Detection & Diagnostic Systems Multimedia

Millimeter-Wave Remote Biometric Identification and Tracking (RBIT) System for Security Applications

Argonne National Laboratory and Northwestern University

Argonne, in collaboration with Northwestern University, has developed the first heart/respiration/movement biometric system that remotely identifies and tracks moving individuals by using a novel mmW sensor, facial-recognition video sensor/camera, and feature-extraction and data-analysis software. The portable RBIT system can be rapidly deployed for security screening, health of soldiers in battlefield and disaster rescue efforts…

Transcript of Video

Our product enables remote wireless detection of human vital signs without the subject's knowledge. While a terrorist may seem cool and composed on the outside, his involuntary response would give him away. In addition, live friendly soldiers can be found in dangerous areas without risking personnel for recon.

Our system works by generating low-energy millimeter waves. They are able to penetrate through many common materials, including clothing and masonry and efficiently reflect off optically rough surfaces and yet they provide a high degree of sensitivity to minute displacements. As the wave reflects from the target, its phase is shifted proportionally to the movement of the target. The phase can be extracted and the displacement of the target can be demodulated.

When this system is aimed at a human's chest, we can see the chest's movements. There are components of chest motion due to heartbeats, breathing, and gross movement. Using signal processing, we eliminate noise and motions due to gross movement and analyze only those due to heartbeats and breathing.

Since the antenna has to be well-aimed at the subject's chest, we employ cameras to identify and track the subject, and a pan/tilt base to move the whole assembly for constant aiming. A color camera locks onto the subject's face, learns a hue histogram of the face, and performs mean-shift tracking. At the same time, a depth camera segments the image based on the depth of the target so that only pixels at the right depth are considered for processing.

The millimeter-wave system itself can be seen here. First, and most importantly, is the millimeter-wave sensor and antenna assembly, which is what gathers the in-phase and quadrature components of the signal and returns it to the computer for processing. This unique, low power all solid state millimeter wave sensor itself can be battery operated. Then, there are the cameras, which perform the real-time detection and tracking of a human subject, and return that data to the pan and tilt base, which is what moves the entire assembly, and follows the subject as he moves around and maintains the aim at his chest.

In the following demonstration, the screen is split into three parts. In the top left is the system assembly which shall be moving to follow the subject. In the top right is the subject as seen by the color tracking camera, where the oval and rectangle are the location and scale of his head, and the purple circle is the location the antenna is aiming. In the bottom half is the measure of displacement as it is being calculated by the computer in real time. The red circles represent calculated heartbeats. Also on the graph are the locations of heartbeats obtained from an ECG that was gathered simultaneously, shown in green. This is a ground truth, used for verification, and was not used at all in the algorithm.

The sensor starts off aimed at the subject's chest. While he is seated, the subject holds his breath for a while in order to demonstrate that individual heartbeats are visible. As the subject stands up, the system follows his movements and maintains good aim. Now the subject demonstrates the fidelity of the system during lateral motion. Here the subject demonstrates the accuracy of the system during large gross movement. The subject’s backward movement is larger than that due to heartbeat and respiration, as is evidenced by the scale of the displacement on the bottom of the screen. Even with such large movements, the system continues to process the data with good accuracy.

This system works on various body types, such as males and females, with larger and smaller frames and builds.

Although our system is currently intended for homeland and national security applications, this novel technology has a wide range of applications for stand-off monitoring of non-stationary targets, including home healthcare, collision avoidance, condition monitoring of operating machinery, and non-destructive evaluation.