Exploring the Generalizability of Geomagnetic Navigation: A Deep Reinforcement Learning approach with Policy Distillation

2025-02-07

AI Summary

This work addresses the weak generalization capability of autonomous vehicle geomagnetic navigation across regions in GPS-denied environments. We propose a deep reinforcement learning method that synergistically integrates multi-teacher policy distillation and reward shaping. Specifically, diverse teacher policies are trained across multiple geographic domains, and knowledge is distilled into a unified, robust navigation policy via potential-function-guided imitation and intrinsic motivation mechanisms. To our knowledge, this is the first approach to jointly leverage multi-domain policy distillation and physics-informed reward design, significantly enhancing transfer performance and exploration efficiency in unseen regions. Simulation results demonstrate substantial improvements: 23% higher cross-domain navigation success rate, 18% shorter path length, and 31% reduction in heading deviation—outperforming state-of-the-art evolutionary geomagnetic navigation methods in overall performance.

Technology Category

• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Intelligent Robots: Localization, Mapping, and Navigation
• Multiagent Systems: Multiagent Learning

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

The advancement in autonomous vehicles has empowered navigation and exploration in unknown environments. Geomagnetic navigation for autonomous vehicles has drawn increasing attention with its independence from GPS or inertial navigation devices. While geomagnetic navigation approaches have been extensively investigated, the generalizability of learned geomagnetic navigation strategies remains unexplored. The performance of a learned strategy can degrade outside of its source domain where the strategy is learned, due to a lack of knowledge about the geomagnetic characteristics in newly entered areas. This paper explores the generalization of learned geomagnetic navigation strategies via deep reinforcement learning (DRL). Particularly, we employ DRL agents to learn multiple teacher models from distributed domains that represent dispersed navigation strategies, and amalgamate the teacher models for generalizability across navigation areas. We design a reward shaping mechanism in training teacher models where we integrate both potential-based and intrinsic-motivated rewards. The designed reward shaping can enhance the exploration efficiency of the DRL agent and improve the representation of the teacher models. Upon the gained teacher models, we employ multi-teacher policy distillation to merge the policies learned by individual teachers, leading to a navigation strategy with generalizability across navigation domains. We conduct numerical simulations, and the results demonstrate an effective transfer of the learned DRL model from a source domain to new navigation areas. Compared to existing evolutionary-based geomagnetic navigation methods, our approach provides superior performance in terms of navigation length, duration, heading deviation, and success rate in cross-domain navigation.

Problem

• Generalizability of geomagnetic navigation
• Deep reinforcement learning approach
• Cross-domain navigation strategy

Innovation

• Deep Reinforcement Learning approach
• Policy Distillation technique
• Reward shaping mechanism