Ad Hoc Mesh Network Localization Using Ultra-Wideband for Mobile Robotics

Area of Science:

• Robotics
• Localization Systems
• Wireless Sensor Networks

Background:

• Accurate localization is critical for autonomous systems, especially in GPS-denied environments.
• Existing research on ad hoc ultra-wideband (UWB) localization with mobile and stationary nodes is limited.
• Non-static environments with dynamic actors pose significant challenges for localization accuracy.

Purpose of the Study:

• To demonstrate a high-accuracy GPS-denied ad hoc localization system for bicycle-modeled robots.
• To enable robots to localize themselves within a mesh network of mobile and stationary UWB nodes.
• To address the complexities of localization in dynamic, non-static environments.

Main Methods:

• Initial localization using a simplified iterative multi-iteration ad hoc localization system (AHLos).
• Position refinement via an unscented Kalman filter (UKF) with a constant turn rate and velocity magnitude (CTRV) model.
• Fusion of robot odometry and UWB range measurements from Decawave receivers.
• Network-based collaborative localization where nodes share position estimates.

Main Results:

• Successful demonstration of ad hoc localization for unmanned ground vehicles (UGVs) in a non-static environment.
• Improved localization accuracy through the fusion of odometry and UWB ranging data.
• Effective self-localization and mutual localization among network nodes.

Conclusions:

• The proposed system enables robust and accurate ad hoc localization in GPS-denied, dynamic environments.
• Collaborative localization through a mesh network significantly enhances the performance of individual nodes.
• The system effectively addresses the challenges of dynamic environments and node visibility changes.