Increasing performance of a receive pipeline of a radar with memory optimization

Patent metadata

Field	Value
Publication number	US-11454697-B2
Application number	US-202017120424-A
Country	US
Kind code	B2
Filing date	Dec 14, 2020
Priority date	Feb 10, 2017
Publication date	Sep 27, 2022
Grant date	Sep 27, 2022

Abstract

A radar includes transmitters, receivers, a memory, and a processor. The transmitters transmit radio signals, and the receivers receive reflected radio signals. The processor produces samples by correlating reflected radio signals with time-delayed replicas of transmitted radio signals. The processor stores this information as a first data structure, with information related to signals reflected from objects as a function of time (one dimension of the data structure) at various distances (a second dimension of the data structure) for various receivers (a third dimension of the data structure). The first data structure is processed to compute velocity and angle estimates, which are stored in second and third data structures, respectively. One or more memory optimizations are used to increase performance. Before storing the second and third data structures in a memory, the second and third data structures are sparsified to only include the outputs in specific regions of interest.

First claim

The invention claimed is:

1. A radar sensing system comprising:
a transmit pipeline comprising a plurality of transmitters configured to transmit radio signals;
a receive pipeline comprising a plurality of receivers configured to receive radio signals that include the transmitted radio signals transmitted by the transmitters and reflected from objects in an environment;
a sparsifier processor; and
a memory;
wherein the receive pipeline is configured to correlate the received radio signals with a plurality of time-delayed replicas of the transmitted radio signals to produce first samples stored in the memory, wherein the first samples are arranged in the memory as a first data structure that comprises an array of ranges of different objects in the environment as measured by different receivers;
wherein the receive pipeline is configured to process the first data structure to estimate Doppler/velocity of objects in the environment for given ranges and given receivers to produce a second data structure that comprises an array of Doppler/velocity estimates of the different objects as defined by their corresponding given ranges for the given receivers from the first data structure;
wherein the receive pipeline is configured to output the second data structure to the sparsifier processor; and
wherein the sparsifier processor is configured to analyze the second data structure to compute a compressed version of the second data structure that comprises selected portions of data contained in the second data structure, and wherein the compressed version of the second data structure is stored in the memory for further processing.

2. The radar sensing system of claim 1 , wherein the receive pipeline is configured to process the second data structure to estimate angles of objects in the environment for given ranges and Doppler/velocity estimates to produce a third data structure that comprises an array of angle estimates of the different objects as defined by their corresponding Doppler/velocity estimates and given ranges from the second data structure.

3. The radar sensing system of claim 2 , wherein the sparsifier processor is configured to analyze the third data structure to compute a compressed version of the third data structure that comprises selected portions of data contained in the third data structure, and wherein the compressed version of the third data structure is stored in the memory for further processing.

4. The radar sensing system of claim 3 , wherein samples within the selected portions of data have similar Doppler estimates, ranges and/or angles.

5. The radar sensing system of claim 3 , wherein the selected portions of data correspond to certain regions of interest in the second data structure and the third data structure.

6. The radar sensing system of claim 3 , wherein the receive pipeline comprises the sparsifier processor.

7. The radar sensing system of claim 3 , wherein the sparsifier processor is configured to analyze the third data structure to: 1) identify voxels of interest corresponding to targets of low absolute velocity by identifying those voxels with Doppler/velocity estimates below a threshold, 2) tag or identify a plurality of regions, including the identified voxels of interest, as belong to a static image, and 3) marking the identified voxels of interest as belonging to the static image.

8. The radar sensing system of claim 7 , wherein the sparsifier processor is configured to output the static image to an external memory by specifying a first static voxel to be captured for each angle slice, along with a window width to be captured.

9. The radar sensing system of claim 1 , wherein the sparsifier processor is configured to compute parity checks on data either before or after the sparsification operation, and wherein the sparsifier processor is configured to compute parity checks by checking bit/byte values based upon a remainder found from dividing data mapped into a polynomial with a known code generator polynomial.

10. The radar sensing system of claim 1 , wherein the first data structure is a first three-dimensional data structure corresponding to three dimensions including in one dimension the ranges of the different objects, wherein the second data structure is a second three-dimensional data structure corresponding to three dimensions including in one dimension the Doppler/velocity estimates of the different objects, and wherein the third data structure is a third three-dimensional data structure corresponding to three dimensions including in one dimension the angle estimates of the different objects.

11. The radar sensing system of claim 1 , wherein the plurality of transmitters is configured for installation and use on a vehicle, and wherein the plurality of receivers is are configured for installation and use on the vehicle.

12. The radar sensing system of claim 1 , wherein the receive pipeline is configured to process the second data structure to estimate angles of objects in the environment for given ranges and for given Doppler shifts.

13. A method for optimizing memory performance and access of a data structure in a receive pipeline of a radar sensing system, the method comprising:
transmitting, with a plurality of transmitters, radio signals;
receiving, with a plurality of receivers, radio signals that include the transmitted radio signals transmitted by the transmitters and reflected from objects in an environment;
correlating the received radio signals with a plurality of time-delayed replicas of the transmitted radio signals to produce first samples stored in a memory as a first data structure that comprises an array of ranges of different objects in the environment as measured by different receivers;
processing the first data structure to estimate Doppler shifts at given ranges and for given receivers to produce a second data structure that comprises an array of Doppler estimates of the different objects as defined by their corresponding given ranges for the given receivers from the first data structure; and
analyzing, with a sparsifier processor, the second data structure to compute a compressed version of the second data structure that comprises selected portions of data contained in the second data structure, and wherein the compressed version of the second data structure is stored in the memory for further processing.

14. The method of claim 13 further comprising computing one or more common block exponents for the first data structure to reduce a bit-width required to represent complex numbers in the samples of the first data structure.

15. The method of claim 14 , wherein the one or more common block exponents are computed on-the-fly without analyzing all samples in the block.

16. The method of claim 14 , wherein computing the block exponent is performed for multiple blocks of the first data structure.

17. The method of claim 14 , wherein samples associated with each range-bin are chosen as separate blocks.

18. The method of claim 13 further comprising processing the second data structure to estimate angles of objects in the environment for given ranges and Doppler estimates to produce a third data structure that comprises an array of angle estimates of the different objects as defined by their corresponding Doppler estimates and given ranges from the second data structure.

19. The method of claim 18 further comprising analyzing, with the sparsifier processor, the third data structure to compute a compressed version of the third data structure that comprises selected portions of data con

Assignees

• Uhnder Inc

Inventors

• Maher Monier
• Bordes Jean Pierre
• Davis Curtis

Classifications

• G01S7/4021

  of receivers · CPC title

• G06F3/0661

  Format or protocol conversion arrangements · CPC title

• G01S7/032 Primary

  Constructional details for solid-state radar subsystems · CPC title

• G01S13/42

  Simultaneous measurement of distance and other co-ordinates (indirect measurement G01S13/46) · CPC title

• G01S13/58

  Velocity or trajectory determination systems; Sense-of-movement determination systems · CPC title