FTIN: Frequency-Time Integration Network for Inertial Odometry

2025-07-21

AI Summary

Existing inertial odometry methods predominantly rely on time-domain CNNs, which struggle to capture long-range temporal dependencies in IMU sequences, thereby limiting localization accuracy. To address this, we propose the first frequency-time joint learning framework: it employs Fourier transform to extract global, energy-concentrated frequency-domain features; leverages CNNs to model local temporal patterns; and introduces a Scalar LSTM to fuse cross-domain features—effectively balancing long-range dependency modeling with redundancy suppression. This architecture pioneers the integration of frequency-domain representations into end-to-end inertial navigation learning. Evaluated on public benchmarks including RoNIN, it significantly outperforms state-of-the-art methods: achieving a 43.0% reduction in absolute trajectory error and a 13.1% reduction in relative pose error over RoNIN.

Technology Category

Intelligent Robots: State EstimationSearch and Optimization: Learning to SearchComputer Vision: Learning & Optimization for CV

Application Category

Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingGraph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphsUser Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

In recent years, machine learning has achieved significant advancements in inertial odometry. However, most existing inertial odometry methods primarily rely on CNNs in the time domain. These methods often struggle to capture long-term dependency in inertial measurement unit data, thereby constraining the potential for further improvements in localization accuracy. To address these issues, we propose a novel network architecture that integrates both frequency-domain and time-domain information. Specifically, we leverage the global view and energy compaction properties of frequency-domain learning to effectively model long-term dependency and reduce redundancy in IMU data. Additionally, we introduce a Scalar LSTM to capture sequential dependencies in the time domain, enabling cross-domain information fusion and providing a stable and reliable reference for localization. Experimental evaluations on multiple public datasets (e.g., RIDI, RoNIN, OxIOD, RNIN, TLIO, and IMUNet) demonstrate the effectiveness of the proposed frequency-time domain fusion strategy. Notably, on the RoNIN dataset, our method achieves a 43.0% reduction in absolute trajectory error and a 13.1% reduction in relative trajectory error compared to RoNIN ResNet.

Problem

• Capturing long-term dependency in IMU data
• Reducing redundancy in inertial measurement data
• Improving localization accuracy via frequency-time fusion

Innovation

• Frequency-time domain fusion for odometry
• Scalar LSTM captures sequential dependencies
• Reduces redundancy with frequency-domain learning