An Anchor-Based Pedestrian Navigation Approach Using Only Inertial Sensors.

Area of Science:

• Robotics and Autonomous Systems
• Geomatics Engineering
• Indoor Positioning Systems

Background:

• Inertial Measurement Units (IMUs) are crucial for pedestrian navigation but suffer from cumulative positioning errors.
• Pre-deployed anchors improve accuracy but increase system complexity, hindering applications like emergency response.
• Existing anchor-based methods often require significant prior knowledge of building structures.

Purpose of the Study:

• To propose a novel anchor-based pedestrian navigation approach that does not require pre-deployed anchors or additional sensors.
• To leverage virtual anchors, inferred from pedestrian trajectories, as a proxy for building structure characteristics.
• To enable robust and simplified indoor positioning for applications such as emergency response.

Main Methods:

• Defining anchors as intersection points of perpendicular corridors within building structures.
• Extracting virtual anchors from pedestrian trajectories to represent observed real anchors.
• Employing a Rao-Blackwellized Particle Filter (RBPF) for joint estimation of pedestrian positions and anchor maps.
• Minimizing assumptions on building structures for broader applicability.

Main Results:

• The proposed method effectively compensates for IMU-derived positioning errors using inferred anchors.
• Virtual anchors derived from trajectories serve as reliable indicators of building structures.
• The Rao-Blackwellized Particle Filter successfully performs joint trajectory and map estimation.
• Experimental validation demonstrates the method's effectiveness and robustness in real-world scenarios.

Conclusions:

• The developed approach offers a simplified and effective solution for anchor-based pedestrian navigation.
• It overcomes the limitations of pre-deployed anchors, making it suitable for dynamic environments and emergency response.
• The method exhibits robustness even when initial assumptions about building structures are not perfectly met, preventing positioning failure.