Deep reinforcement learning for localisability-aware mapless navigation

Gao, Yan ORCID: https://orcid.org/0000-0001-5890-9717, Wu, Jing ORCID: https://orcid.org/0000-0001-5123-9861, Wei, Changyun, Grech, Raphael and Ji, Ze ORCID: https://orcid.org/0000-0002-8968-9902 2025. Deep reinforcement learning for localisability-aware mapless navigation. IET Cyber-Systems and Robotics

PDF - Accepted Post-Print Version Download (31MB)

Abstract

Mapless navigation refers to the task of searching for a collision-free path without relying on a pre-defined map. Most current works of mapless navigation assume accurate ground-truth localisation is available. However, this is not true, especially for indoor environments, where simultaneous localisation and mapping (SLAM) is needed for location estimation, which highly relies on the richness of environment features. In this work, we propose a novel deep reinforcement learning (DRL)-based mapless navigation method without relying on the assumption of the availability of localisation. Our method utilises RGB-D-based ORB-SLAM2 for robot localisation. Our policy effectively guides the robot’s movement toward the target while enhancing robot pose estimation by considering the quality of the observed features along the selected paths. To facilitate policy training, we propose a compact state representation based on the spatial distributions of map points, which enhances the robot’s awareness of areas with reliable map points. Furthermore, we suggest incorporating the relative pose error into the reward function. In this way, the policy will be more responsive to each single action. In addition, rather than utilising a pre-set threshold, we adopt a dynamic threshold to improve the policy’s adaptability to variations in SLAM performance across different environments. The experiments in localisation-challenging environments have demonstrated the remarkable performance of our proposed method. It outperforms the related DRL-based methods in terms of success rate.

Item Type:	Article
Status:	In Press
Schools:	Schools > Computer Science & Informatics Schools > Engineering
Publisher:	Wiley Open Access
ISSN:	2631-6315
Date of First Compliant Deposit:	2 June 2025
Date of Acceptance:	2 April 2025
Last Modified:	03 Jun 2025 12:00
URI:	https://orca.cardiff.ac.uk/id/eprint/178643

Actions (repository staff only)

Edit Item