GI-SLAM: Gaussian-Inertial SLAM

2025-03-24

AI Summary

Existing 3D Gaussian Splatting (3DGS)-based SLAM methods exhibit strong geometric and appearance modeling capabilities but commonly neglect IMU measurements, limiting tracking accuracy, robustness, and real-time performance. To address this, we propose GI-SLAM—the first general-purpose, real-time Gaussian SLAM system with tightly coupled IMU integration, supporting monocular, stereo, and RGB-D modalities. Our key innovation lies in formulating IMU preintegration and motion constraints as differentiable loss terms embedded within the joint 3DGS optimization framework, enabling end-to-end differentiable tight coupling of photometric, geometric, and inertial errors. We further model IMU uncertainty using an error-state Kalman filter to enhance adaptability in dynamic and texture-deprived scenarios. Evaluated on EuRoC and TUM-RGBD benchmarks, GI-SLAM achieves state-of-the-art real-time performance, reducing absolute trajectory error (ATE) by approximately 32% compared to prior 3DGS-SLAM approaches.

Technology Category

Intelligent Robots: State EstimationComputer Vision: Motion & TrackingSearch and Optimization: Sampling/Simulation-based Search

Application Category

Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applicationsSearch and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAGSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

3D Gaussian Splatting (3DGS) has recently emerged as a powerful representation of geometry and appearance for dense Simultaneous Localization and Mapping (SLAM). Through rapid, differentiable rasterization of 3D Gaussians, many 3DGS SLAM methods achieve near real-time rendering and accelerated training. However, these methods largely overlook inertial data, witch is a critical piece of information collected from the inertial measurement unit (IMU). In this paper, we present GI-SLAM, a novel gaussian-inertial SLAM system which consists of an IMU-enhanced camera tracking module and a realistic 3D Gaussian-based scene representation for mapping. Our method introduces an IMU loss that seamlessly integrates into the deep learning framework underpinning 3D Gaussian Splatting SLAM, effectively enhancing the accuracy, robustness and efficiency of camera tracking. Moreover, our SLAM system supports a wide range of sensor configurations, including monocular, stereo, and RGBD cameras, both with and without IMU integration. Our method achieves competitive performance compared with existing state-of-the-art real-time methods on the EuRoC and TUM-RGBD datasets.

Problem

• Integrates inertial data into 3DGS SLAM for improved accuracy
• Enhances camera tracking robustness with IMU loss in deep learning
• Supports diverse sensor setups with or without IMU integration

Innovation

• Integrates IMU data with 3D Gaussian Splatting
• Uses IMU loss for enhanced camera tracking
• Supports diverse sensor configurations flexibly