GridiLoc: A Backtracking Grid Filter for Fusing the Grid Model with PDR Using Smartphone Sensors

Area of Science:

• Computer Science
• Robotics
• Geomatics Engineering

Background:

• Pedestrian Dead Reckoning (PDR) with map filtering improves indoor localization but struggles in complex environments.
• Close corners and neighboring passages in complex indoor spaces can cause tracking failures.

Purpose of the Study:

• To develop a reliable and accurate pedestrian indoor localization method using smartphone sensors and a grid model.
• To address the limitations of existing methods in complex indoor spaces and dead-ending scenarios.

Main Methods:

• Proposed GridiLoc method fusing smartphone sensors and a grid model.
• Utilized a backtracking grid filter to enhance localization accuracy and manage dead ends.
• Introduced a topological graph to represent candidate backtracking points, reducing computational time.
• Implemented automatic calibration of the PDR step length model using historical data.

Main Results:

• GridiLoc demonstrated higher localization accuracy and reliability compared to standard map filtering approaches.
• The method effectively handles dead-ending issues in complex indoor environments.
• GridiLoc maintained acceptable computational complexity.

Conclusions:

• GridiLoc offers a robust solution for accurate pedestrian indoor localization, especially in challenging environments.
• The fusion of sensor data, grid modeling, and backtracking significantly improves performance.
• The method provides a practical and efficient approach for real-world indoor navigation applications.