Radar-based method and apparatus for generating a model of an object relative to a vehicle

1. A method of generating a model of one or more objects relative to a vehicle, the method comprising:

receiving radar information in the form of in-phase quadrature (IQ) data;

converting the IQ data to one or more first range-doppler maps; and

evaluating the one or more first range-doppler maps with a machine learning model to generate the model that captures detection of the one or more objects relative to the vehicle.

2. A method according to claim 1 further comprising receiving information relating to a speed of the vehicle, wherein evaluating the one or more first range-doppler maps comprises evaluating the one or more first range-doppler maps and the information relating to the speed of the vehicle with the machine learning model to generate the model that captures the detection of the one or more objects relative to the vehicle.

3. A method according to claim 2 further comprising:

converting the information relating to the speed of the vehicle to a grayscale image; and

combining the one or more first range-doppler maps and the grayscale image into a combined image,

wherein evaluating the one or more first range-doppler maps and information relating to the speed of the vehicle comprises evaluating the combined image with the machine learning model to generate the model that captures the detection of the one or more objects relative to the vehicle.

4. A method according to claim 1 wherein receiving radar information comprises receiving radar information for each of a plurality of channels, and wherein converting the IQ data to one or more first range-doppler maps comprises converting the IQ data to a plurality of first range-doppler maps with each first range-doppler map associated with the radar information for a respective channel.

5. A method according to claim 1 further comprising creating a second range-doppler map associated with a respective first range-doppler map, wherein creating the second range-doppler map comprises creating the second range-doppler map based upon noise data selected from the IQ data in an instance in which the one or more objects are not present.

6. A method according to claim 5 wherein the machine learning model comprises a range-doppler bin estimation neural network, and wherein evaluating the one or more first range-doppler maps comprises evaluating the one or more first range-doppler maps and the second range-doppler map with the range-doppler bin estimating neural network to generate range and doppler bins corresponding to the range and velocity of a respective object of the one or more objects.

7. A method according to claim 6 further comprising identifying phase and magnitude components of the range-doppler bin corresponding to the range and the velocity of the respective object.

8. A method according to claim 7 wherein the machine learning model comprises an azimuth angle regression neural network, and wherein evaluating the one or more first range-doppler maps comprises evaluating the one or more first range-doppler maps and the second range-doppler map with the azimuth angle regression neural network to generate the azimuth angle of the respective object.

9. A method according to claim 1 wherein the vehicle comprises an air vehicle that is airborne and the one or more objects comprises an aircraft landing surface, and wherein the model of the one or more objects comprises a pose of the aircraft landing surface including one or more of an altitude, pitch or roll of the aircraft landing surface.

10. A method according to claim 1 wherein the vehicle comprises an air vehicle that is airborne and the one or more objects comprises another air vehicle that is also airborne.

11. A method according to claim 1 wherein the vehicle comprises an air vehicle that is on the ground and the one or more objects comprises an obstacle that is also on the ground.

12. A computing device configured to generate a model of one or more objects relative to a vehicle, the computing device comprising:

a communication interface configured to receive radar information in the form of in-phase quadrature (IQ) data; and

processing circuitry configured to convert the IQ data to one or more first range-doppler maps and to evaluate the one or more first range-doppler maps with a machine learning model to generate the model that captures detection of the one or more objects relative to the vehicle.

13. A computing device according to claim 12 wherein the communication interface is further configured to receive information relating to a speed of the vehicle, and wherein the processing circuitry is further configured to evaluate the one or more first range-doppler maps and the information relating to the speed of the vehicle with the machine learning model to generate the model that captures the detection of the one or more objects relative to the vehicle.

14. A computing device according to claim 13 wherein the processing circuitry is further configured to:

convert the information relating to the speed of the vehicle to a grayscale image; and

combine the one or more first range-doppler maps and the grayscale image into a combined image,

wherein the processing circuitry is also configured to evaluate the one or more first range-doppler maps and information relating to the speed of the vehicle by evaluating the combined image with the machine learning model to generate the model that captures the detection of the one or more objects relative to the vehicle.

15. A computing device according to claim 12 wherein the communication interface is configured to receive radar information by receiving radar information for each of a plurality of channels, and wherein the processing circuitry is configured to convert the IQ data to one or more first range-doppler maps by converting the IQ data to a plurality of first range-doppler maps with each first range-doppler map associated with the radar information for a respective channel.

16. A computing device according to claim 12 wherein the processing circuitry is further configured to create a second range-doppler map associated with a respective first range-doppler map, wherein the processing circuitry is configured to create the second range-doppler map by creating the second range-doppler map based upon noise data selected from the IQ data in an instance in which the object is not present.

17. A computing device according to claim 16 wherein the machine learning model comprises a range-doppler bin estimation neural network, and wherein the processing circuitry is configured to evaluate the one or more first range-doppler maps by evaluating the one or more first range-doppler maps and the second range-doppler map with the range-doppler bin estimating neural network to generate range and doppler bins corresponding to the range and velocity of a respective object of the one or more objects.

18. A computing device according to claim 17 wherein the processing circuitry is further configured to identify phase and magnitude components of the range-doppler bin corresponding to the range and the velocity of the respective object.

19. A computing device according to claim 16 wherein the machine learning model comprises an azimuth angle regression neural network, and wherein the processing circuitry is configured to evaluate the one or more first range-doppler maps by evaluating the one or more first range-doppler maps and the second range-doppler map with the azimuth angle regression neural network to generate the azimuth angle of a respective object of the one or more objects.

20. A computing device according to claim 12 wherein the vehicle comprises an air vehicle that is airborne and the one or more objects comprises an aircraft landing surface, and wherein the model of the one or more objects comprises a pose of the aircraft landing surface including one or more of an altitude, pitch or roll of the aircraft landing surface.