1 of 6

Human Activity Recognition Using Radar

Vinay Joshi

Dr. Hongfei Xue

2 of 6

Motivations

In recent years, radar technology has advanced significantly that it can surpass the camera. Radars can easily easily detect human activity even when obstacles block the view. This highlights the radar significance in the real world.

​

  • We want to build applications that utilize radar technology in real-life scenarios.
  • We are developing a learning model that can analyze radar data to determine the activities.
  • This could encourage other people to also research and build application around radar technology.
  • I want to learn more about radar technology and hopefully spot the B2 bomber in the sky.

3 of 6

Backgrounds

  • Concepts
    • Fast Fourier transform (Angle, Range and Doppler)
    • Machine learning(SVM, logistic regression and Decision Tree)
  • Tools
    • Python
    • Tensorflow
    • MMwave studio(radar software)
    • AWR 1843 Boost(radar)
  • Literature
    • Human activity classification based on micro-Doppler signatures using an artificial neural network(link)
    • Hand Gesture Recognition Using Micro-Doppler Signatures With Convolutional Neural Network(link)
    • Experiments with mmWave Automotive Radar Test-Bed(link)

​

​

​

​

4 of 6

Current Progress

  • Current Progress
    • Began data collection with various activities, including walking and hand signaling.
    • We need to process the data to extract information such as angular Fourier transform, short-time Fourier transform, and 3D point cloud.
    • For human classification, we have decided to use a decision tree since the dataset we are collecting is relatively small.
  • Goals
    • The goal is to utilize radar to classify human activity using micro-Doppler and display it in real time.

5 of 6

Future Plans

My future plans for this project involve implementing real-life applications for radar-based human activity detection. Some potential use cases include:

  • Senior care and Monitoring
    • Utilizing radar to monitor the activities of elderly individuals.
  • Finding people stuck under a rubble after earthquake
    • Using radar to locate individuals trapped under rubble after earthquakes, as radar waves can penetrate walls and detect human activity.
  • Creating a Surveillance environment
    • Increasing radar surveillance and its effectiveness around buildings.
  • Military uses
    • Developing UAVs (Unmanned Aerial Vehicles) and Javelin systems equipped with radar for enemy detection and targeting, particularly effective during nighttime operations.

6 of 6

Thank You

Q & A