Type-agnostic RF signal representations
This document describes techniques and devices for type-agnostic radio frequency (RF) signal representations. These techniques and devices enable use of multiple different types of radar systems and fields through type-agnostic RF signal representations. By so doing, recognition and application-layer analysis can be independent of various radar parameters that differ between different radar systems and fields.
1. At least one non-transitory computer-readable storage medium having instructions stored thereon that, responsive to execution by at least one computer processor, cause the computer processor to:
receive type-specific raw data representing a reflection signal caused by movement of an object within a type-specific radar field, the reflection signal comprising a superposition of reflections of a plurality of points of the object;
transform the type-specific raw data into a type-agnostic signal representation that is independent of parameters of the type-specific radar field, the transformation according to a model of the object as a set of scattering centers, each of the scattering centers corresponding to one of the points of the object; and
determine, based on the type-agnostic signal representation, a gesture or action performed by the object.
2. The computer-readable storage media of claim 1 , wherein:
the instructions further cause the computer processor to receive a complex signal based on the type-specific raw data, the complex signal having amplitude and phase information from which a phase of the type-specific raw data can be extracted and unwrapped; and
the type-agnostic signal representation is based on the phase of the type-specific raw data.
3. The computer-readable storage media of claim 1 , wherein the type-agnostic signal representation comprises a range-Doppler-time profile, range-time profile, micro-Doppler profile, or fast-time spectrogram for the type-specific raw data.
4. The computer-readable storage media of claim 1 , wherein:
the instructions further cause the computer processor to extract a type-agnostic feature from the type-agnostic signal representation, the type-agnostic feature comprising a signal transformation, engineered feature, computer-vision feature, machine-learned feature, or inferred target feature; and
the determination of the gesture or action performed by the object is based on the type-agnostic feature.
5. The computer-readable storage media of claim 1 , wherein:
the instructions further cause the processor to determine a gesture classification, motion parameter tracking, regression estimate, or gesture probability; and
the determination of the gesture or action performed by the object is based on the gesture classification, motion parameter tracking, regression estimate, or gesture probability.
6. The computer-readable storage media of claim 1 , wherein:
the instructions further cause the processor to:
receive other type-specific raw data representing another reflection signal caused by movement of the object within another type-specific radar field; and
transform the other type-specific raw data into another type-agnostic signal representation; and
the determination of the gesture or action performed by the object is further based on the other type-agnostic signal representation.
7. The computer-readable storage media of claim 1 , wherein the parameters of the type-specific radar field comprise modulation, frequency, amplitude, or phase parameters.
8. The computer-readable storage media of claim 1 , wherein the parameters comprise the type-specific radar field being a single tone, stepped frequency modulated, linear frequency modulated, impulse, or chirped.
9. A computer-implemented method comprising:
receiving type-specific raw data representing a reflection signal caused by movement of an object within a type-specific radar field, the reflection signal comprising a superposition of reflections of a plurality of points of the object;
transforming the type-specific raw data into a type-agnostic signal representation that is independent of parameters of the type-specific radar field, the transformation according to a model of the object as a set of scattering centers, each of the scattering centers corresponding to one of the points of the object;
determining, based on the type-agnostic signal representation, a gesture or action performed by the object; and
passing the determined gesture or action to an application effective to control or alter a display, function, or capability associated with the application.
10. The method of claim 9 , wherein the type-agnostic signal representation is independent of modulation, frequency, amplitude, or phase of the type-specific radar field.
11. The method of claim 9 , wherein the type-agnostic signal representation is independent of the type-specific radar field being single tone, stepped frequency modulated, linear frequency modulated, impulse, or chirped.
12. The method of claim 9 , wherein the type-agnostic signal representation comprises a range-Doppler profile, a range profile, a micro-Doppler profile, or a fast-time spectrogram.
13. The method of claim 9 , further comprising receiving a complex signal based on the type-specific raw data, the complex signal having amplitude and phase information from which a phase of the type-specific raw data can be extracted and unwrapped; and
wherein the type-agnostic signal representation is based on the phase of the type-specific raw data.
14. The method of claim 9 , further comprising determining a gesture classification, motion parameter tracking, regression estimate, or gesture probability; and
wherein the determination of the gesture or action performed by the object is based on the gesture classification, motion parameter tracking, regression estimate, or gesture probability.
15. An apparatus comprising:
at least one computer processor;
a type-specific radar system configured to provide a type-specific radar field, the type-specific radar field provided though a modulation scheme or a type of hardware radar-emitting element, the type-specific radar system comprising:
at least one radar-emitting element configured to provide the type-specific radar field; and
at least one antenna element configured to receive a reflection signal caused by an object moving in the type-specific radar field; and
at least one computer-readable storage medium having instructions stored thereon that, responsive to execution by the computer processor, cause the computer processor to:
receive type-specific raw data representing the reflection signal caused by movement of the object within the type-specific radar field, the reflection signal comprising a superposition of reflections of a plurality of points of the object;
transform the type-specific raw data into a type-agnostic signal representation that is independent of parameters of the type-specific radar field, the transformation according to a model of the object as a set of scattering centers, each of the scattering centers corresponding to one of the points of the object;
determine, based on the type-agnostic signal representation, a gesture or action performed by the object; and
pass the determined gesture or action to an application effective to control or alter a display, function, or capability associated with the application.
16. The apparatus of claim 15 , wherein the type-agnostic signal representation comprises a range-Doppler profile, a range profile, a micro-Doppler profile, or a fast-time spectrogram.
17. The apparatus of claim 15 , wherein:
the instructions further cause the processor to receive a complex signal based on the type-specific raw data, the complex signal having amplitude and phase information from which a phase of the type-specific raw data can be extracted and unwrapped; and
the type-agnostic signal representation is based on the phase of the type-specific raw data.
18. The apparatus of claim 15 , wherein:
the instructions further cause the processor to determine a gesture classification, motion parameter tracking, regression estimate, or gesture probability; and
the determination of the gesture or action performed by the object is based on the gesture classification, motion parameter tracking, regression estimate, or gesture probability.
19. The apparatus of claim 15 , wherein the apparatus is a mobile computing device having the display.
20. The apparatus of claim 19 , wherein the determined gesture or action is a gesture controlling a user interface associated with the application and presented on the display.