LocSpeck: A Collaborative and Distributed Positioning System for Asymmetric Nodes Based on UWB Ad-Hoc Network and...

Area of Science:

• Robotics and Automation
• Wireless Communication
• Indoor Navigation

Background:

• Traditional indoor positioning systems often rely on fixed infrastructure like ultra-wideband (UWB) anchors.
• Asymmetric sensor capabilities and computational resources among nodes can degrade positioning accuracy for less capable devices.
• Existing systems may lack flexibility in network topology and node management.

Purpose of the Study:

• To introduce LocSpeck, a novel collaborative and distributed indoor positioning system.
• To enable accurate positioning for dynamic nodes in ad-hoc networks without fixed anchors.
• To investigate the benefits of node collaboration for improving positioning of sensor-limited devices.

Main Methods:

• Utilizes inter-node relative range measurements from UWB transceiver chips (e.g., DW1000) and Wi-Fi fingerprinting.
• Employs a distributed estimator for peer-to-peer collaboration, avoiding centralized computation.
• Integrates relative range, internal sensor data, and Wi-Fi Received Signal Strength Indicator (RSSI) readings locally on each node.

Main Results:

• Demonstrated a system capable of online node addition/removal and dynamic role assignment (anchor/rover).
• Real-world tests with smartphones showed that weaker nodes benefit significantly from collaboration with nearby nodes.
• Performance was evaluated against a commercial UWB-based system in realistic indoor navigation scenarios.

Conclusions:

• LocSpeck provides a flexible and robust indoor positioning solution for dynamic, ad-hoc networks.
• Collaborative filtering effectively mitigates the limitations of asymmetric sensor capabilities in mobile nodes.
• The system eliminates the need for fixed UWB infrastructure, offering a more adaptable positioning approach.