KISS-IMU: Self-supervised Inertial Odometry with Motion-balanced Learning and Uncertainty-aware Inference

2026-03-06

AI Summary

This work proposes a self-supervised inertial odometry framework that eliminates the reliance on ground-truth pose labels, thereby enhancing generalization in unknown environments. The method trains solely on IMU data, leveraging lightweight LiDAR point cloud registration via ICP and pose graph optimization to generate pseudo-supervisory signals. To improve robustness and accuracy, it incorporates a motion-aware balanced learning strategy and an uncertainty-driven adaptive inference mechanism. Extensive experiments across diverse real-world platforms—including quadrupedal robots—demonstrate the effectiveness of the approach. Notably, the system achieves strong generalization performance in inertial odometry without requiring ground-truth annotations or multimodal joint learning, offering a practical solution for deployment in unstructured or GPS-denied settings.

Technology Category

• Intelligent Robots: State Estimation
• Search and Optimization: Learning to Search
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Inertial measurement units (IMUs), which provide high-frequency linear acceleration and angular velocity measurements, serve as fundamental sensing modalities in robotic systems. Recent advances in deep neural networks have led to remarkable progress in inertial odometry. However, the heavy reliance on ground truth data during training fundamentally limits scalability and generalization to unseen and diverse environments. We propose KISS-IMU, a novel self-supervised inertial odometry framework that eliminates ground truth dependency by leveraging simple LiDAR-based ICP registration and pose graph optimization as a supervisory signal. Our approach embodies two key principles: keeping the IMU stable through motion-aware balanced training and keeping the IMU strong through uncertainty-driven adaptive weighting during inference. To evaluate performance across diverse motion patterns and scenarios, we conducted comprehensive experiments on various real-world platforms, including quadruped robots. Importantly, we train only the IMU network in a self-supervised manner, with LiDAR serving solely as a lightweight supervisory signal rather than requiring additional learnable processes. This design enables the framework to ensure robustness without relying on joint multi-modal learning or ground truth supervision. The supplementary materials are available at https://sparolab.github.io/research/kiss_imu.

Problem

• inertial odometry
• self-supervised learning
• ground truth dependency
• generalization
• scalability

Innovation

• self-supervised learning
• inertial odometry
• motion-balanced training
• uncertainty-aware inference
• LiDAR supervision