Neural Inertial Odometry from Lie Events

2025-05-14

AI Summary

Existing Neural Displacement Priors (NDPs) exhibit limited generalization and robustness across varying IMU sampling rates and trajectory morphologies. To address this, we propose Lie Events—a novel IMU preintegration change-sampling mechanism triggered by SE(3) Lie algebra norm threshold crossing—marking the first adaptation of event-camera principles to IMU modeling. Leveraging zero-crossing sampling and normalized Lie polarity, our approach achieves intrinsic invariance to motion scale, velocity, and trajectory shape. The method integrates Lie-group preintegration, event-driven sampling, and uncertainty-aware NDPs. Evaluated on standard inertial odometry benchmarks, it reduces trajectory error by 21% while requiring only minimal preprocessing. This yields significantly improved localization accuracy and cross-device generalization under dynamic conditions.

Technology Category

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

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Large-scale security measurements

Abstract

Neural displacement priors (NDP) can reduce the drift in inertial odometry and provide uncertainty estimates that can be readily fused with off-the-shelf filters. However, they fail to generalize to different IMU sampling rates and trajectory profiles, which limits their robustness in diverse settings. To address this challenge, we replace the traditional NDP inputs comprising raw IMU data with Lie events that are robust to input rate changes and have favorable invariances when observed under different trajectory profiles. Unlike raw IMU data sampled at fixed rates, Lie events are sampled whenever the norm of the IMU pre-integration change, mapped to the Lie algebra of the SE(3) group, exceeds a threshold. Inspired by event-based vision, we generalize the notion of level-crossing on 1D signals to level-crossings on the Lie algebra and generalize binary polarities to normalized Lie polarities within this algebra. We show that training NDPs on Lie events incorporating these polarities reduces the trajectory error of off-the-shelf downstream inertial odometry methods by up to 21% with only minimal preprocessing. We conjecture that many more sensors than IMUs or cameras can benefit from an event-based sampling paradigm and that this work makes an important first step in this direction.

Problem

• Reducing drift in inertial odometry using neural displacement priors
• Improving robustness across varying IMU sampling rates
• Enhancing trajectory accuracy with Lie event-based inputs

Innovation

• Replace raw IMU data with Lie events
• Generalize level-crossing to Lie algebra
• Reduce trajectory error by 21%