Linear prediction-based bistatic detector for automotive radar
The disclosure provides systems, apparatuses, and techniques for operating automotive MIMO radars in crowded multi-path environments to obtain reliable detections by linearly predicting whether a bistatic condition occurred. To avoid saturating computing resources processing bistatic detections, the described techniques enable a radar system to quickly identify and discard from the field-of-view radar detections that are likely a result of bistatic conditions. By ignoring unusable radar returns that are likely a result of bistatic conditions, an example radar system can focus on processing radar returns from static conditions, for example, in providing radar-based detections as output to an automotive system that is driving a vehicle in an autonomous or a semi-autonomous mode. In so doing, the example radar system provides a highly accurate static object detector that is sufficiently quick in detecting bistatic conditions for use in vehicle-safety systems as well as autonomous and semi-autonomous control.
1. A method, comprising:
obtaining, using a radar system installed on a vehicle, radar returns reflected off one or more objects in an environment of the vehicle;
determining, by the radar system, based on the radar returns, a linear prediction as to whether a bistatic condition exists with the one or more objects in the environment from using a detector γ, wherein:
the detector γ is equal to
P
F
⊥
g
2
2
M
-
P
≷
H
1
H
0
ρ
,
H 0 is a hypothesis that the bistatic condition did not happen, H 1 is a hypothesis that the bistatic condition did happen, the radar returns are sampled at time t and denoted as x 1 (t), x 2 (t), . . . , x M (t), M is a quantity of elements of a MIMO array of the radar system, P is greater than zero and less than M and comprises a set of complex coefficients {c 1 , c 2 , . . . , c p }, g equals F multiplied by c, g equals
[
x
P
+
1
x
P
+
2
⋮
x
M
]
,
F equals
[
x
P
x
P
-
1
…
x
1
x
P
+
1
x
P
…
x
2
⋮
⋮
⋱
⋮
x
M
-
1
x
M
-
2
…
x
M
-
P
]
,
and c equals
[
c
1
c
2
⋮
c
P
]
;
and
responsive to determining that the linear prediction from using the detector γ indicates a bistatic condition does not exist with the one or more objects in the environment:
refraining from discarding the radar returns; and
outputting a range and range rate associated with the one or more objects for localizing the vehicle in the environment relative to the one or more objects.
2. The method of claim 1 , wherein the radar returns comprise first radar returns reflected off one or more first objects in the environment of the vehicle, the method further comprising:
obtaining, using the radar system, second radar returns reflected off one or more second objects in the environment of the vehicle;
determining, by the radar system, based on the second radar returns, another linear prediction as to whether a bistatic condition exists with the one or more second objects in the environment; and
responsive to determining that the other linear prediction indicates a bistatic condition exists with the one or more second objects in the environment:
discarding the second radar returns; and
refraining from outputting another range and range rate associated with the one or more second objects to prevent the vehicle from localizing in the environment relative to the one or more second objects.
3. The method of claim 2 , wherein further responsive to determining that the other linear prediction indicates a bistatic condition exists with the one or more objects in the environment, the method further comprises:
immediately obtaining, using the radar system, third radar returns reflected off one or more third objects in the environment of the vehicle;
determining, based on the third radar returns, a third linear prediction as to whether a bistatic condition exists with the one or more third objects in the environment; and
responsive to determining that the third linear prediction indicates a bistatic condition does not exist with the one or more third objects in the environment:
refraining from discarding the third radar returns; and
outputting, to a control interface of the vehicle, a third range and range rate associated with the one or more third objects for localizing the vehicle in the environment relative to the one or more third objects and to prevent the vehicle from localizing in the environment relative to the one or more second objects.
4. The method of claim 2 , further comprising:
setting, in a memory of the radar system, a parameter to indicate the linear prediction; and
refraining from discarding the first radar returns in response to setting the parameter to indicate the linear prediction indicates a bistatic condition does not exist with the one or more objects in the environment.
5. The method of claim 4 , further comprising:
setting, in the memory, another parameter to indicate the other linear prediction; and
discarding the second radar returns in response to setting the other parameter to indicate the other linear prediction indicates a bistatic condition exists with the one or more objects in the environment.
6. The method of claim 2 , wherein discarding the second radar returns comprises:
clearing a portion of a synthetic array comprising the second radar returns.
7. The method of claim 6 , wherein clearing a portion of the synthetic array comprises clearing all of the synthetic array.
8. The method of claim 2 , wherein discarding the second radar returns comprises:
ignoring the second radar returns for radar range and range rate calculations performed by the radar system.
9. The method of claim 1 , wherein the one or more objects comprise one or more static objects with a velocity that is less than a movement threshold.
10. The method of claim 1 , further comprising:
determining the linear prediction as to whether a bistatic condition exists with the one or more objects in the environment based on a detector γ′, wherein:
γ′ is equal to
P
F
⊥
g
2
2
≷
H
1
H
0
ρ
(
M
-
P
)
.
11. The method of claim 10 , further comprising:
determining the linear prediction as to whether a bistatic condition exists with the one or more objects in the environment based on a detector γ″, wherein:
γ″ is equal to
P
F
g
2
g
2
≶
H
1
H
0
S
N
R
S
N
R
+
1
,
and
SNR is equal to signal-to-noise ratio of the radar system.
12. A system, comprising:
a radar system for installation on a vehicle, the radar system having at least one processor configured to:
obtain radar returns reflected off one or more objects in an environment of the vehicle;
determine, based on the radar returns, a linear prediction as to whether a bistatic condition exists with the one or more objects in the environment from using a detector γ, wherein:
the detector γ is equal to
P
F
⊥
g
2
2
M
-
P
≷
H
1
H
0
ρ
,
H 0 is a hypothesis that the bistatic condition did not happen, H 1 is a hypothesis that the bistatic condition did happen, the radar returns are sampled at time t and denoted as x 1 (t), x 2 (t), . . . , x M (t), M is a quantity of elements of a MIMO array of the radar system, P is greater than zero and less than M and comprises a set of complex coefficients {c 1 , c 2 , . . . , c P }, g equals F multiplied by c, g equals
[
X
P
+
1
X
P
+
2
⋮
X
M
]
,
F equals
[
X
P
X
P
-
1
⋯
X
1
X
P
+
1
X
P
⋯
X
2
⋮
⋮
⋱
⋮
X
M
-
1
X
M
-
2
⋯
X
M
-
P
]
,
and c equals
[
c
1
c
2
⋮
c
P
]
;
and
responsive to determining that the linear prediction from using the detector γ indicates a bistatic condition does not exist with the one or more objects in the environment:
refrain from discarding the radar returns; and
output a range and range rate associated with the one or more objects for localizing the vehicle in the environment relative to the one or more objects.
13. The system of claim 12 , further comprising:
an automotive system for controlling the vehicle, the automotive system configured to operate the vehicle in an autonomous or semi-autonomous mode based in part on the range and range rate.
14. The system of claim 12 , further comprising a vehicle that includes the radar system.
15. The system of claim 14 , wherein the at least one processor is configured to obtain the radar returns by forming a synthetic array based on multiple-input-multiple-output channels of a phased antenna array.
16. A non-transitory computer-readable storage medium comprising instructions that, when executed, configure a processor of a vehicle to:
obtain radar returns reflected off one or more objects in an environment of the vehicle;
determine, based on the radar returns, a linear prediction as to whether a bistatic condition exists with the one or more objects in the environment from using a detector γ, wherein:
the detector γ is equal to
P
F
⊥
g
2
2
M
-
P
≷
H
1
H
0
ρ
,
H 0 is a hypothesis that the bistatic condition did not happen, H 1 is a hypothesis that the bistatic condition did happen, the radar returns are sampled at time t and denoted as x 1 (t), x 2 (t), . . . , x M (t), M is a quantity of elements of a MIMO array of the radar system, P is greater than zero and less than M and comprises a set of complex coefficients {c 1 , c 2 , . . . , c P }, g equals F multiplied by c, g equals
[
X
P
+
1
X
P
+
2
⋮
X
M
]
,
F equals
[
X
P
X
P
-
1
⋯
X
1
X
P
+
1
X
P
⋯
X
2
⋮
⋮
⋱
⋮
X
M
-
1
X
M
-
2
⋯
X
M
-
P
]
,
and c equals
[
c
1
c
2
⋮
c
P
]
;
and
responsive to determining that the linear prediction from using the detector γ indicates a bistatic condition does not exist with the one or more objects in the environment:
refrain from discarding the radar returns; and
output a range and range rate associated with the one or more objects for localizing the vehicle in the environment relative to the one or more objects.
17. The computer-readable storage medium of claim 16 , wherein the radar returns comprise first radar returns reflected off one or more first objects in the environment of the vehicle, the instructions, when executed, further configure the processor to:
obtain second radar returns reflected off one or more second objects in the environment of the vehicle;
determine, based on the second radar returns, another linear prediction as to whether a bistatic condition exists with the one or more second objects in the environment; and
responsive to determining that the other linear prediction indicates a bistatic condition exists with the one or more second objects in the environment:
discard the second radar returns; and
refrain from outputting another range and range rate associated with the one or more second objects to prevent the vehicle from localizing in the environment relative to the one or more second objects.
18. The computer-readable storage medium of claim 17 , wherein the instructions, when executed, further configure the processor to:
further responsive to determining that the other linear prediction indicates a bistatic condition exists with the one or more objects in the environment:
immediately obtain third radar returns reflected off one or more third objects in the environment of the vehicle;
determine, based on the third radar returns, a third linear prediction as to whether a bistatic condition exists with the one or more third objects in the environment; and
responsive to determining that the third linear prediction indicates a bistatic condition does not exist with the one or more third objects in the environment:
refrain from discarding the third radar returns; and
output, to a control interface of the vehicle, a third range and range rate associated with the one or more third objects for localizing the vehicle in the environment relative to the one or more third objects and to prevent the vehicle from localizing in the environment relative to the one or more second objects.
19. The computer-readable storage medium of claim 17 , wherein the instructions, when executed, further configure the processor to:
set a parameter to indicate the linear prediction indicates a bistatic condition does not exist with the one or more objects in the environment; and
refrain from discarding the first radar returns in response to setting the parameter to indicate the linear prediction indicates a bistatic condition does not exist with the one or more objects in the environment.