Learning IMU Bias with Diffusion Model

TL;DR

This paper introduces a probabilistic diffusion model to accurately predict IMU bias as a distribution, addressing the stochastic nature of bias influenced by factors like temperature and vibration, thus improving motion sensing accuracy.

Contribution

It presents a novel diffusion-based approach for modeling IMU bias as a distribution, surpassing traditional regression methods in capturing bias variability.

Findings

Enhanced bias prediction accuracy

Better modeling of stochastic bias behavior

Improved motion tracking performance

Abstract

Motion sensing and tracking with IMU data is essential for spatial intelligence, which however is challenging due to the presence of time-varying stochastic bias. IMU bias is affected by various factors such as temperature and vibration, making it highly complex and difficult to model analytically. Recent data-driven approaches using deep learning have shown promise in predicting bias from IMU readings. However, these methods often treat the task as a regression problem, overlooking the stochatic nature of bias. In contrast, we model bias, conditioned on IMU readings, as a probabilistic distribution and design a conditional diffusion model to approximate this distribution. Through this approach, we achieve improved performance and make predictions that align more closely with the known behavior of bias.