SWA-PF: Semantic-Weighted Adaptive Particle Filter for Memory-Efficient 4-DoF UAV Localization in GNSS-Denied Environments

2025-09-17

AI Summary

To address critical limitations of vision-based localization in GNSS-denied environments—including poor real-time performance, high environmental sensitivity, and weak generalization—this paper proposes a semantic-weighted adaptive particle filter for efficient and robust 4-degree-of-freedom (4-DoF) pose estimation. The method innovatively introduces a semantic weighting mechanism that fuses semantic features from UAV-captured imagery and low-resolution satellite imagery, integrated with multimodal matching and rapid map registration techniques. This significantly enhances localization robustness under dynamic conditions and across temporal domains. Evaluated on the self-constructed MAFS multi-altitude flight dataset, the proposed approach achieves a tenfold improvement in computational efficiency, global positioning error below 10 meters, and single-pose estimation time of only several seconds—demonstrating superior accuracy, strong generalization capability, and real-time performance.

Technology Category

• Intelligent Robots: State Estimation
• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Vision-based Unmanned Aerial Vehicle (UAV) localization systems have been extensively investigated for Global Navigation Satellite System (GNSS)-denied environments. However, existing retrieval-based approaches face limitations in dataset availability and persistent challenges including suboptimal real-time performance, environmental sensitivity, and limited generalization capability, particularly in dynamic or temporally varying environments. To overcome these limitations, we present a large-scale Multi-Altitude Flight Segments dataset (MAFS) for variable altitude scenarios and propose a novel Semantic-Weighted Adaptive Particle Filter (SWA-PF) method. This approach integrates robust semantic features from both UAV-captured images and satellite imagery through two key innovations: a semantic weighting mechanism and an optimized particle filtering architecture. Evaluated using our dataset, the proposed method achieves 10x computational efficiency gain over feature extraction methods, maintains global positioning errors below 10 meters, and enables rapid 4 degree of freedom (4-DoF) pose estimation within seconds using accessible low-resolution satellite maps. Code and dataset will be available at https://github.com/YuanJiayuuu/SWA-PF.

Problem

• Memory-efficient 4-DoF UAV localization in GNSS-denied environments
• Overcoming suboptimal real-time performance and environmental sensitivity
• Addressing limited generalization in dynamic or varying environments

Innovation

• Semantic-weighted adaptive particle filter
• Integrates UAV and satellite semantic features
• Optimized particle filtering for efficient localization