Radar based sensing for retail applications

1 . A radar system for monitoring an area in a retail environment, said area comprising one or more objects, the system comprising:

at least one antenna array comprising a plurality of electromagnetic antennas;

a transmitter subsystem for applying RF (radio-frequency) signals to said plurality of electromagnetic antennas;

a receiver for receiving RF signals from said plurality of electromagnetic antennas; and

at least one processor, wherein the at least one processor is configured to:

convert the received RF signals into one or more images;

analyze the one or more images based on an object identification method or movement tracking method to generate data, said data comprises one or more of: number of said one or more objects in said area, the type of said objects; movements of said one or more objects in said area,

wherein said object identification method comprises:

subtracting an initial image representing the leakage radiation between antenna elements in said plurality of electromagnetic antennas and a captured background radiation image of an empty area;

finding local peaks in the subtracted image; and

clustering the peaks into groups according to a known geometry of the identified objects and identifying the number of objects in the area of the monitored environment.

2 . The system of claim 1 , wherein said one or more objects are one or more customers and wherein the at least one processor is configured to identify these one or more customers and the movement of these customers.

3 . The system of claim 1 , wherein the identification method further comprises:

filtering the identified local peaks by thresholding the intensity value interpolated from a pre-calculated table depending on the relative distance of the identified objects from the radar system.

4 . The system of claim 1 , wherein the identification method further comprises:

classifying the formed clusters into lanes on the area to identify which object of said one or more objects is missing from each lane by comparing the number of objects in the current image to previous image.

5 . The system of claim 2 , wherein the movement tracking method comprises:

averaging a number of generated images of the one or more images over time to filter out non-moving objects;

filtering out the averaged images false targets identifications;

detecting real targets by finding peaks in the image intensity;

clustering the peaks into a number of possible identified objects; and

tracking the identified object's location using a prediction filter.

6 . The system of claim 5 , wherein the movement tracking method further comprises classifying the identified objects into two defined groups: passersby customers and interested customers.

7 . The system of claim 1 , wherein said area comprises one or more shelves holding one or more items and wherein the at least one processor is configured to estimate the occupancy of items on the shelf over time.

8 . The system of claim 7 , wherein the movement tracking method further comprises monitoring the items on the shelf being moved by the customer.

9 . The system of claim 1 , wherein said at least one antenna array is a MIMO (Multiple Input Multiple Output) antenna array.

10 . The system of claim 1 , further comprising at least two linear arrays of antennas, wherein at least two of said at least two linear arrays are orthogonal to each other.

11 . The radar system of claim 1 , further comprising at least four linear arrays of antennas, wherein at least two of said at least four linear arrays of antennas are orthogonal to each other and at least two of said at least four linear arrays of antennas are parallel to each other.

12 . A method for monitoring an area in a retail environment, said area comprising one or more objects, the method comprising:

using at least one processor to:

receive one or more RF signals from a radar system, wherein said radar system comprises at least one antenna array comprising a plurality of electromagnetic antennas;

convert the one or more received RF signals into one or more images; and

analyze the one or more images based on an object identification method or movement tracking method to generate data, said data comprises one or more of: number of said one or more objects in said area, the type of said objects, movements of said one or more objects in said area,

wherein said object identification method comprises:

subtracting an initial image representing the leakage radiation between antenna elements and a captured background radiation image of an empty area;

finding local peaks in the subtracted image; and

clustering the peaks into groups according to a known geometry of the identified objects and identifying the number of objects in the area of the environment.

13 . The method of claim 12 , wherein said objects are items on shelves and one or more customers in said area.

14 . The method of claim 12 , comprising further analyzing the generated data to yield one or more measurements of:

customer behavior; customer's height; customer's location; time spent in front of the shelf; items picked from the shelf; items returned to the shelf; direction of arrival and direction of departure; general and specific statics, such as behavior statistics.

15 . The method of claim 14 , comprising comparing the one or more measurements to one or more predefined thresholds.

16 . The method of claim 15 , comprising generating an alert when the quantity of a given item drops below a threshold.

17 . The method of claim 12 , wherein said object identification method comprises:

filtering the identified local peaks by thresholding the intensity value interpolated from a pre-calculated table depending on the relative distance of the identified objects from the radar system.

18 . The method of claim 12 , wherein the movement tracking method comprises:

averaging a number of generated images of the one or more images over time to filter out non-moving objects;

filtering out the averaged images false targets identifications;

detecting real targets by finding peaks in the image intensity;

clustering the peaks into a number of possible identified objects; and

tracking the identified object's location using a prediction filter.