knowledge of relevant semantic factors. To do this, we train an encoder-decoder deep neural network to predict multiple application-relevant factors such as the trajectories of other agents and the ego car. Furthermore, we propose a hazard signal based on other vehicles' future trajectories and the planned route which is used in conjunction with the learned latent representation as input to a down-stream policy. We demonstrate that using the multi-head encoder-decoder neural network results in a more informative representation than a standard single-head model. In particular, the proposed representation learning and the hazard signal help reinforcement learning to learn faster, with increased performance and less data than baseline methods.

Original languageEnglish
Pages (from-to)701-710
Number of pages10
JournalIEEE Transactions on Intelligent Vehicles
Issue number3
Early online dateFeb 2022
Publication statusPublished - Sep 2022
MoE publication typeA1 Journal article-refereed


  • Automobiles
  • Autonomous vehicles
  • Hazards
  • Image reconstruction
  • Multi-task learning
  • Policy learning
  • Reinforcement learning
  • Representation learning
  • Task analysis
  • Trajectory
  • Vehicle dynamics


Dive into the research topics of 'Increasing the Efficiency of Policy Learning for Autonomous Vehicles by Multi-Task Representation Learning'. Together they form a unique fingerprint.

Cite this