Granted Patent B2

US 12,541,015 · App. 17/474,902 · Granted Feb 3, 2026

Inventor: Theagenis J. Abatzoglou (Dana Point, CA)

Assignee: Deer & Company

G01S13/06G01S13/582

Abstract

Claims (81)

1 . A method for processing radar signals, comprising:

obtaining a first plurality of samples of a radar signal, by:

for each of a plurality of receiver antennas of a sparse multiple-input and multiple-output (MIMO) receiver array, collecting a plurality of samples of a radar signal, wherein the radar signal comprises a plurality of mutually orthogonal radar sub-signals that are each generated by one of a plurality of transmitter antennas of a sparse MIMO transmitter array;

selecting a transmitter-receiver antenna pair by:

determining, for each virtual antenna associated with a potential transmitter-receiver antenna pair, a signal-to-noise ratio associated with the virtual antenna relative to an object; and

selecting the transmitter-receiver pair associated with the virtual antenna having the highest associated signal-to-noise ratio relative to the object; and

generating the first plurality of samples of the radar signal by processing the plurality of samples of the virtual antenna corresponding to the selected transmitter-receiver antenna pair, wherein:

the radar signal comprises a plurality of radar signal pulses; and

the first plurality of samples of the radar signal comprises a plurality of samples of each of the plurality of radar signal pulses; and

processing the first plurality of samples of the radar signal to estimate a motion parameter of an object comprising values for a range, a velocity, and an acceleration of the object by:

generating a second plurality of samples of the radar signal by calculating the Fourier transforms of each of the plurality of radar signal pulses comprising the first plurality of samples of the radar signal;

processing the second plurality of samples of the radar signal to generate a coarse search estimate of the motion parameter of the object; and

further processing the second plurality of samples of the radar signal to generate a fine search estimate of the motion parameter of the object using:

the coarse search estimate of the motion parameter of the object as a starting point.

2 . The method of claim 1 , wherein generating the coarse search estimate of the motion parameter of the object comprises:

generating a plurality of potential motion parameters of the object, wherein:

each of the plurality of potential motion parameters comprises potential values for the range, the velocity, and the acceleration of the object;

the potential values for the range, the velocity, and the acceleration are each within an allowed range interval, an allowed velocity interval, and an allowed acceleration interval, respectively; and

the plurality of potential motion parameters are distributed such that at least one of the plurality of potential motion parameters is within a region of convergence of Newton's method for any combination of values within the allowed range, the allowed velocity, and the allowed acceleration intervals;

calculating a value of a likelihood function for each of the generated plurality of potential motion parameters, wherein the likelihood function is for the range, the velocity, and the acceleration of the object for the generated second plurality of samples of the radar signal; and

selecting from the plurality of generated potential motion parameters a potential motion parameter that yielded a highest value of the likelihood function.

3 . The method of claim 2 , wherein generating the plurality of potential motion parameters of the object comprises:

determining, based on the second plurality of samples of the radar signal, a lattice comprised of parallelepipeds, wherein:

the vertices of each of the parallelepipeds correspond to a particular value for the range, the velocity, and the acceleration of the object; and

the vertices of each of the parallelepipeds are spaced such that at any point within a parallelepiped, at least one of the vertices of that parallelepiped is within the region of convergence of the respective point; and

enumerating the vertices of the parallelepipeds comprising the lattice whose values for range, velocity, and acceleration are within the allowed position range, the allowed velocity range, and the allowed acceleration range, respectively.

4 . The method of claim 1 , wherein:

the first plurality of samples of the radar signal corresponds to a radar system on a vehicle;

the fine search estimate of the motion parameter of the object is generated within 100 milliseconds of the radar signal being transmitted by the radar system; and

the fine search estimate of the motion parameter of the object are within 10 percent of the true value of the motion parameter of the object.

5 . A system for estimating a motion parameter of an object using a radar system, comprising:

a transmitter configured to generate and transmit a radar signal, wherein the transmitter comprises a sparse multiple-input and multiple-output (MIMO) transmitter array that comprises a plurality of transmitter antennas, wherein the radar signal includes mutually orthogonal radar sub-signals that are each generated by one of the plurality of transmitter antennas;

a receiver configured to obtain a first plurality of samples of reflections of the transmitted radar signal, and comprising a sparse MIMO receiver array that comprises a plurality of receiver antennas; and

at least one processor configured to:

for each of the plurality of receiver antennas, collect the respective plurality of samples;

select a transmitter-receiver antenna pair by:

determining, for each virtual antenna associated with a potential transmitter-receiver antenna pair, a signal-to-noise ratio associated with the virtual antenna relative to an object; and

selecting the transmitter-receiver pair associated with the virtual antenna having the highest associated signal-to-noise ratio relative to the object;

generate a first plurality of samples of reflections of the transmitted radar signal by processing the plurality of samples of the virtual antenna corresponding to the selected transmitter-receiver antenna pair;

generate a second plurality of samples of reflections of the transmitted radar signal by calculating the Fourier transforms of the first plurality of samples of reflections of the transmitted radar signal;

process the second plurality of samples to generate a coarse search estimate of the motion parameter of the object, wherein the motion parameter of the object comprises values for a range, a velocity, and an acceleration of the object; and

further process the second plurality of samples to generate a fine search estimate of the range, the velocity, and the acceleration of the object using:

the coarse search estimate of the range, the velocity, and the acceleration of the object as a starting point.

6 . The system of claim 5 , wherein generating the coarse search estimate of the motion parameter of the object comprises:

generating a plurality of potential motion parameters of the object, wherein:

each of the plurality of potential motion parameters comprises potential values for the range, the velocity, and the acceleration of the object;

the potential values for the range, the velocity, and the acceleration are each within an allowed range interval, an allowed velocity interval, and an allowed acceleration interval, respectively; and

the plurality of potential motion parameters are distributed such that at least one of the plurality of potential motion parameters is within a region of convergence of Newton's method for any combination of values within the allowed range, the allowed velocity, and the allowed acceleration intervals; and

selecting a potential motion parameter from the plurality of motion parameters that yields a highest value for a likelihood function, wherein the likelihood function is for the range, the velocity, and the acceleration of the object for the generated second plurality of samples of reflections of the transmitted radar signal.

7 . The system of claim 6 , wherein generating the plurality of potential motion parameters of the object comprises:

determining, based on the second plurality of samples of reflections of the transmitted radar signal, a lattice comprised of parallelepipeds, wherein:

the vertices of each of the parallelepipeds correspond to a particular value for the range, the velocity, and the acceleration of the object; and

the vertices of each of the parallelepipeds are spaced such that at any point within a parallelepiped, at least one of the vertices of that parallelepiped is within the region of convergence of the respective point; and

enumerating the vertices of the parallelepipeds comprising the lattice whose values for range, velocity, and acceleration are within the allowed position range, the allowed velocity range, and the allowed acceleration range, respectively.

8 . The system of claim 5 , wherein:

the radar system is a component of a vehicle;

the transmitter and the receiver are a component of the radar system;

the fine search estimate of the motion parameter of the object is generated within 100 milliseconds of the radar signal being transmitted by the transmitter; and

the fine search estimate of the motion parameter of the object are within 10 percent of the true value of the motion parameter of the object.

9 . A non-transitory computer readable medium comprising instructions that, when executed by one or more processors, cause the one or more processors to process radar signals to estimate a motion parameter of an object by:

obtaining a first plurality of samples of a radar signal by:

for each of a plurality of receiver antennas of a sparse multiple-input and multiple-output (MIMO) receiver array, collecting a plurality of samples of a radar signal, wherein the radar signal comprises a plurality of mutually orthogonal radar sub-signals that are each generated by one of a plurality of transmitter antennas of a sparse MIMO transmitter array;

selecting a transmitter-receiver antenna pair by:

determining, for each virtual antenna associated with a potential transmitter-receiver antenna pair, a signal-to-noise ratio associated with the virtual antenna relative to an object; and

selecting the transmitter-receiver pair associated with the virtual antenna having the highest associated signal-to-noise ratio relative to the object; and

generating the first plurality of samples of the radar signal by processing the plurality of samples of the virtual antenna corresponding to the selected transmitter-receiver antenna pair, wherein:

generating a second plurality of samples of the radar signal by calculating the Fourier transforms of the first plurality of samples of the radar signal;

processing the second plurality of samples of the radar signal to generate a coarse search estimate of the motion parameter of the object, wherein the motion parameter of the object comprises values for a range, a velocity, and an acceleration of the object; and

further processing the second plurality of samples of the radar signal to generate a fine search estimate of the range, the velocity, and the acceleration of the object using:

the coarse search estimate of the range, the velocity, and the acceleration of the object as a starting point.

10 . The non-transitory computer readable medium of claim 9 , wherein generating the coarse search estimate of the motion parameter of the object comprises:

generating a plurality of potential motion parameters of the object, wherein:

each of the plurality of potential motion parameters comprises potential values for the range, the velocity, and the acceleration of the object;

the potential values for the range, the velocity, and the acceleration are each within an allowed range interval, an allowed velocity interval, and an allowed acceleration interval, respectively; and

the plurality of potential motion parameters are distributed such that at least one of the plurality of potential motion parameters is within a region of convergence of Newton's method for any combination of values within the allowed range, the allowed velocity, and the allowed acceleration intervals; and

selecting a potential motion parameter from the plurality of motion parameters that yields a highest value for a likelihood function, wherein the likelihood function is for the range, the velocity, and the acceleration of the object for the generated second plurality of samples of the radar signal.

11 . The non-transitory computer readable medium of claim 10 , wherein generating the plurality of potential motion parameters of the object comprises:

determining, based on the second plurality of samples of the radar signal, a lattice comprised of parallelepipeds, wherein:

the vertices of each of the parallelepipeds correspond to a particular value for the range, the velocity, and the acceleration of the object; and

the vertices of each of the parallelepipeds are spaced such that at any point within a parallelepiped, at least one of the vertices of that parallelepiped is within the region of convergence of the respective point; and

enumerating the vertices of the parallelepipeds comprising the lattice whose values for range, velocity, and acceleration are within the allowed position range, allowed velocity range, and allowed acceleration range, respectively.

Assignments (4)

Continuity (1)

References Cited (25)