Identification of Human Motion Using Radar Sensor in an Indoor Environment

Area of Science:

• Engineering
• Computer Science
• Signal Processing

Background:

• Human motion identification is crucial for applications like healthcare and security.
• Existing methods often rely on cameras or wearable sensors, which have limitations.

Purpose of the Study:

• To develop a non-invasive method for human motion identification using millimeter wave radar.
• To accurately detect motion transitions and classify specific human actions.

Main Methods:

• Utilized a millimeter wave radar sensor for data acquisition.
• Implemented parallel signal processing: statistical analysis for motion change detection and deep learning for action classification.
• Employed radar spectrograms, representing distance change over time, as input for both processes.

Main Results:

• The method accurately identifies motion change moments with a 3% error rate.
• Human actions are classified with over 95% accuracy.
• Performance was validated using indoor radar sensor data.

Conclusions:

• The proposed millimeter wave radar-based method offers a highly accurate and reliable approach to human motion identification.
• This technology has potential applications in various fields requiring unobtrusive human activity monitoring.