A Bayesian Filtering Approach for Error Mitigation in Ultra-Wideband Ranging.

Area of Science:

• Robotics
• Sensor Networks
• Signal Processing

Background:

• Ultra-wideband (UWB) sensors offer high accuracy and real-time performance for multi-robot cooperative tracking and positioning.
• Non-line-of-sight (NLOS) conditions caused by indoor obstacles degrade UWB ranging accuracy.

Purpose of the Study:

• To develop a novel Bayesian filtering approach for mitigating UWB ranging errors in indoor environments.
• To enhance the accuracy of cooperative tracking and positioning in multi-robot systems under NLOS conditions.

Main Methods:

• Constructed nonparametric UWB sensor models (received signal strength and time of arrival) to capture probabilistic noise characteristics.
• Developed a Bayesian filtering approach for UWB ranging error mitigation, applicable to peer-to-peer ranging and integrated state estimation.

Main Results:

• The proposed method accurately identifies line-of-sight (LOS) and NLOS scenarios with various obstacles (wood, metal) probabilistically.
• Significantly improved ranging and tracking accuracy in inter-robot ranging and mobile robot tracking experiments.
• Demonstrated effective error mitigation for UWB communication in challenging indoor environments.

Conclusions:

• The novel Bayesian filtering approach effectively addresses UWB ranging errors caused by indoor obstructions.
• The method enhances the reliability and accuracy of multi-robot systems for cooperative tasks.
• Low computational overhead makes the approach suitable for real-time robotic applications.