Focus on Impact: Indoor Exploration with Intrinsic Motivation

TL;DR

This paper introduces an intrinsic motivation-based exploration method for indoor environments, using impact-driven rewards and neural pseudo-counts, outperforming traditional extrinsic reward approaches in simulation and real-world tasks.

Contribution

The work presents a novel intrinsic reward mechanism based on impact and neural pseudo-counts, enabling effective indoor exploration without prior environment knowledge.

Findings

Outperforms DRL methods with intrinsic rewards in simulated environments.

Surpasses agents trained with dense extrinsic rewards in exploration tasks.

Demonstrates successful real-world deployment and point-goal navigation adaptation.

Abstract

Exploration of indoor environments has recently experienced a significant interest, also thanks to the introduction of deep neural agents built in a hierarchical fashion and trained with Deep Reinforcement Learning (DRL) on simulated environments. Current state-of-the-art methods employ a dense extrinsic reward that requires the complete a priori knowledge of the layout of the training environment to learn an effective exploration policy. However, such information is expensive to gather in terms of time and resources. In this work, we propose to train the model with a purely intrinsic reward signal to guide exploration, which is based on the impact of the robot's actions on its internal representation of the environment. So far, impact-based rewards have been employed for simple tasks and in procedurally generated synthetic environments with countable states. Since the number of states observable by the agent in realistic indoor environments is non-countable, we include a neural-based density model and replace the traditional count-based regularization with an estimated pseudo-count of previously visited states. The proposed exploration approach outperforms DRL-based competitors relying on intrinsic rewards and surpasses the agents trained with a dense extrinsic reward computed with the environment layouts. We also show that a robot equipped with the proposed approach seamlessly adapts to point-goal navigation and real-world deployment.