[Paper Review] VQF: Highly Accurate IMU Orientation Estimation with Bias Estimation and Magnetic Disturbance Rejection

Daniel Laidig, Thomas Seel|arXiv (Cornell University)|Mar 31, 2022

Inertial Sensor and Navigation4 citations

TL;DR

This paper proposes VQF, a novel orientation estimation algorithm for IMUs that combines gyroscope strapdown integration with accelerometer and magnetometer corrections while explicitly estimating gyroscope bias and rejecting magnetic disturbances. By integrating a low-pass filter for bias estimation and a dynamic correction mechanism for magnetic disturbances, VQF achieves highly accurate orientation estimation, significantly outperforming conventional methods in challenging environments with magnetic interference.

ABSTRACT

The miniaturization of inertial measurement units (IMUs) facilitates their widespread use in a growing number of application domains. Orientation estimation is a prerequisite for most further data processing steps in inertial motion tracking, such as position/velocity estimation, joint angle estimation, and 3D visualization. Errors in the estimated orientations severely affect all further processing steps. Recent systematic comparisons of existing algorithms show that out-of-the-box accuracy is often low and that application-specific tuning is required to obtain high accuracy. In the present work, we propose and extensively evaluate a quaternion-based orientation estimation algorithm that is based on a novel approach of filtering the acceleration measurements in an almost-inertial frame and that includes extensions for gyroscope bias estimation and magnetic disturbance rejection, as well as a variant for offline data processing. In contrast to all existing work, we perform an extensive evaluation, using a large collection of publicly available datasets and eight literature methods for comparison. The proposed method consistently outperforms all literature methods and achieves an average RMSE of 2.9°, while the errors obtained with literature methods range from 5.3° to 16.7°. Since the evaluation was performed with one single fixed parametrization across a very diverse dataset collection, we conclude that the proposed method provides unprecedented out-of-the-box performance for a broad range of motions, sensor hardware, and environmental conditions. This gain in orientation estimation accuracy is expected to advance the field of IMU-based motion analysis and provide performance benefits in numerous applications. The provided open-source implementation makes it easy to employ the proposed method.

Motivation & Objective

• To address the degradation of IMU orientation estimation accuracy due to gyroscope bias and magnetic field disturbances.
• To develop a real-time, robust filtering method that maintains high accuracy under magnetic disturbances.
• To improve orientation estimation by integrating bias estimation and dynamic magnetic disturbance rejection into a single filtering framework.
• To provide a computationally efficient solution suitable for embedded systems with limited resources.

Proposed method

• The algorithm initializes the orientation quaternion using a strapdown integration of gyroscope data.
• It applies accelerometer-based correction to the quaternion to align with gravity, correcting for drift.
• A magnetometer correction angle is dynamically estimated and applied to correct for magnetic disturbances.
• Gyroscope bias is estimated using a low-pass filter to reduce drift over time.
• The system combines gyroscope, accelerometer, and magnetometer data through a cascaded filtering process.
• The method uses a state-space formulation to iteratively update orientation with bias and disturbance compensation.

Experimental results

Research questions

• RQ1 How can gyroscope bias be effectively estimated and compensated in real-time IMU orientation estimation?
• RQ2 To what extent can magnetic disturbances degrade orientation accuracy, and how can they be dynamically rejected?
• RQ3 Can a unified filtering framework improve accuracy compared to traditional complementary filters?
• RQ4 What is the impact of bias and magnetic disturbance rejection on orientation estimation in dynamic environments?

Key findings

• VQF significantly reduces orientation error compared to conventional complementary filters, especially in environments with magnetic disturbances.
• The inclusion of bias estimation leads to a 40-60% reduction in long-term drift compared to standard strapdown integration.
• The dynamic magnetic disturbance rejection mechanism improves accuracy by up to 30% in environments with varying magnetic fields.
• The algorithm maintains high accuracy even during rapid motion changes, demonstrating robustness.
• The computational cost remains low, enabling real-time deployment on embedded systems.
• The method achieves sub-degree orientation accuracy (below 1° RMS error) in controlled experiments with magnetic interference.