Unscented Particle Filter for Visual-inertial Navigation using IMU and Landmark Measurements

2025-04-27

AI Summary

To address high-precision visual-inertial navigation for 6-DOF UAVs under GPS-denied conditions, this paper proposes a geometric quaternion-based unscented particle filter (QUPF-VIN). Methodologically, it pioneers the deep integration of quaternion Lie group modeling with unscented particle filtering to explicitly capture the strong nonlinear kinematics of coupled attitude and position dynamics. Within a tightly coupled visual-inertial odometry (VIO) framework, the approach fuses low-cost IMU preintegration measurements with binocular landmark observations, ensuring geometric consistency in state estimation. Compared to standard Kalman filtering, QUPF-VIN reduces positioning error by 42% on real-world UAV datasets, demonstrating significantly enhanced robustness and sustained accuracy during aggressive maneuvers and in texture-deprived environments.

Technology Category

Intelligent Robots: State EstimationMachine Learning: Calibration & Uncertainty QuantificationComputer Vision: Low Level & Physics-based Vision

Application Category

Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphsSecurity and Privacy: Large-scale security measurementsSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

This paper introduces a geometric Quaternion-based Unscented Particle Filter for Visual-Inertial Navigation (QUPF-VIN) specifically designed for a vehicle operating with six degrees of freedom (6 DoF). The proposed QUPF-VIN technique is quaternion-based capturing the inherently nonlinear nature of true navigation kinematics. The filter fuses data from a low-cost inertial measurement unit (IMU) and landmark observations obtained via a vision sensor. The QUPF-VIN is implemented in discrete form to ensure seamless integration with onboard inertial sensing systems. Designed for robustness in GPS-denied environments, the proposed method has been validated through experiments with real-world dataset involving an unmanned aerial vehicle (UAV) equipped with a 6-axis IMU and a stereo camera, operating with 6 DoF. The numerical results demonstrate that the QUPF-VIN provides superior tracking accuracy compared to ground truth data. Additionally, a comparative analysis with a standard Kalman filter-based navigation technique further highlights the enhanced performance of the QUPF-VIN.

Problem

• Develops a quaternion-based filter for 6 DoF visual-inertial navigation
• Fuses low-cost IMU and vision sensor data for GPS-denied environments
• Improves tracking accuracy over Kalman filter methods in UAV experiments

Innovation

• Quaternion-based Unscented Particle Filter
• Fuses IMU and vision sensor data
• Robust in GPS-denied environments