Radar Technology for Smart Surveillance and Environmental Sensing

The radar sensor is considered as an efficient sensor for human-related applications for its real-time capturing capability that suits the fast human reactions while moving and its capability to perform under harsh visionary conditions e.g., smoke, dust, darkness and sunlight. The radar capturing capability has been extended to capture the micro body motions e.g., limbs motion while walking and chest activity while breathing. The detection of such micro motions have introduced the radar as a feasible sensor for many applications in the fields of smart surveillance and healthcare. Since the past years have witnessed great development in the radar sensor manufacturing in terms of processing power, real-time performance and size-efficiency, feasibility of radar integration has been of main interest for our researchers in two main domains:

  • Integration feasibility of radar in real-life scenarios (Public places – Industrial fields – Healthcare)
  • Offering a multi-sensor perception system if required 
  • Investigating radar processing to offer a highly-descriptive data for machine learning techniques

What we offer:

  • Studies on radar feasibility in Use-Case scenarios that require human interaction.
  • Consultancy on radar hardware selection and customization based on the application constraints.
  • Professional lab tests for problem evaluation in contact with our experts in human locomotion
Sensor comparison in terms of detection capability to support multi-sensor perception systems

Micro-Doppler Based Human-Robot Classification Using Ensemble and Deep Learning Approaches

Person Identification and Body Mass Index: A Deep Learning-Based Study on Micro-Dopplers

Real-Time Capable Micro-Doppler Signature Decomposition of Walking Human Limbs

DimRad: A Radar-Based Perception System for Prosthetic Leg Barrier Traversing

  • Radar-based solution to monitor the breathing micro motion activity in standing position in contactless scenario

    The radar capability of capturing the body micro motions has been extended to capture the chest movement and estimate the vital signs in terms of breath and heart rates. The estimations are done on real-time basis in contact-less scenarios. Our algorithm is developed to extract the chest micro motions and estimate the breath rate while the human is in standing position. The estimations are done while the human is on a distance of 1- 2 m from the radar. The utilized hardware has shown great integration feasibility, and has been integrated with a thermal camera to formulate a Corona Access Checker to be used in public places as Airports. The Corona Access Checker is used as to estimate the human vital signs in terms of body temperature and breathing rate in contact-less scenarios. 

    More Information: German press release "Access Checker"

  • Radar capability of detecting the range of the targets in view has been extended to numerous applications that require dimensioning feature e.g., autonomous driving. The ranging capability with the mobile motion can be used in formulating a two-dimensional map of the sagittal plane, that can be used for extracting the dimensions of static objects e.g., stairs. The MIMO radar transmission protocol can be used for real-time sagittal plane scanning. We have developed a perception system composed of a MIMO radar and an IMU, that can be integrated in a robotic module with an angular motion and form a scan of the sagittal plane. One example for integration is for prosthetic legs to support automated obstacle traversing e.g., stair climbing  

  • Different moving targets are captured by the radar with unique signatures on both the velocity and range dimensions. Such unique signatures can be used in multiple fields that require smart surveillance in terms of human differentiation from another moving target e.g., Human-Robot collaboration. We have developed a technique based on a shallow neural network that is sued for differentiating human from a moving robot to define a human-free area. Such application has been of great feasibility integration for conveyor belts to support automated industrial process. Our results were published in 2018 in the IEEE Radar Conference (RadarConf’18).

  • Using Treadmill to apply a study on Human-ID on real-time basis in noisey and perfect scenarios to achieve high identification accuracy of 22 subjects based on half gait cycle

    Human identification is a required application in many security fields e.g., monitoring home entrance. Relying on only camera systems is subjected to many environmental limitations as darkness, sunlight and visionary-blocking circumstances as smoke and dust. Moreover, cameras are always violating the privacy aspects of street passengers and neighbors. Accordingly, the radar capability in capturing the micro-motion signatures of each human on real-time basis has been considered a feasible and smart solution for such application. We have presented a study based on a deep neural network to differentiate between 22 different subjects based on their walking signatures. A study has also shown a direct correlation between the body mass index (BMI) and the walking style. Our results were published in 2019 in the IEEE Radar Conference (RadarConf’19).  

  • Estimating human limbs motion trajectories based on combining MoCap data and radar signal model

    Unique micro-Doppler signature (μ-D) of a human body motion can be analyzed as the superposition of different body parts μ-D signatures. Extraction of human limbs μ-D signatures in real-time can be used to detect, classify and track human motion especially for safety application. We could use our MoCap system to develop a realistic simulation radar signal model. We could achieve limbs trajectories decomposition on real-time basis using machine learning to differentiate 4 main classes (base – arms – legs – feet). Our results were published in 2017 in the IEEE Radar Conference (RadarConf’17).