Location-Based Safety Management with Wearable Seamless Indoor-Outdoor Positioning

The gist The system presents a practical safety-oriented positioning system that uses resource-constrained wearable tags and a device–cloud split architecture for continuous outdoor–indoor worker tracking.

System Architecture

The system is designed as an end-to-end positioning platform for LBSM using resource-constrained wearables, integrating five main components: wearable tags, a positioning server, a database, a dashboard server, and a map tool To meet the stringent requirements of positioning accuracy, latency, and efficiency, its architecture distributes processing to balance computational demand with hardware efficiency Each tag performs local sensing, lightweight fusion, and communication while the positioning server handles computationally intensive sensor fusion using device data and environmental information stored in the map database. The tags communicate with the server over low-bandwidth links using selectable IoT protocols, including CoAP, HTTP, and MQTT.

Wearable Tag

The wearable platform is designed for LBSM under strict size, power, and cost constraints while supporting continuous sensing during daily worker activities. As shown in Fig. 2, the system is built around a 32-bit dualcore processor that coordinates sensor acquisition, on-board processing, wireless communication, memory, interfaces, and power control. The device integrates a dual-band L1/L5 GNSS receiver and an I2C-based sensor suite. The 6-DoF IMU combines a 16-bit triaxial accelerometer and gyroscope for motion estimation while a 3-DoF magnetometer provides geomagnetic orientation cues. Power management is handled through a protected switching converter for the main 3.3 V rail and an ultra-low-dropout regulator for the 1.8 V sensor rail reducing thermal loss and improving sensor stability.

On-board Processing

The device software follows a layered, object-oriented architecture to support modularity, portability, and real-time operation. The top software layer implements the main application logic through reusable libraries and event-driven tasks. On-board PDR is a core function of the wearable IoT device especially under bandwidth-constrained communication transforming high-rate IMU data into compact, low-rate motion increments which greatly reduce communication load while retaining the information needed for cloud-based map-aware fusion. Algorithm 1 illustrates the main pipeline of the on-board PDR implemented on the wearable tag where S(·) detects stance onsets, G(·) estimates gyroscope bias, H(·) updates orientation with bias-corrected AHRS, and D(·) estimates stride length.

Positioning Server

The positioning server serves as the system’s central processing engine using a distributed worker architecture where each processing worker handles one wearable tag and maintains its own parameters and state variables for consistent trajectory estimation. When a packet arrives, the server retrieves the corresponding building context and indoor layouts from the map database then executes the Map-Aware Particle Filter (MAPF). MAPF replaces the fixed predict–update–resample cycle of conventional particle filters with an event-driven design that updates whenever motion, GNSS, or pressure data arrive matching the system’s asynchronous sensing flow. Algorithm 2 summarizes the procedure where C(·) enforces wall-collision and entrance-based outdoor–indoor transition constraints; αk = C(σh)Sv weights GNSS updates using horizontal accuracy and velocity consistency; B(·) detects vertical transitions using DBSCAN and estimates the current floor lk and transition context qk using an SVM-based classifier.

Database and Dashboard

The proposed LBSM system uses two connected MongoDB repositories: a general database and an indoor-map database supporting seamless data exchange across storage, localization, and visualization. The general database stores dynamic operational data including user profiles, device records, sensor measurements, and positioning and context outputs used for worker tracking. The indoor-map database stores spatial data for map-aware fusion and visualization such as building metadata, floor plans, occupancy maps, reachable areas, landmarks. The dashboard provides a unified web interface for device registration worker assignment and live monitoring across indoor and outdoor spaces visualizing the worksite in both 2D and 3D.

Evaluation Results

Experimental results show that the system offers a practical balance among accuracy, latency, and resource efficiency for scalable LBSM. The system achieves 1.32 m mean error and 1.80 m RMSE indoors and 3.30 m mean error and 3.64 RMSE outdoors over a ∼1400 m testbed. The system maintains low typical latency over IoT-grade links, although protocol choice significantly affects robustness to intermittent connectivity. The platform-side alert-generation latency was 1.76 s for geofence-based access detection indicating timely and reliable location-based access alerts. The device provides approximately 10–12 h of operation depending on network and environmental conditions.

Future Work

Future work will enhance robustness in challenging deployments by exploring NPU-assisted data-driven PDR on the wearable tag improving energy efficiency and thermal stability for longterm use. Efforts will move the platform toward continuous, scalable, and field-deployable safety management. The system supports map management, real-time monitoring, geofencing-based access alerts, and safety/performance analytics.

References

[1] W. W. S. Chung, S. Tariq, S. R. Mohandes, and T. Zayed “Iotbased application for construction site safety monitoring” International Journal of Construction Management vol 23 no 1 pp 58–74 2023

[2] V.-T. Nguyen, T.-A. Mac, V.-A. Nguyen, A.-V. Vu, and D. Han “Edge-centric indoor–outdoor positioning using LoRa communication for emergency response” in Proceedings of the 2026 International Conference on Indoor Positioning and Indoor Navigation (IPIN) Rome Italy Oct 2026 to appear

[3] R. Jin, H. Zhang, D. Liu, and X. Yan “Iot-based detecting, locating and alarming of unauthorized intrusion on construction sites” Automation in Construction vol 118 p 103278 2020

[4] J. Chen, B. Zhou, S. Bao, X. Liu, Z. Gu, L. Li, Y. Zhao, J. Zhu, and Q. Li “A data-driven inertial navigation/bluetooth fusion algorithm for indoor localization” IEEE Sensors Journal vol 22 no 6 2021

[5] Y. Hu, F. Qian, Z. Yin, Z. Li, Z. Ji, Y. Han, Q. Xu, and W. Jiang “Experience: Practical indoor localization for malls” in Proceedings of the 28th Annual International Conference on Mobile Computing And Networking 2022 pp 82–93

[6] P. Mayer, M. Magno, and L. Benini “Self-sustaining ultrawideband positioning system for event-driven indoor localization” IEEE Internet of Things Journal vol 11 no 1 pp 1272–1284 2023

[7] S. Zhang, H. Tang, L. Chen, and Y. Gao “A seamless pedestrian localization system based on gnss/imu/uwb/map integration” IEEE Internet of Things Journal vol 12 no 11 pp 16 653–16 667 2025

[8] J. Wang, C. Shi, M.

1. A more robust on-board processing pipeline could be developed by exploring NPU-assisted data-driven PDR on the wearable tag to enhance motion summary generation, which would allow the device to handle higher data rates or more complex fusion algorithms locally, potentially reducing reliance on cloud communication and improving real-time responsiveness.

2. The system can transition from simple geofencing alerts to proactive hazard avoidance by integrating map matching against walls and entrance constraints within the server's Map-Aware Particle Filter (MAPF), enabling the system to not just alert a worker entering a restricted zone but to suggest or enforce safe paths around known obstacles.

3. The location-based analytics module can be significantly enhanced by leveraging the accumulated traces [which] can also reveal behavioral patterns, such as collaboration through overlapping paths or shared work areas, allowing the AI system to automatically flag potentially unsafe collaborative activities based on spatial proximity and movement patterns.

4. To improve resilience in low-connectivity scenarios, the communication layer could be optimized by tuning protocol selection based on predicted link quality; for instance, using CoAP achieves the lowest latency when connectivity is normal, while adapting dynamically to switch to MQTT if a connection degrades, ensuring that critical safety alerts are delivered with minimal delay.

5. The map tool's capability can be expanded by integrating machine learning models for automated floor classification and map refinement from crowdsourced data, moving beyond manual alignment to create a navigable spatial model of the indoor environment that is continuously updated and verified by the system.