Data Fusion Methods for Indoor Positioning Systems Based on Channel State Information Fingerprinting

Area of Science:

• Wireless communication
• Indoor positioning systems
• Signal processing

Background:

• Indoor signals face challenges like multipath effects and dynamic environments, impacting location services.
• Channel State Information (CSI) offers enhanced wireless channel metrics over traditional Received Signal Strength (RSS) fingerprinting.
• Existing CSI-based methods struggle with robustness and stability due to indoor multipath effects.

Purpose of the Study:

• To propose a novel data fusion method for robust indoor positioning using CSI.
• To address temporal variations in CSI measurements within complex indoor environments like parking lots.
• To reduce the training and calibration overhead associated with indoor positioning systems.

Main Methods:

• A positive knowledge transfer-based heterogeneous data fusion method was developed.
• The method represents temporal variations in CSI-based fingerprint measurements.
• Extensive experiments were conducted in real-world indoor parking lot scenarios.

Main Results:

• The proposed algorithm demonstrated efficient and consistent positioning accuracy across various temporal variations.
• Significant improvements in indoor parking location accuracy were achieved.
• Computationally robust and efficient location estimates were provided for dynamic environments.

Conclusions:

• The developed method effectively enhances indoor positioning accuracy in complex environments.
• The algorithm offers reliable performance despite dynamic environmental changes.
• Cramer-Rao lower bound analysis confirmed the factors influencing location error variance.