ADUGS-VINS: Generalized Visual-Inertial Odometry for Robust Navigation in Highly Dynamic and Complex Environments

2024-11-28

AI Summary

To address the degradation of visual-inertial odometry (VIO) accuracy in dynamic, complex environments—caused by moving objects and frequent occlusions—this paper proposes a robust VIO framework. Our method integrates promptable foundation models (e.g., SAM) into VIO for the first time, coupled with an enhanced SORT tracker to achieve precise association and motion decoupling of dynamic objects. We further design a dynamic feature suppression module and a motion consistency verification module, enabling generalized perception of unknown-category and partially occluded dynamic objects, as well as robust pose estimation, all within a tightly coupled VINS-Mono architecture. Evaluated on multiple public benchmarks and real-world highly dynamic scenes—including dense pedestrian-vehicle interactions—our approach reduces average position error by 32% while maintaining centimeter-level localization stability, significantly outperforming existing state-of-the-art methods.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Visual-inertial odometry (VIO) is widely used in various fields, such as robots, drones, and autonomous vehicles. However, real-world scenes often feature dynamic objects, compromising the accuracy of VIO. The diversity and partial occlusion of these objects present a tough challenge for existing dynamic VIO methods. To tackle this challenge, we introduce ADUGS-VINS, which integrates an enhanced SORT algorithm along with a promptable foundation model into VIO, thereby improving pose estimation accuracy in environments with diverse dynamic objects and frequent occlusions. We evaluated our proposed method using multiple public datasets representing various scenes, as well as in a real-world scenario involving diverse dynamic objects. The experimental results demonstrate that our proposed method performs impressively in multiple scenarios, outperforming other state-of-the-art methods. This highlights its remarkable generalization and adaptability in diverse dynamic environments, showcasing its potential to handle various dynamic objects in practical applications.

Problem

• Improves VIO accuracy in dynamic environments
• Handles diverse and occluded dynamic objects
• Enhances pose estimation in complex scenes

Innovation

• Enhanced SORT algorithm integration
• Promptable foundation model in VIO
• Improved pose estimation in dynamic environments