Indoor Localization Method of Personnel Movement Based on Non-Contact Electrostatic Potential Measurements

Area of Science:

• Human-computer interaction
• Machine learning
• Smart city technologies

Background:

• Indoor localization is crucial for smart city applications like smart homes and elderly care.
• Electrostatic localization offers a passive, label-free approach balancing accuracy, power, privacy, and environmental impact.
• Existing electrostatic methods lack generality due to unique transfer functions in different environments.

Purpose of the Study:

• To develop a generalized indoor localization method using on-site measured electrostatic signals and symbolic regression.
• To overcome the limitations of environmental variability in electrostatic localization.
• To enable accurate, non-contact, and privacy-preserving human tracking.

Main Methods:

• Designed and implemented a remote, non-contact human electrostatic potential sensor.
• Developed a prototype test system for indoor localization experiments.
• Utilized symbolic regression machine learning algorithms for signal processing and localization.

Main Results:

• Achieved indoor localization of moving people in a 5 m × 5 m space.
• Demonstrated an 80% positioning accuracy with a median absolute error of 0.4–0.6 m.
• Enabled on-site calibration, eliminating the need for prior scene-specific physical information.

Conclusions:

• The proposed method offers a robust and generalizable indoor localization solution based on electrostatic signals.
• It requires low computational complexity and minimal training data.
• This approach is suitable for various smart city applications demanding accurate and passive localization.