Vehicle radar sensor utilizing non-uniform frequency modulated continuous wave (FMCW) chirps
AssigneeAurora Operations, Inc.
InventorsChunshu Li, Stephen Crouch
Life of this patent
Estimating the term…
Abstract
A vehicle radar sensor utilizes Frequency Modulated Continuous Wave (FMCW) radar signals that incorporate non-uniform FMCW chirps having chirp profiles that differ from one another to sense one or more parameters of one or more objects in a field of view of the radar sensor. The chirp profiles may differ from one another in various manners, e.g., based on starting frequency, repetition interval, duration and/or slope, and among other advantages, may be used to enhance sensing of various parameters such as range, Doppler/velocity and/or angle.
Claims20
A radar sensor for a vehicle, comprising:
- a radar transmitter configured to transmit a first radar signal, the first radar signal including a frame associated with a plurality of frequency modulated continuous wave (FMCW) chirps, wherein the frame is a coherent processing interval, and wherein the plurality of FMCW chirps includes a plurality of uniform FMCW chirps having a particular chirp profile and a plurality of non-uniform FMCW chirps having chirp profiles that differ from one another and from the particular chirp profile;
- a radar receiver configured to receive a second radar signal that is a reflected signal of the first radar signal;
- and
- control logic coupled to the radar receiver and configured to process the second radar signal, based on the uniform and non-uniform FMCW chirps in the frame, to sense one or more parameters of an object in a field of view of the radar transmitter;
- wherein the control logic uses the uniform and non-uniform FMCW chirps to sense the one or more parameters of the object by:
- generating a uniform data cube by performing a range transformation with fast-time samples of the second radar signal and using the uniform FMCW chirps;
- generating a non-uniform data cube by performing a range transformation with fast-time samples of the second radar signal and using the non-uniform FMCW chirps, wherein the FMCW chirps used to generate the non-uniform data cube are different from the FMCW chirps used to generate the uniform data cube;
- detecting a plurality of candidate objects using the uniform data cube;
- determining a first parameter for the plurality of candidate objects using the uniform data cube and without using the non-uniform data cube;
- and
- enhancing the first parameter for the plurality of candidate objects using the non-uniform data cube.
A method of operating a radar sensor for a vehicle, comprising:
- transmitting a first radar signal, the first radar signal including a frame associated with a plurality of frequency modulated continuous wave (FMCW) chirps, wherein the frame is a coherent processing interval, and wherein the plurality of FMCW chirps includes a plurality of uniform FMCW chirps having a particular chirp profile and a plurality of non-uniform FMCW chirps having chirp profiles that differ from one another and from the particular chirp profile;
- receiving a second radar signal that is a reflected signal of the first radar signal;
- and
- processing the second radar signal, based on the uniform and non-uniform FMCW chirps in the frame, to sense one or more parameters of an object in a field of view of the radar sensor, including:
- generating a uniform data cube by performing a range transformation with fast-time samples of the second radar signal and using the uniform FMCW chirps;
- generating a non-uniform data cube by performing a range transformation with fast-time samples of the second radar signal and using the non-uniform FMCW chirps, wherein the FMCW chirps used to generate the non-uniform data cube are different from the FMCW chirps used to generate the uniform data cube;
- detecting a plurality of candidate objects using the uniform data cube;
- determining a first parameter for the plurality of candidate objects using the uniform data cube and without using the non-uniform data cube;
- and
- enhancing the first parameter for the plurality of candidate objects using the non-uniform data cube.
An autonomous vehicle control system, comprising:
- a radar transmitter configured to transmit a first radar signal, the first radar signal including a frame associated with a plurality of frequency modulated continuous wave (FMCW) chirps, wherein the frame is a coherent processing interval, and wherein the plurality of FMCW chirps includes a plurality of uniform FMCW chirps having a particular chirp profile and a plurality of non-uniform FMCW chirps having chirp profiles that differ from one another and from the particular chirp profile;
- a radar receiver configured to receive a second radar signal that is a reflected signal of the first radar signal;
- and
- control logic coupled to the radar receiver and configured to process the second radar signal, based on the uniform and non-uniform FMCW chirps in the frame, to sense one or more parameters of an object in a field of view of the radar transmitter;
- wherein the control logic uses the uniform and non-uniform FMCW chirps to sense the one or more parameters of the object by:
- generating a uniform data cube by performing a range transformation with fast-time samples of the second radar signal and using the uniform FMCW chirps;
- generating a non-uniform data cube by performing a range transformation with fast-time samples of the second radar signal and using the non-uniform FMCW chirps, wherein the FMCW chirps used to generate the non-uniform data cube are different from the FMCW chirps used to generate the uniform data cube;
- with the uniform data cube, performing a Doppler Fast Fourier Transform (FFT) transformation, performing a beamforming FFT transformation, and detecting a plurality of candidate objects, wherein performing the Doppler FFT transformation introduces Doppler ambiguities into the uniform data cube and performing the beamforming FFT transformation introduces angle ambiguities into the uniform data cube;
- and
- with the non-uniform data cube, disambiguating the Doppler and angle ambiguities by performing Doppler and beamforming Discrete Fourier Transform (DFT) transformations focused on the plurality of candidate objects.
Dependencies are inferred from the claim text.