Unsupervised Indoor Positioning System Based on Environmental Signatures.

Area of Science:

• Robotics
• Sensor Technology
• Geomatics

Background:

• Indoor positioning systems (IPS) commonly rely on mobile sensors but often require extensive calibration for accuracy.
• Existing methods face challenges with precision and user-friendliness due to calibration demands.

Purpose of the Study:

• To propose a novel unsupervised indoor positioning (UIP) system that negates the need for cumbersome calibration.
• To enhance indoor localization accuracy and reduce convergence time through an innovative approach.

Main Methods:

• UIP utilizes unique environmental features as landmarks for positioning.
• Combines dead reckoning with landmark-based SLAM (Simultaneous Localization and Mapping).
• Operates without requiring manual or complex sensor calibration.

Main Results:

• The developed UIP system demonstrates accurate indoor positioning capabilities.
• The integration of environmental signatures and SLAM effectively reduces positioning errors.
• Faster convergence times were observed compared to traditional methods.

Conclusions:

• Unsupervised indoor positioning (UIP) presents a viable, calibration-free alternative for mobile sensor-based localization.
• The system's reliance on environmental signatures offers a promising unconventional approach to indoor navigation.
• Further research into UIP could significantly advance the field of indoor positioning.