Do We Really Need Immediate Resets? Rethinking Collision Handling for Efficient Robot Navigation

TL;DR

This paper proposes a Multi-Collision reset Budget framework that allows robots to retry obstacle configurations within the same episode, enhancing exploration and navigation efficiency during training.

Contribution

It introduces a novel collision handling approach that decouples local collision responses from global resets, improving learning speed and success rates in robot navigation.

Findings

Accelerates early-stage exploration in training.

Improves success rate and navigation efficiency.

Small collision budgets yield the largest gains.

Abstract

Should a single collision necessarily terminate an entire navigation episode? In most deep reinforcement learning (DRL) frameworks for robot navigation, this remains the standard practice: every collision immediately triggers a global environment reset and is penalized as a complete task failure. While a collision during deployment naturally indicates task failure, applying the same treatment during training prevents the agent from exploring challenging obstacle configurations, which slows learning progress in the early training phase. In this work, we challenge this convention and propose a Multi-Collision reset Budget (MCB) framework that decouples local collision termination from global environment resets, allowing the agent to retry difficult configurations within the same episode. Experiments on multiple simulated and real-world robotic platforms show that the framework accelerates early-stage exploration and improves both success rate and navigation efficiency over conventional single-collision reset baselines, with a small collision budget producing the largest gains.