Feature extraction for radar
Data is received characterizing a measurement for a scene received by a plurality of sensor elements forming a sensor array. Intermediate scattering coefficients can be determined using the received data by applying to the received data a mapping of a plurality of scene sub-domains to the plurality of sensor elements. A parameter of an object within the scene can be estimated using the intermediate scattering coefficients. Related apparatus, systems, techniques, and articles are also described.
1. A method for implementation by at least one data processor forming part of at least one computing system, the method comprising:
receiving, by the at least one data processor, data characterizing a measurement for a scene received by a plurality of sensor elements forming a sensor array;
determining, using the at least one data processor and the received data, intermediate scattering coefficients by at least applying to the received data a mapping of a plurality of scene sub-domains to the plurality of sensor elements, the intermediate scattering coefficients determined for a plurality of measurement sub-domains, a measurement sub-domain of the plurality of measurement sub-domains including a grouping of measurements received by at least a subset of the plurality of sensor elements forming the sensor array, a scene sub-domain of the plurality of scene sub-domains including a plurality of associated voxels that are a subset of voxels in the scene; and
estimating, using the at least one data processor and the intermediate scattering coefficients, a parameter of an object within the scene.
2. The method of claim 1 , wherein the intermediate scattering coefficients are a product of the measurement and the mapping of the plurality of scene sub-domains to the plurality of sensor elements.
3. The method of claim 1 , wherein the intermediate scattering coefficients form one or more images or point-clouds.
4. The method of claim 1 , wherein the intermediate scattering coefficients represent reflective properties of the plurality of scene sub-domains with respect to interrogation by sensor elements corresponding to the measurement sub-domain.
5. The method of claim 4 , wherein the measurement sub-domain comprises a grouping of measurements having proximal transmitting sensor elements, proximal receiving sensor elements, and frequencies within a predefined frequency range.
6. The method of claim 5 , wherein the sensor array comprises a plurality of antenna panels, the transmitting sensor elements are proximal when the transmitting sensor elements reside on a first common antenna panel and the receiving sensor elements are proximal when the receiving sensor elements reside on a second common antenna panel.
7. The method of claim 1 , further comprising
classifying the object within the scene based on the parameter.
8. The method of claim 1 , wherein the parameter characterizes a surface of the object within the scene.
9. The method of claim 1 , wherein the parameter is a statistic of the intermediate scattering coefficients.
10. The method of claim 1 , wherein estimating the parameter uses a labeled dataset of intermediate scattering coefficients.
11. The method of claim 1 , wherein the parameter includes a class of the object within the scene.
12. The method of claim 1 , wherein the scattering coefficients are not computed for individual voxels.
13. The method of claim 1 , wherein the plurality of sensors elements reside on a plurality of panels driven by a frequency modulated continuous wave signal, the plurality of panels being independently moveable.
14. The method of claim 1 , wherein the mapping of the plurality of scene sub-domains to the plurality of sensor elements includes a block diagonal transfer matrix.
15. The method of claim 1 , wherein a separation between one of the plurality of measurement sub-domains and the plurality of scene sub-domains is greater than a separation between a voxel and an associated scene sub-domain center and greater than a separation between one of the plurality of sensor elements and an associated measurement sub-domain center.
16. The method of claim 1 , wherein applying the mapping to the received data includes computing a match-filter solution using the mapping.
17. The method of claim 1 , wherein applying the mapping to the received data includes computing, using the mapping, a least squares solution or a compressed sensing solution.
18. A non-transitory computer readable storage medium comprising executable instructions which when executed by at least one data processor forming part of at least one computing system, result in operations comprising:
receiving, by the at least one data processor, data characterizing a measurement for a scene received by a plurality of sensor elements forming a sensor array;
determining, using the at least one data processor and the received data, intermediate scattering coefficients by at least applying to the received data a mapping of a plurality of scene sub-domains to the plurality of sensor elements, the intermediate scattering coefficients determined for a plurality of measurement sub-domains, a measurement sub-domain of the plurality of measurement sub-domains including a grouping of measurements received by at least a subset of the plurality of sensor elements forming the sensor array, a scene sub-domain of the plurality of scene sub-domains including a plurality of associated voxels that are a subset of voxels in the scene; and
estimating, using the at least one data processor and the intermediate scattering coefficients, a parameter of an object within the scene.
19. A system comprising:
a plurality of panels having one or more sensor elements;
memory and at least one data processor forming part of at least one computing system which implements operations comprising:
receiving, by the at least one data processor, data characterizing a measurement for a scene received by a plurality of sensor elements forming a sensor array;
determining, using the at least one data processor and the received data, intermediate scattering coefficients by at least applying to the received data a mapping of a plurality of scene sub-domains to the plurality of sensor elements, the intermediate scattering coefficients determined for a plurality of measurement sub-domains, a measurement sub-domain of the plurality of measurement sub-domains including a grouping of measurements received by at least a subset of the plurality of sensor elements forming the sensor array, a scene sub-domain of the plurality of scene sub-domains including a plurality of associated voxels that are a subset of voxels in the scene; and
estimating, using the at least one data processor and the intermediate scattering coefficients, a parameter of an object within the scene.
20. The system of claim 19 , wherein the intermediate scattering coefficients are a product of the measurement and the mapping of the plurality of scene sub-domains to the plurality of sensor elements.
21. The system of claim 19 , wherein the intermediate scattering coefficients form one or more images or point-clouds.
22. The system of claim 19 , wherein the intermediate scattering coefficients represent reflective properties of the plurality of scene sub-domains with respect to interrogation by sensor elements corresponding to the measurement sub-domain.
23. The system of claim 19 , wherein the measurement sub-domain comprises a grouping of measurements having proximal transmitting sensor elements, proximal receiving sensor elements, and frequencies within a predefined frequency range.
24. The system of claim 19 , wherein the sensor array comprises a plurality of antenna panels, the transmitting sensor elements are proximal when the transmitting sensor elements reside on a first common antenna panel and the receiving sensor elements are proximal when the receiving sensor elements reside on a second common antenna panel.
25. The system of claim 19 , the operations further comprising classifying the object within the scene based on the parameter.
26. The system of claim 19 , wherein the parameter characterizes a surface of the object within the scene.
27. The system of claim 19 , wherein the parameter is a statistic of the intermediate scattering coefficients.
28. The system of claim 19 , wherein estimating the parameter uses a labeled dataset of intermediate scattering coefficients.
29. The system of claim 19 , wherein the parameter includes a class of the object within the scene.
30. The system of claim 19 , wherein the scattering coefficients are not computed for individual voxels.
31. The system of claim 19 , wherein the plurality of sensors elements reside on a plurality of panels driven by a frequency modulated continuous wave signal, the plurality of panels being independently moveable.
32. The system of claim 19 , wherein the mapping of the plurality of scene sub-domains to the plurality of sensor elements includes a block diagonal transfer matrix.
33. The system of claim 19 , wherein a separation between one of the plurality of measurement sub-domains and the plurality of scene sub-domains is greater than a separation between a voxel and an associated scene sub-domain center and greater than a separation between one of the plurality of sensor elements and an associated measurement sub-domain center.
34. The system of claim 19 , wherein applying the mapping to the received data includes computing a match-filter solution using the mapping.
35. The system of claim 19 , wherein applying the mapping to the received data includes computing, using the mapping, a least squares solution or a compressed sensing solution.