A method for testing time-frequency performance of roadside perception system in connected vehicle network

Abstract

The present invention discloses a method for testing the time-frequency performance of a roadside perception system of an Internet of Vehicles, and belongs to the field of Internet of Vehicles. The testing method comprises the following steps: S1: starting the test process and data collection, S2: processing and outputting the test data, and S3: analyzing and evaluating the time-frequency performance. Compared with the prior art, the beneficial effects of the present invention are: based on the true value vehicle data, the system data to be tested, and the ROI data, by selecting the test data with geographical location characteristics and calculating the perception delay, the perception delay of the roadside perception system of the Internet of Vehicles can be evaluated. It is also possible to evaluate whether the system frequency of the perception system is stable by calculating TDEV and MTIE. It comprehensively reflects the time-frequency performance of the roadside perception system.

Description

Technical Field

[0001] The present invention belongs to the field of vehicle networking, and specifically relates to a method for testing the time-frequency performance of a roadside perception system of a vehicle networking. Background Art

[0002] Vehicle-road collaboration uses advanced wireless communications and the new generation of the Internet to implement all-round dynamic real-time information interaction between vehicles and roads, and carries out active vehicle safety control and road collaborative management based on the collection and integration of dynamic traffic information in all time and space, fully realizing the effective collaboration between people, vehicles and roads, ensuring traffic safety, and improving traffic efficiency, thus forming a safe, efficient and environmentally friendly road traffic system.

[0003] Roadside Sensing System (RSS) is an important means to support connected autonomous driving, improve traffic efficiency and alleviate congestion. Providing autonomous vehicles with beyond-visual-range perception, blind spot warning, driving intention and other information through the RSS system is one of the important technical means to make up for the limitations of single-vehicle autonomous driving perception.

[0004] In vehicle-road collaborative applications, the roadside perception system realizes real-time vectorization and tracking of global targets, and its precise perception capability is the key to the roadside perception system.

[0005] Based on this, there are a variety of solutions for evaluating the performance of roadside perception systems in the prior art.

[0006] For example, an existing patent with publication number CN114383649A discloses a roadside perception system testing method based on high-precision positioning, including a test system and a roadside perception system, the test system including a mobile carrier, the test system obtains traffic participant information with the mobile carrier as a reference, and outputs reference state information of the traffic participant information by processing the obtained traffic participant information; the roadside perception system obtains the state information to be tested of the traffic participant; the reference state information and the state information to be tested are compared and analyzed, the error between the reference state information and the state information to be tested is calculated, and a performance evaluation report of the roadside perception system is given according to the calculated error.

[0007] For example, an existing patent with publication number CN112816954A discloses a roadside perception system evaluation method based on true value, which includes the following steps: establishing a true value perception device group, and collecting roadside perception data synchronously with the perception devices of the roadside perception system RSS to be tested in a selected test time interval; processing the original data sent back by the true value perception device group, completing target type identification and target trajectory identification, and completing perception data labeling; generating true values ​​based on the labeled data, and the true value data includes the target type, position, speed, acceleration, and trajectory of traffic participants; in the selected test time interval, comparing the structured perception data output by the RSS to be tested with the true value data, and outputting the statistical evaluation results of the perception performance.

[0008] For example, an existing patent with publication number CN112382079A discloses a roadside perception simulation method and system for vehicle-road collaboration, which provides a virtual environment for simulation testing before actual road testing of roadside perception. With the help of the simulation environment, the relationship between the sensor and the environment before actual installation can be analyzed, and the corresponding working effect can be visualized. In addition, the training data and test data required for machine learning are directly output through the simulation system.

[0009] The above patents all test and evaluate the perception capabilities of overall traffic participants, including traffic targets, traffic events, traffic flows, etc. Although they can test the perception capabilities of the roadside perception system to a certain extent, they cannot evaluate the time-frequency performance of the roadside perception system.

[0010] The time-frequency performance of the roadside perception system not only reflects the stability of the system itself, but is also one of the important factors affecting the actual perception effect. It is particularly important to test and evaluate it. Summary of the invention

[0011] To solve the above problems, the primary purpose of the present invention is to provide a method for testing the time-frequency performance of a roadside perception system of an Internet of Vehicles. The method is based on true vehicle data, system data to be tested, and ROI data. By selecting test data with geographic location characteristics and calculating the perception delay, the method can evaluate the perception delay of the roadside perception system of the Internet of Vehicles.

[0012] Another object of the present invention is to provide a method for testing the time-frequency performance of a roadside perception system of an Internet of Vehicles, which can evaluate whether the system frequency of the perception system is stable by calculating TDEV and MTIE.

[0013] To achieve the above object, the technical solution of the present invention is as follows:

[0014] A method for testing the time-frequency performance of a roadside perception system of an Internet of Vehicles, the method comprising the following steps:

[0015] S1: Start the test process and data collection: determine the ROI range of the detection area, select a true value vehicle to perform driving test in the test area, the true value vehicle records the true value data of the vehicle, and the system to be tested perceives the target of the true value vehicle in the test area and outputs structured perception data;

[0016] S2: Processing and output of test data: The true value car and the system to be tested process the perception data, and output the perception timestamp ts, the MEC sending timestamp tt, and the data of the target object information corresponding to the perception timestamp and the MEC sending timestamp tt. The data output by the true value car is Sgt, and the data output by the system to be tested is Sdut1 and Sdut2. The content of Sdut1 data contains the perception data of the system to be tested on all targets in the perception range, while Sdut2 only contains the data perceived by the true value car;

[0017] S3: Time-frequency performance analysis and evaluation.

[0018] Furthermore, the step S1 includes:

[0019] Step S11: Determine the scope of the test area. Record the longitude and latitude information of each point in the order of "southeast-east-northeast-north-northwest-west-southwest-southeast";

[0020] Step S12: Select a true value vehicle, select a place outside the sensing range of the system to be tested as the starting point of the true value vehicle, and the true value vehicle starts from the starting point to test and record the true value data of the vehicle;

[0021] Step S13: the real value vehicle travels at a normal urban traffic speed, enters along the lane from the entrance direction, and passes through the designated test area;

[0022] Step S14: the true value vehicle leaves the test area and stops collecting true value data of the vehicle;

[0023] Step S15: If the test area contains a turning road, repeat S11-S13 until all lanes are traversed and the route is driven according to the lane rules (i.e., drive straight on a straight road and turn on a turning road. If the lane can drive straight and turn at the same time, drive straight first and then turn in the next test);

[0024] If there are only straight roads in the test area, repeat steps S11-S13 until all lanes are traversed, and then repeat steps S11-S13 at least 2 additional times, and require that in the additional repeated step S12, the real value vehicle travels in a curved "S shape" and crosses at least 2 lanes through the test area; in the present application, the real value vehicle travels in a curved "S shape" and repeats steps S11-S13 to ensure that in the test scenario of a long straight road, data on vehicle turning can be collected to ensure the integrity of the collected data.

[0025] Furthermore, in step S12, the true value vehicle is outside the sensing range of the system to be tested, and is required to be at a distance of more than 400 meters from the location of the system to be tested.

[0026] Furthermore, in step S13, the real value vehicle's travel speed is 20-60 km / h.

[0027] Furthermore, in step S14, after the true value vehicle leaves the test area, it needs to travel an additional distance of at least 150 meters before stopping data collection of the vehicle to ensure the integrity of the true value data.

[0028] Furthermore, in step S2, the perception timestamp is the time when the sensor device collects the target information, wherein the sensor device includes the sensor device of the true value vehicle and the sensor device of the perception system. The MEC sending timestamp tt is: the time when the perception system outputs the perception result after processing the collected target information through the MEC.

[0029] Furthermore, in step S2, the target object information includes but is not limited to the longitude, latitude, altitude, speed, and heading angle of the target object.

[0030] Further, in step S2, Sgt, Sdut1 and Sdut2 output after each test are in a one-to-one correspondence and are defined as a set of test group data.

[0031] Furthermore, the step S3 includes:

[0032] S31: Unify the positioning reference points of the roadside truth system and the system to be tested;

[0033] S32: Check the data of Sgt, Sdut1 and Sdut2, retain the valid data, remove the invalid data, and draw the trajectory diagram of Sgt and Sdut2;

[0034] S33: Selecting data with geographical location characteristics as test group data;

[0035] S34: Calculate the mean △t of the difference between the MEC sending timestamp tt and the sensing timestamp ts of each frame from the data of the system to be tested Sdut2, that is, △t=mean(tt-ts);

[0036] S35: based on the sensing timestamp ts, extract the continuous position information in Sgt and Sdut2, and set them as trajectory data Pgt and Pdut respectively;

[0037] S36: Set the transformed Pdut trajectory to Where R 0 and t 0 are the trajectory rotation and translation matrices respectively;

[0038] S37: Find the transformed Pdut trajectory through KNN search method The Euclidean distance between each point in the Pgt trajectory and the shortest point in the Pgt trajectory is defined as the residual Δ ag ;

[0039] The expression is

[0040] S38: Optimize the residual Δ by LM optimization method ag , to obtain the minimum Δ ag , and calculate the corresponding R * and t * ;

[0041] S39: Based on the result of the previous step, the trajectory of the system to be tested after fitting with the Pgt trajectory is obtained as follows:

[0042] S310: Calculation The time error between the corresponding point and Pgt is The difference between the corresponding perception timestamp ts and the perception timestamp ts corresponding to Pgt, and the time errors of all points are averaged to obtain the average value δt;

[0043] S311: Calculate the perception delay Dsen=△t+δt;

[0044] S312: Calculate TDEV based on the tt timestamp in the Sdut1 data. TDEV is a measure of the change in the signal time interval within a given integration time τ, which can provide information about the signal phase and spectral content. Set the ideal sampling interval of the system to be tested to τ 0 =1 / f obj , where f obj is the default frequency of the system to be tested. When the observation interval is τ=nτ 0 When TDEV is , it can be estimated by the following formula:

[0045]

[0046] Where: ——Round down;

[0047] x i ——instantaneous time error;

[0048] N——total number of samples;

[0049] τ 0 - sampling interval;

[0050] τ——integration time;

[0051] n – the number of sampling intervals within the integration time.

[0052] S313: Calculate MTIE based on the tt timestamp in the Sdut1 data. MTIE is the total length of the measurement period T, for a given observation interval τ = nτ 0 The maximum value of the peak value of the delay time between the timing signal time interval output by the roadside perception system and the ideal time interval represents the stability of the frequency. It can be estimated by the following formula:

[0053]

[0054] Where: x i ——instantaneous time error;

[0055] N——total number of samples;

[0056] τ 0 ——Sampling interval.

[0057] Furthermore, in step S31, the positioning reference point, i.e., the origin of the local tangent plane coordinate system, is used to determine the coordinate position of the true value vehicle.

[0058] Furthermore, in step S32, the judgment condition for valid data is: the trajectories of Sgt and Sdut2 are continuous without obvious interruption, and the time interval of Sgt completely covers the time interval of Sdut1 and Sdut2; otherwise, it is regarded as invalid data. Drawing the trajectory graph of Sgt and Sdut2 can realize the visualization of Sgt and Sdut2 data, making it easy to find whether there is obvious deviation or interruption in the trajectories of the two.

[0059] Furthermore, in step S33, each test group data includes Sgt, Sdut1 and Sdut2 under the same perception timestamp.

[0060] Furthermore, in step S33, the data with geographic location characteristics is data with vehicle turning behavior, including curve test data or "S-shaped" test data, so as to ensure the integrity of the test data.

[0061] Furthermore, in step S36, two points at any position of the Pdut trajectory can be matched by trajectory rotation and translation.

[0062] In step S38, R * and t * are respectively R in step S36 0 and t 0 The corresponding specific value.

[0063] Furthermore, in step S312, the observation interval is τ=nτ0 The smaller the value of TDEV is, the more stable the frequency of the perception system is; in step S313, the observation interval τ = nτ 0 The smaller the MTIE value is, the more stable the frequency of the perception system is.

[0064] Compared with the prior art, the beneficial effects of the present invention are: based on the real-value vehicle data, the system data to be tested, and the ROI data, by selecting the test data with geographical location characteristics and calculating the perception delay, the perception delay of the roadside perception system of the Internet of Vehicles can be evaluated. It is also possible to evaluate whether the system frequency of the perception system is stable by calculating TDEV and MTIE, which comprehensively reflects the time-frequency performance of the roadside perception system. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0066] The present invention is implemented as follows:

[0067] A method for testing the time-frequency performance of a roadside perception system of an Internet of Vehicles, the method comprising the following steps:

[0068] S1: Start the test process and data collection: determine the ROI range of the detection area, select a true value vehicle to perform driving test in the test area, the true value vehicle records the true value data of the vehicle, and the system to be tested perceives the target of the true value vehicle in the test area and outputs structured perception data;

[0069] S2: Processing and output of test data: The true value car and the system to be tested process the perception data, and output the perception timestamp ts, the MEC sending timestamp tt, and the data of the target object information corresponding to the perception timestamp and the MEC sending timestamp tt. The data output by the true value car is Sgt, and the data output by the system to be tested is Sdut1 and Sdut2. The content of Sdut1 data contains the perception data of the system to be tested on all targets in the perception range, while Sdut2 only contains the data perceived by the true value car;

[0070] S3: Time-frequency performance analysis and evaluation.

[0071] Furthermore, the step S1 includes:

[0072] Step S11: Determine the scope of the test area. Record the longitude and latitude information of each point in the order of "southeast-east-northeast-north-northwest-west-southwest-southeast";

[0073] Step S12: Select a true value vehicle, select a place outside the sensing range of the system to be tested as the starting point of the true value vehicle, and the true value vehicle starts from the starting point to test and record the true value data of the vehicle;

[0074] Step S13: the real value vehicle travels at a normal urban traffic speed, enters along the lane from the entrance direction, and passes through the designated test area;

[0075] Step S14: the true value vehicle leaves the test area and stops collecting true value data of the vehicle;

[0076] Step S15: If the test area contains a turning road, repeat S11-S13 until all lanes are traversed and the route is driven according to the lane rules (i.e., drive straight on a straight road and turn on a turning road. If the lane can drive straight and turn at the same time, drive straight first and then turn in the next test);

[0077] If there are only straight roads in the test area, repeat steps S11-S13 until all lanes are traversed, and then repeat steps S11-S13 at least 2 additional times, and require that in the additional repeated step S12, the real value vehicle travels in a curved "S shape" and crosses at least 2 lanes through the test area; in the present application, the real value vehicle travels in a curved "S shape" and repeats steps S11-S13 to ensure that in the test scenario of a long straight road, data on vehicle turning can be collected to ensure the integrity of the collected data.

[0078] Furthermore, in step S12, the true value vehicle is outside the sensing range of the system to be tested, and is required to be at a distance of more than 400 meters from the location of the system to be tested.

[0079] Furthermore, in step S13, the real value vehicle's travel speed is 20-60 km / h.

[0080] Furthermore, in step S14, after the true value vehicle leaves the test area, it needs to travel an additional distance of at least 150 meters before stopping data collection of the vehicle to ensure the integrity of the true value data.

[0081] Furthermore, in step S2, the perception timestamp is the time when the sensor device collects the target information, wherein the sensor device includes the sensor device of the true value vehicle and the sensor device of the perception system. The MEC sending timestamp tt is: the time when the perception system outputs the perception result after processing the collected target information through the MEC.

[0082] Furthermore, in step S2, the target object information includes but is not limited to the longitude, latitude, altitude, speed, and heading angle of the target object.

[0083] Further, in step S2, Sgt, Sdut1 and Sdut2 output after each test are in one-to-one correspondence and are defined as a set of test group data. The data definitions of Sgt, Sdut1 and Sdut2 are shown in Table 1 below:

[0084] Table 1:

[0085]

[0086]

[0087] Furthermore, the step S3 includes:

[0088] S31: Unify the positioning reference points of the roadside truth system and the system to be tested;

[0089] S32: Check the data of Sgt, Sdut1 and Sdut2, retain the valid data, remove the invalid data, and draw the trajectory diagram of Sgt and Sdut2;

[0090] S33: Selecting data with geographical location characteristics as test group data;

[0091] S34: Calculate the mean △t of the difference between the MEC sending timestamp tt and the sensing timestamp ts of each frame from the data of the system to be tested Sdut2, that is, △t=mean(tt-ts);

[0092] S35: based on the sensing timestamp ts, extract the continuous position information in Sgt and Sdut2, and set them as trajectory data Pgt and Pdut respectively;

[0093] S36: Set the transformed Pdut trajectory to Where R 0 and t 0 are the trajectory rotation and translation matrices respectively;

[0094] S37: Find the transformed Pdut trajectory through KNN search method The Euclidean distance between each point in the Pgt trajectory and the shortest point in the Pgt trajectory is defined as the residual Δ ag ;

[0095] The expression is

[0096] S38: Optimize the residual Δ by LM optimization method ag , to obtain the minimum Δ ag , and calculate the corresponding R * and t * ;

[0097] S39: Based on the result of the previous step, the trajectory of the system to be tested after fitting with the Pgt trajectory is obtained as follows:

[0098] S310: Calculation The time error between the corresponding point and Pgt is The difference between the corresponding perception timestamp ts and the perception timestamp ts corresponding to Pgt, and the time errors of all points are averaged to obtain the average value δt;

[0099] S311: Calculate the perception delay Dsen=△t+δt;

[0100] S312: Calculate TDEV based on the tt timestamp in the Sdut1 data. TDEV is a measure of the change in the signal time interval within a given integration time τ, which can provide information about the signal phase and spectral content. Set the ideal sampling interval of the system to be tested to τ 0 =1 / f obj , where f obj is the default frequency of the system to be tested. When the observation interval is τ=nτ 0 When TDEV is , it can be estimated by the following formula:

[0101]

[0102] Where: ——Round down;

[0103] x i ——instantaneous time error;

[0104] N——total number of samples;

[0105] τ 0 - sampling interval;

[0106] τ——integration time;

[0107] n – the number of sampling intervals within the integration time.

[0108] S313: Calculate MTIE based on the tt timestamp in the Sdut1 data. MTIE is the total length of the measurement period T, for a given observation interval τ = nτ 0 The maximum value of the peak value of the delay time between the timing signal time interval output by the roadside perception system and the ideal time interval represents the stability of the frequency. It can be estimated by the following formula:

[0109]

[0110] Where: x i ——instantaneous time error;

[0111] N——total number of samples;

[0112] τ 0 ——Sampling interval.

[0113] Furthermore, in step S31, the positioning reference point, i.e., the origin of the local tangent plane coordinate system, is used to determine the coordinate position of the true value vehicle.

[0114] Furthermore, in step S32, the judgment condition for valid data is: the trajectories of Sgt and Sdut2 are continuous without obvious interruption, and the time interval of Sgt completely covers the time interval of Sdut1 and Sdut2; otherwise, it is regarded as invalid data. Drawing the trajectory graph of Sgt and Sdut2 can realize the visualization of Sgt and Sdut2 data, making it easy to find whether there is obvious deviation or interruption in the trajectories of the two.

[0115] Furthermore, in step S33, each test group data includes Sgt, Sdut1 and Sdut2 under the same perception timestamp.

[0116] Furthermore, in step S33, the data with geographic location characteristics is data with vehicle turning behavior, including curve test data or "S-shaped" test data, so as to ensure the integrity of the test data.

[0117] Furthermore, in step S36, two points at any position of the Pdut trajectory can be matched by trajectory rotation and translation.

[0118] In step S38, R * and t * are respectively R in step S36 0 and t 0 The corresponding specific value.

[0119] Furthermore, in step S312, the observation interval is τ=nτ 0 The smaller the value of TDEV is, the more stable the frequency of the perception system is; in step S313, the observation interval τ = nτ 0 The smaller the MTIE value is, the more stable the frequency of the perception system is.

[0120] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for testing the time-frequency performance of a roadside perception system of an Internet of Vehicles, characterized in that: The test method includes the following steps: S1: Start the test process and data collection: determine the ROI range of the detection area, select a true value vehicle to perform driving test in the test area, the true value vehicle records the true value data of the vehicle, and the system to be tested perceives the target of the true value vehicle in the test area and outputs structured perception data; S2: Processing and output of test data: The true value car and the system to be tested process the perception data, and output the perception timestamp ts, the MEC sending timestamp tt, and the data of the target object information corresponding to the perception timestamp and the MEC sending timestamp tt. The data output by the true value car is Sgt, and the data output by the system to be tested is Sdut1 and Sdut2. The content of Sdut1 data contains the perception data of all targets in the perception range of the system to be tested, and Sdut2 only contains the data perceived by the true value car; S3: Time-frequency performance analysis and evaluation; The step S3 includes: S31: Unify the positioning reference points of the roadside truth system and the system to be tested; S32: Check the data of Sgt, Sdut1 and Sdut2, retain the valid data, remove the invalid data, and draw the trajectory diagram of Sgt and Sdut2; S33: Selecting data with geographical location characteristics as test group data; S34: Calculate the mean △t of the difference between the MEC sending timestamp tt and the sensing timestamp ts of each frame from the data of the system to be tested Sdut2, that is, △t=mean(tt-ts); S35: based on the sensing timestamp ts, extract the continuous position information in Sgt and Sdut2, and set them as trajectory data Pgt and Pdut respectively; S36: Set the transformed Pdut trajectory to Where R0 and t0 are trajectory rotation and translation matrices respectively; S37: Find the transformed Pdut trajectory through KNN search method The Euclidean distance between each point in the Pgt trajectory and the shortest point in the Pgt trajectory is defined as the residual Δ ag ; The expression is S38: Optimize the residual Δ by LM optimization method ag , to obtain the minimum Δ ag , and calculate the corresponding R * and t * ; S39: Based on the result of the previous step, the trajectory of the system to be tested after fitting with the Pgt trajectory is obtained as follows: S310: Calculation The time error between the corresponding point and Pgt is The difference between the corresponding perception timestamp ts and the perception timestamp ts corresponding to Pgt, and the time errors of all points are averaged to obtain the average value δt; S311: Calculate the perception delay Dsen=△t+δt.

2. A method for testing the time-frequency performance of a roadside perception system of an Internet of Vehicles as claimed in claim 1, characterized in that: The step S1 includes: Step S11: Determine the scope of the test area; record the latitude and longitude information of each point in the order of "southeast-east-northeast-north-northwest-west-southwest-south-southeast"; Step S12: Select a true value vehicle, select a place outside the sensing range of the system to be tested as the starting point of the true value vehicle, and the true value vehicle starts from the starting point to test and record the true value data of the vehicle; Step S13: the real value vehicle travels at a normal urban traffic speed, enters along the lane from the entrance direction, and passes through the designated test area; Step S14: the true value vehicle leaves the test area and stops collecting true value data of the vehicle; Step S15: If the test area contains a turning road, repeat S11-S13 until all lanes are traversed and the route is completed according to the lane rules, that is, go straight on the straight road and turn on the turning road. If the lane can go straight and turn at the same time, go straight first and then turn in the next test; If there is only a straight lane in the test area, repeat steps S11-S13 until all lanes are traversed, and then repeat steps S11-S13 at least 2 more times, and require that in the additionally repeated step S12, the real value vehicle travels in a curved "S-shape" and crosses at least 2 lanes through the test area; In the present application, the real value vehicle travels in an "S-shaped" curve to repeat steps S11-S13, ensuring that in the test scenario of a long straight road, the vehicle turning data can also be collected, thereby ensuring the integrity of the collected data.

3. The time-frequency performance testing method of the vehicle network roadside perception system according to claim 1, characterized in that: In step S2, the target object information includes but is not limited to the longitude, latitude, altitude, speed, and heading angle of the target object.

4. The time-frequency performance testing method of the vehicle network roadside perception system according to claim 1, characterized in that: In step S2, Sgt, Sdut1 and Sdut2 output after each test are in one-to-one correspondence and are defined as a set of test group data.

5. The time-frequency performance testing method of the vehicle network roadside perception system according to claim 1, characterized in that: In step S32, the conditions for determining valid data are: the trajectories of Sgt and Sdut2 are continuous without obvious interruption, and the time interval of Sgt completely covers the time intervals of Sdut1 and Sdut2; otherwise, it is considered as invalid data.

6. The time-frequency performance testing method of the vehicle network roadside perception system according to claim 1, characterized in that: In step S33, the data with geographic location characteristics is data with vehicle turning behavior, including curve test data or "S-shaped" test data.

7. The time-frequency performance testing method of a vehicle networking roadside perception system according to claim 1, characterized in that: In step S38, R * and t * They are the specific values ​​corresponding to R0 and t0 in step S36 respectively.

8. The time-frequency performance testing method of a vehicle network roadside perception system according to claim 1, characterized in that: Step S3 also includes S312: calculating TDEV based on the MEC sending timestamp tt in the Sdut1 data; TDEV is a measure of the change in the signal time interval within a given integration time τ, and setting the ideal sampling interval of the system to be tested to τ0 = 1 / f obj , where f obj is the default frequency of the system to be tested. When the observation interval is τ=nτ0, the estimation formula of TDEV is: Where: ——Round down; x i ——instantaneous time error; N——total number of samples; τ0——sampling interval; τ——integration time; n – the number of sampling intervals within the integration time.

9. The time-frequency performance testing method of a vehicle network roadside perception system according to claim 1, characterized in that: Step S3 also includes S313: calculating MTIE based on the MEC sending timestamp tt in the Sdut1 data, where MTIE is the maximum value of the peak value of the delay time between the timing signal time interval output by the roadside perception system and the ideal time interval for a given observation interval τ=nτ0 during the measurement period with a total length of T, and is estimated by the following formula: Where: x i ——instantaneous time error; N——total number of samples; τ0——sampling interval.

Images