IP Library › Granted Patent US 11,474,228
Granted Patent B2
US 11,474,228 · App. 16/559,603 · Granted Oct 18, 2022

Radar-based detection of objects while in motion

Inventors: Asaf Tzadok (New Castle, NY); Alberto Valdes Garcia (Chappaqua, NY)
Assignee: International Business Machines Corporation
G01S13/426G01S7/417G01S13/04G01S13/886G01S13/9029
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 11,474,228
App. No.
16/559,603
Granted
Oct 18, 2022
Kind
B2
Abstract

A radar-based security screening system for detecting objects is described. The screening system includes a radar transmitter, a radar receiver, and a processing unit. In use, the radar transmitter steers a radar beam across a screening volume. The radar receiver, in turn, receives a return signal from an object over time as the object moves in the screening volume to create a three-dimensional temporal signature for the object. The processing unit classifies the three-dimensional temporal signature utilizing a classification process based on a deep neural network model, and provides an alert when the object is classified as an object of interest. During screening, a screened person is not required to remain still in a confined volume and is not exposed to harmful radiation.

Claims (25)

1. A method for detecting and classifying an object, the method comprising the steps of:

steering a radar beam across a screening volume;

receiving a return signal from the object over time as the object moves in the screening volume to create a three-dimensional temporal signature for the object;

classifying the three-dimensional temporal signature utilizing a classification process based at a deep neural network model, the deep neural network model trained to process two four-dimensional tensors, each four-dimensional tensor based on a three-dimensional grid of voxels, each voxel describing the returned signal from a particular horizontal direction, a particular vertical direction and a particular depth in the screening volume; and

providing an alert when the object is classified as an object of interest.

2. The method of claim 1 , wherein the method further comprises the step of providing a level of confidence when an object is classified as an object of interest.

3. The method of claim 1 , wherein the received return signal is stored in a memory device as a data structure represented by the four-dimensional tensor.

4. The method of claim 1 , further comprising the step of inviting feedback on the classification process.

5. The method of claim 4 , further comprising the step of refining the classification process based at least in part on the feedback.

6. The method of claim 1 , wherein classifying the three-dimensional temporal signature utilizes video data from a video camera.

7. The method of claim 1 , wherein the step of providing an alert comprises raising, lowering, closing, or locking an object.

8. The method of claim 1 , wherein the radar beam has a wavelength between about one and about ten millimeters.

9. A screening system comprising:

a radar transmitter operative to steer a radar beam across a screening volume;

a radar receiver operative to receive a return signal from an object over time as the object moves in the screening volume to create a three-dimensional temporal signature for the object; and

a processing unit operative to classify the three-dimensional temporal signature utilizing a classification process based at least in part on a deep neural network model, the deep neural network model trained to process two four-dimensional tensors, each four-dimensional tensor based on a three-dimensional grid of voxels, each voxel describing the returned signal from a particular horizontal direction, a particular vertical direction and a particular depth in the screening volume, and to provide an alert when the object is classified as an object of interest.

10. The screening system of claim 9 , further comprising a memory storing computer readable program code executable by the logic circuitry.

11. The screening system of claim 9 , wherein the radar transmitter comprises a phased-array antenna.

12. The screening system of claim 9 , wherein the processing unit is further operative to provide a level of confidence when an object is classified as an object of interest.

13. The screening system of claim 9 , wherein the processing unit is further operative to invite feedback on the classification process.

14. The screening system of claim 13 , wherein the processing unit is further operative to refine the classification process based at least in part on the feedback.

15. The screening system of claim 9 , further comprising an alarm that may be activated by the processing unit upon classifying an object of interest.

16. The screening system of claim 9 , wherein the radar receiver is operative to resolve one or more angles of arrival of the return signal.

17. The screening system of claim 9 , further comprising a video camera providing video data, wherein the processing unit is further operative to utilize the video data in the classification process.

18. The screening system of claim 9 , wherein the radar beam has a wavelength between about one and about ten millimeters.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 3, 2019
From: TZADOK, ASAF; VALDES GARCIA, ALBERTO
To: INTERNATIONAL BUSINESS MACHINES CORPORATION
Reel/Frame 050253/0406 →
Continuity (1)
Related Publication 20210063559A1 · Mar 4, 2021