Steadily Learn to Drive with Virtual Memory

TL;DR

This paper introduces LVM, a reinforcement learning algorithm that uses virtual memory via latent dynamic models to improve data efficiency and stability in autonomous driving tasks with high-dimensional inputs.

Contribution

LVM compresses high-dimensional data into latent states and uses imagined trajectories for efficient policy learning, reducing oscillations and improving performance.

Findings

LVM outperforms existing methods in data efficiency.

LVM demonstrates improved training stability.

LVM achieves better control performance in autonomous driving.

Abstract

Reinforcement learning has shown great potential in developing high-level autonomous driving. However, for high-dimensional tasks, current RL methods suffer from low data efficiency and oscillation in the training process. This paper proposes an algorithm called Learn to drive with Virtual Memory (LVM) to overcome these problems. LVM compresses the high-dimensional information into compact latent states and learns a latent dynamic model to summarize the agent's experience. Various imagined latent trajectories are generated as virtual memory by the latent dynamic model. The policy is learned by propagating gradient through the learned latent model with the imagined latent trajectories and thus leads to high data efficiency. Furthermore, a double critic structure is designed to reduce the oscillation during the training process. The effectiveness of LVM is demonstrated by an image-input autonomous driving task, in which LVM outperforms the existing method in terms of data efficiency, learning stability, and control performance.