TriHelper: Zero-Shot Object Navigation with Dynamic Assistance

TL;DR

TriHelper is a new framework that improves zero-shot object navigation by dynamically assisting agents with collision avoidance, exploration, and detection, leading to better success rates and efficiency in unknown environments.

Contribution

The paper introduces TriHelper, a novel multi-component framework that addresses specific navigation challenges dynamically, improving zero-shot object navigation performance.

Findings

Outperforms existing methods in success rate and exploration efficiency

Each helper component significantly enhances navigation capabilities

Effective in diverse datasets like HM3D and Gibson

Abstract

Navigating toward specific objects in unknown environments without additional training, known as Zero-Shot object navigation, poses a significant challenge in the field of robotics, which demands high levels of auxiliary information and strategic planning. Traditional works have focused on holistic solutions, overlooking the specific challenges agents encounter during navigation such as collision, low exploration efficiency, and misidentification of targets. To address these challenges, our work proposes TriHelper, a novel framework designed to assist agents dynamically through three primary navigation challenges: collision, exploration, and detection. Specifically, our framework consists of three innovative components: (i) Collision Helper, (ii) Exploration Helper, and (iii) Detection Helper. These components work collaboratively to solve these challenges throughout the navigation process. Experiments on the Habitat-Matterport 3D (HM3D) and Gibson datasets demonstrate that TriHelper significantly outperforms all existing baseline methods in Zero-Shot object navigation, showcasing superior success rates and exploration efficiency. Our ablation studies further underscore the effectiveness of each helper in addressing their respective challenges, notably enhancing the agent's navigation capabilities. By proposing TriHelper, we offer a fresh perspective on advancing the object navigation task, paving the way for future research in the domain of Embodied AI and visual-based navigation.