Collecting Human Motion Data in Large and Occlusion-Prone Environments using Ultra-Wideband Localization

Abstract: With robots increasingly integrating into human environments, understanding and predicting human motion is essential for safe and efficient interactions. Modern human motion and activity prediction approaches require high quality and quantity of data for training and evaluation, usually collected from motion capture systems, onboard or stationary sensors. Setting up these systems is challenging due to the intricate setup of hardware components, extensive calibration procedures, occlusions, and substantial costs. These constraints make deploying such systems in new and large environments difficult and limit their usability for in-the-wild measurements. In this paper we investigate the possibility to apply the novel Ultra-Wideband (UWB) localization technology as a scalable alternative for human motion capture in crowded and occlusion-prone environments. We include additional sensing modalities such as eye-tracking, onboard robot LiDAR and radar sensors, and record motion capture data as ground truth for evaluation and comparison. The environment imitates a museum setup, with up to four active participants navigating toward random goals in a natural way, and offers more than 130 minutes of multi-modal data. Our investigation provides a step toward scalable and accurate motion data collection beyond vision-based systems, laying a foundation for evaluating sensing modalities like UWB in larger and complex environments like warehouses, airports, or convention centers.

Ultra-Wideband Localization for Human Motion Data Collection

The paper by Kaden et al. explores the application of Ultra-Wideband (UWB) localization technology as a scalable method for collecting human motion data in environments prone to crowding and occlusion, specifically a museum-like setting. The study integrates additional sensing modalities such as eye-tracking, LiDAR, and radar sensors onboard a robot to evaluate the effectiveness of UWB localization compared to traditional motion capture systems.

Overview of the Methodology

The investigation centers around a room setup designed to simulate a museum environment, thoughtfully arranged to direct visual attention akin to exhibit navigation. The authors introduce various sensor modalities, including UWB tags and smartphones for tracking participants' movements, along with eye-tracking glasses that record gaze information. Complementing these are a quadruped robot equipped with LiDAR and radar sensors to capture point cloud data. Motion capture technology serves as a ground truth reference.

The data collection scenarios encompass constructing room layouts to mimic exhibition spaces, navigating to various goal points to study motion patterns, and employing a robot to introduce dynamic interactions. These scenarios enable the gathering of over 130 minutes of multimodal data, providing rich datasets for analysis.

Key Findings and Results

The incorporation of UWB technology demonstrates promising results, as evidenced by recorded trajectories showing a mean 2D displacement error of approximately 41 centimeters against ground truth motion capture data. Moreover, the study showcases the mapping capabilities of onboard sensors in generating environmental point cloud maps despite occlusions—a testament to the potential utility in dynamic, complex settings like museums. The LiDAR sensor captures detailed geometry, while radar data contributes velocity information crucial for tracking dynamic entities.

Implications and Future Directions

This study presents a significant advancement towards scalable motion data collection in environments large enough to render traditional methods impractical. The implications of this research stretch across multiple domains, including robotics, indoor navigation, and human motion prediction. The adoption of consumer-friendly UWB technology promises easier integration into such spaces, leveraging device interoperability and enhancing user experiences in real-world applications.

However, further exploration is necessary to fully validate UWB's precision and applicability in diverse occluded environments. Future work may focus on refining UWB devices for even greater accuracy and expanding the capabilities of radar for human motion tracking without reliance on motion capture systems. This research is a foundational step towards large-scale, practical human-robot interaction evaluations.

Conclusion

This paper underscores the potential of UWB localization technology in conjunction with various sensing modalities to efficiently collect human motion data in complex indoor settings. By doing so, it lays the groundwork for extensive dataset creation essential for advancing human motion prediction algorithms and facilitating seamless robot-human interaction.