Radar authentication method and apparatus, and computer storage medium

1. A radar authentication method, comprising:

obtaining first output data of a to-be-authenticated radar, wherein the first output data is based on a first emitted wave of the to-be-authenticated radar and a first reflected wave received at the to-be-authenticated radar;

obtaining a prediction model through training based on second output data of a target radar, wherein the second output data comprises real data of a plurality of channels of the target radar, wherein the prediction model comprises at least one sub-model, wherein each sub-model corresponds to one of the plurality of channels;

invoking the prediction model to obtain predicted data of the to-be-authenticated radar based on the first output data, wherein invoking the prediction model comprises invoking a target sub-model corresponding to a target channel in the plurality of channels to obtain predicted data of a target channel of the to-be-authenticated radar based on real data of other channels of the to-be-authenticated radar; and

verifying, based on the predicted data of the target channel of the to-be-authenticated radar and real data of the target channel of the to-be-authenticated radar, whether the to-be-authenticated radar and the target radar are the same radar, wherein the to-be-authenticated radar and the target radar are not the same radar when at least one of the following is obtained:

a cosine similarity between the predicted data of the target channel of the to-be-authenticated radar and the real data of the target channel of the to-be-authenticated radar is less than a first threshold, and a mean square error between the predicted data of the target channel of the to-be-authenticated radar and the real data of the target channel of the to-be-authenticated radar is greater than a second threshold; or

a fraction of variance unexplained that is obtained through calculation based on the predicted data of the target channel of the to-be-authenticated radar and the real data of the target channel of the to-be-authenticated radar is greater than a third threshold.

2. The radar authentication method of claim 1 , further comprising:

obtaining the prediction model through training; or

obtaining the prediction model from another device.

3. The radar authentication method of claim 2 , further comprising obtaining the target sub-model corresponding to the target channel in the plurality of channels through training based on real data of other channels of the target radar and real data of the target channel of the target radar, and wherein the other channels comprise at least one of the plurality of channels other than the target channel.

4. The radar authentication method of claim 3 , further comprising obtaining the target sub-model through training based on the real data of the other channels of the target radar and the real data of the target channel of the target radar using a linear regression algorithm.

5. The radar authentication method of claim 4 , further comprising obtaining the target sub-model through training using a linear regression model, wherein the linear regression model is expressed as follows:

y

k

,

i

=

∑

1

⩽

j

⩽

n

c

k

,

j

⁢

x

i

,

j

+

b

k

,

wherein k represents the target channel, wherein y k,i represents real data i of the target channel k of the target radar, wherein n represents a quantity of the other channels, wherein j represents a channel j among the other channels, wherein c k,j represents a coefficient j of the target sub-model, wherein x i,j represents real data i of the channel j of the target radar, wherein b k represents an increment of the target sub-model, and wherein k, i, j, and n are all positive integers.

6. The radar authentication method of claim 1 , wherein verifying whether the to-be-authenticated radar and the target radar are the same radar comprises further verifying, based on predicted data of a plurality of target channels of the to-be-authenticated radar and real data of the plurality of target channels of the to-be-authenticated radar, whether the to-be-authenticated radar and the target radar are the same radar.

7. The radar authentication method of claim 1 , wherein the target radar is a millimeter wave radar, and wherein the second output data comprises a Doppler frequency.

8. A computer device comprising:

one or more processors; and

a memory coupled to the one or more processors and configured to store program instructions and that when invoked by the one or more processors cause the computer device to:

obtain first output data of a to-be-authenticated radar, wherein the first output data is based on a first emitted wave of the to-be-authenticated radar and a first reflected wave received at the to-be-authenticated radar;

obtain a prediction model through training based on second output data of a target radar, wherein the second output data comprises real data of a plurality of channels of the target radar, wherein the prediction model comprises at least one sub-model, wherein each sub-model corresponds to one of the plurality of channels;

invoke the prediction model to obtain predicted data of the to-be-authenticated radar based on the first output data, wherein invoking the prediction model comprises invoking a target sub-model corresponding to a target channel in the plurality of channels to obtain predicted data of a target channel of the to-be-authenticated radar based on real data of other channels of the to-be-authenticated radar; and

verify, based on the predicted data of the target channel of the to-be-authenticated radar and real data of the target channel of the to-be-authenticated radar, whether the to-be-authenticated radar and the target radar are the same radar, wherein the to-be-authenticated radar and the target radar are not the same radar when at least one of the following is obtained:

a cosine similarity between the predicted data of the target channel of the to-be-authenticated radar and the real data of the target channel of the to-be-authenticated radar is less than a first threshold, and a mean square error between the predicted data of the target channel of the to-be-authenticated radar and the real data of the target channel of the to-be-authenticated radar is greater than a second threshold; or

a fraction of variance unexplained that is obtained through calculation based on the predicted data of the target channel of the to-be-authenticated radar and the real data of the target channel of the to-be-authenticated radar is greater than a third threshold.

9. The computer device of claim 8 , wherein the program instructions that when invoked by the one or more processors further cause the computer device to:

obtain the prediction model by the computer device through training; or

obtain the prediction model from a device other than the computer device.

10. The computer device of claim 9 , wherein the program instructions that when invoked by the one or more processors further cause the computer device to obtain the target sub-model corresponding to the target channel in the plurality of channels through training based on real data of other channels of the target radar and real data of the target channel of the target radar, and wherein the other channels of the target radar comprise at least one of the plurality of channels other than the target channel.

11. The computer device of claim 10 , wherein the program instructions that when invoked by the one or more processors further cause the computer device to obtain the target sub-model through training based on the real data of the other channels of the target radar and the real data of the target channel of the target radar using a linear regression algorithm.

12. The computer device of claim 11 , wherein the program instructions that when invoked by the processor further cause the computer device to obtain the target sub-model through training using a linear regression model, wherein the linear regression model is expressed as follows:

y

k

,

i

=

∑

1

⩽

j

⩽

n

c

k

,

j

⁢

x

i

,

j

+

b

k

,

wherein k represents the target channel, wherein y k,i represents real data i of the target channel k of the target radar, wherein n represents a quantity of the other channels of the target radar, wherein j represents a channel j among the other channels of the target radar, wherein c k,j represents a coefficient j of the target sub-model, wherein x i,j represents real data i of the channel j of the target radar, wherein b k represents an increment of the target sub-model, and wherein k, i, j, and n are all positive integers.

13. The computer device of claim 8 , wherein the program instructions that when invoked by the one or more processors further cause the computer device to verify, based on predicted data of a plurality of target channels of the to-be-authenticated radar and real data of the plurality of target channels of the to-be-authenticated radar, whether the to-be-authenticated radar and the target radar are the same radar.

14. The computer device of claim 8 , wherein the target radar is a millimeter wave radar, and the second output data comprises a Doppler frequency.

15. A computer program product comprising computer-executable instructions that are stored on a non-transitory computer-readable medium and that, when executed by one or more processors, causes a computer device to:

obtain first output data of a to-be-authenticated radar, wherein the first output data is based on a first emitted wave of the to-be-authenticated radar and a first reflected wave received at the to-be-authenticated radar;

obtain a prediction model through training based on second output data of a target radar, wherein the second output data of the target radar comprises real data of a plurality of channels of the target radar, wherein the prediction model comprises at least one sub-model, wherein each sub-model corresponds to one of the plurality of channels;

obtain a target sub-model corresponding to a target channel in the plurality of channels through training using a linear regression model, wherein the linear regression model is expressed as follows:

y

k

,

i

=

∑

1

⩽

j

⩽

n

c

k

,

j

⁢

x

i

,

j

+

b

k

,

wherein k represents the target channel, wherein y k,i represents real data i of the target channel k of the target radar, wherein n represents a quantity of other channels of the target radar, wherein j represents a channel j among the other channels of the target radar, wherein c k,i represents a coefficient j of the target sub-model, wherein x i,j represents real data i of the channel i of the target radar, wherein b k represents an increment of the target sub-model, and wherein k, i, j, and n are all positive integers,

invoke the target sub-model corresponding to the target channel in the plurality of channels to obtain predicted data of a target channel of the to-be-authenticated radar based on real data of other channels of the to-be-authenticated radar; and

verify, based on the predicted data of the target channel of the to-be-authenticated radar and real data of the target channel of the to-be-authenticated radar, whether the to-be-authenticated radar and the target radar are the same radar.

16. The computer program product of claim 15 , wherein the computer-executable instructions that when executed by the one or more processors further causes the computer device to further obtain, from a device other than the computer device, the prediction model.

17. The computer program product of claim 15 , wherein the computer-executable instructions that when executed by the one or more processors further causes the computer device to obtain the target sub-model corresponding to the target channel in the plurality of channels through training based on real data of the other channels of the target radar and real data of the target channel of the target radar, and wherein the other channels of the target radar comprise at least one of the plurality of channels other than the target channel.

18. The computer program product of claim 15 , wherein the computer-executable instructions that when executed by the one or more processors further causes the computer device to obtain the target sub-model through training based on the real data of the other channels of the target radar and the real data of the target channel of the target radar using a linear regression algorithm.

19. The computer program product of claim 15 , wherein the computer-executable instructions that when executed by the one or more processors further causes the computer device to verify, based on predicted data of a plurality of target channels of the to-be-authenticated radar and real data of the plurality of target channels of the to-be-authenticated radar, whether the to-be-authenticated radar and the target radar are the same radar.

20. The computer program product of claim 15 , wherein the target radar is a millimeter wave radar, and the second output data comprises a Doppler frequency.