Improving UWB-Based Localization in IoT Scenarios with Statistical Models of Distance Error

Area of Science:

• * Computer Science
• * Electrical Engineering
• * Robotics

Background:

• * The Internet of Things (IoT) is rapidly expanding, driving demand for precise indoor localization and context awareness.
• * Ultra Wide Band (UWB) technology offers high accuracy for indoor positioning but is sensitive to distance estimation errors.
• * Current localization algorithms often rely on estimated inter-node distances, necessitating error modeling.

Purpose of the Study:

• * To evaluate the performance improvement of indoor localization algorithms by incorporating a statistical model for Ultra Wide Band (UWB) distance errors.
• * To propose a novel statistical model for range estimation error between UWB nodes.
• * To demonstrate the effectiveness of the proposed model in realistic indoor scenarios.

Main Methods:

• * Extensive experimental measurement campaign to collect UWB ranging data.
• * Development of a general analytical framework based on the Least Squares (LS) method.
• * Derivation of a linear statistical model for UWB distance error.
• * Application of the statistical model to enhance existing localization algorithms.

Main Results:

• * The proposed statistical model effectively characterizes the distance error in UWB systems.
• * Integration of the statistical model significantly improved the accuracy of tested localization algorithms.
• * Localization error was reduced by up to 66% in various realistic scenarios.

Conclusions:

• * A novel statistical model for UWB range estimation error can substantially enhance indoor localization accuracy.
• * The proposed method provides a practical framework for improving IoT indoor positioning systems.
• * Accurate error modeling is crucial for realizing the full potential of UWB in indoor localization.