Suite-IN++: A FlexiWear BodyNet Integrating Global and Local Motion Features from Apple Suite for Robust Inertial Navigation

TL;DR

The paper tackles robust pedestrian localization in flexible wearable ecosystems by leveraging multiple consumer devices to form a flexiwear bodynet. It introduces Suite-IN++, a framework that decouples global and local motion features, uses a reliability-based weighted fusion for global motion, and applies attentive local analysis to capture cross-device local dynamics, all trained with contrastive and orthogonality constraints. A new real-life Apple Suite dataset validates the approach, showing substantial improvements in ATE and RTE across diverse walking modes and device configurations, including unseen modes. The results demonstrate strong generalization and real-time deployment potential on mobile devices, highlighting the practical viability of multi-device inertial navigation in GNSS-denied environments.

Abstract

The proliferation of wearable technology has established multi-device ecosystems comprising smartphones, smartwatches, and headphones as critical enablers for ubiquitous pedestrian localization. However, traditional pedestrian dead reckoning (PDR) struggles with diverse motion modes, while data-driven methods, despite improving accuracy, often lack robustness due to their reliance on a single-device setup. Therefore, a promising solution is to fully leverage existing wearable devices to form a flexiwear bodynet for robust and accurate pedestrian localization. This paper presents Suite-IN++, a deep learning framework for flexiwear bodynet-based pedestrian localization. Suite-IN++ integrates motion data from wearable devices on different body parts, using contrastive learning to separate global and local motion features. It fuses global features based on the data reliability of each device to capture overall motion trends and employs an attention mechanism to uncover cross-device correlations in local features, extracting motion details helpful for accurate localization. To evaluate our method, we construct a real-life flexiwear bodynet dataset, incorporating Apple Suite (iPhone, Apple Watch, and AirPods) across diverse walking modes and device configurations. Experimental results demonstrate that Suite-IN++ achieves superior localization accuracy and robustness, significantly outperforming state-of-the-art models in real-life pedestrian tracking scenarios.

Figures (11)

• Figure 1: Our innovative flexiwear bodynet-based approach for robust pedestrian localization: Achieving robust positioning under complex walking modes and flexible device configurations by integrating global and local motion features from a flexiwear bodynet.
• Figure 2: Introduction to the dataset, including illustrations of device wearing, flexible device configuration, and complex walking modes.
• Figure 3: Overview of our flexiwear-bodynet-based positioning framework. (a) shows the device and flexiwear bodynet used for data collection. (b) provides an overview of the Suite-IN++ algorithm, which consists of three key modules: (1) global and local motion feature extraction, (2) a weighted global fusion module (detailed in (c)), and (3) an attentive local analysis module (detailed in (d)). These modules collectively aggregate motion information from the flexiwear bodynet while distinguishing between global and local motion, enhancing the accuracy and consistency of position estimation.
• Figure 4: Selected visualizations. We select 3 examples from each activity range, for each sequence, we label the trajectory length and report the ATE and RTE of our method and the second-best method (the unit of ATE and RTE is $m$). Our method performs best under three different ranges.
• Figure 5: Position Estimation Error. The left is the position estimation error of the sequence where the headphones are taken off midway, the middle is the sequence where the subject sit down for a while during walking, and the right is the sequence where three devices are randomly shaken.