GPR-OdomNet: Difference and Similarity-Driven Odometry Estimation Network for Ground Penetrating Radar-Based Localization

2025-11-21

AI Summary

In challenging environments, ground-penetrating radar (GPR)-based robot localization suffers from low displacement estimation accuracy due to subtle variations in consecutive B-scan images. To address this, we propose a deep neural network that jointly models multi-scale feature differences and similarities. Our method introduces, for the first time, a multi-scale feature contrast mechanism: a custom network extracts hierarchical representations from sequential B-scan images, and both dissimilarity and similarity cues are fused to regress Euclidean displacement distances. Evaluated on the CMU-GPR dataset, our approach achieves a weighted RMSE of 0.449 m—improving upon the current state-of-the-art by 10.2%. It significantly enhances robustness and accuracy for fine-grained displacements, particularly under adverse conditions. This work establishes a new paradigm for all-weather, GPR-based localization.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Security and Privacy: Large-scale security measurements

Abstract

When performing robot/vehicle localization using ground penetrating radar (GPR) to handle adverse weather and environmental conditions, existing techniques often struggle to accurately estimate distances when processing B-scan images with minor distinctions. This study introduces a new neural network-based odometry method that leverages the similarity and difference features of GPR B-scan images for precise estimation of the Euclidean distances traveled between the B-scan images. The new custom neural network extracts multi-scale features from B-scan images taken at consecutive moments and then determines the Euclidean distance traveled by analyzing the similarities and differences between these features. To evaluate our method, an ablation study and comparison experiments have been conducted using the publicly available CMU-GPR dataset. The experimental results show that our method consistently outperforms state-of-the-art counterparts in all tests. Specifically, our method achieves a root mean square error (RMSE), and achieves an overall weighted RMSE of 0.449 m across all data sets, which is a 10.2% reduction in RMSE when compared to the best state-of-the-art method.

Problem

• Estimating precise robot distances from GPR B-scans with minor distinctions
• Leveraging similarity and difference features in GPR images for odometry
• Improving localization accuracy in adverse weather using neural networks

Innovation

• Neural network extracts multi-scale GPR B-scan features
• Analyzes feature similarities and differences for distance
• Estimates Euclidean distance traveled between consecutive B-scans