Self-supervised Deep Reinforcement Learning with Generalized Computation Graphs for Robot Navigation

TL;DR

This paper introduces a generalized computation graph for robot navigation that combines model-free and model-based reinforcement learning, enabling sample-efficient, self-supervised learning from raw images in both simulated and real-world environments.

Contribution

The authors propose a unified computation graph framework that interpolates between model-free and model-based RL, improving sample efficiency and enabling autonomous learning for robot navigation.

Findings

Outperforms single-step and N-step double Q-learning in simulation.

Successfully learns to navigate a complex indoor environment with an RC car.

Achieves autonomous, self-supervised training within a few hours.

Abstract

Enabling robots to autonomously navigate complex environments is essential for real-world deployment. Prior methods approach this problem by having the robot maintain an internal map of the world, and then use a localization and planning method to navigate through the internal map. However, these approaches often include a variety of assumptions, are computationally intensive, and do not learn from failures. In contrast, learning-based methods improve as the robot acts in the environment, but are difficult to deploy in the real-world due to their high sample complexity. To address the need to learn complex policies with few samples, we propose a generalized computation graph that subsumes value-based model-free methods and model-based methods, with specific instantiations interpolating between model-free and model-based. We then instantiate this graph to form a navigation model that learns from raw images and is sample efficient. Our simulated car experiments explore the design decisions of our navigation model, and show our approach outperforms single-step and $N$-step double Q-learning. We also evaluate our approach on a real-world RC car and show it can learn to navigate through a complex indoor environment with a few hours of fully autonomous, self-supervised training. Videos of the experiments and code can be found at github.com/gkahn13/gcg