An adaptive vehicle dynamic response detection method and system based on asphalt pavement sensors

Abstract

This invention discloses an adaptive vehicle dynamic response detection method and system based on asphalt pavement sensors. By selecting data from multiple sensors with the highest data quality, the negative impact of poor sensor data quality on the system is avoided. Then, the detection data from multiple sensors are used for joint decision-making, and the parameters used by each sensor participating in waveform detection are dynamically adjusted according to the decision results, avoiding detection errors caused by sensor parameter drift. Subsequently, if a sensor fails to detect continuously and cannot be recovered by adjusting parameters, the system will remove it from the decision list and update the decision sensor list, avoiding system detection failures due to damage to some sensors. As long as the number of normally functioning sensors is greater than the number required in the decision list, the system can work normally, enhancing the robustness of the entire vehicle detection system and making it suitable for widespread promotion and use.

Description

Technical Field

[0001] This invention relates to the field of rail transit system technology, specifically to an adaptive vehicle dynamic response detection method and system based on asphalt pavement sensors. Background Technology

[0002] To statistically analyze vehicle load and dynamic response data at a specific location on an asphalt pavement, multiple stress-strain and pressure sensors are placed at different locations along the same cross-section of the pavement to detect the stress and strain response of the pavement when a vehicle passes over it. When a vehicle crosses the pavement, the sensor output signals exhibit significant abrupt changes, manifesting as varying waveforms in the continuous time domain. By detecting these abrupt changes in the sensor's time-domain waveform data, it is possible to determine whether a vehicle is passing over the pavement, what type of vehicle is passing, and the stress and strain exerted on the pavement's structural layers during this passage. This allows for the acquisition of all vehicle information and dynamic response data for that road segment.

[0003] Currently, most existing sensors are buried in different locations in the road, resulting in significant differences in sensor data quality. Furthermore, sensors inevitably malfunction during use, and vehicle detection data obtained based on damaged sensors will have large errors. At the same time, the basic parameters of the sensors will drift over time, making it impossible for fixed detection parameters to adapt to changes in the sensors. Therefore, it is necessary to design an adaptive vehicle dynamic response detection method and system based on asphalt pavement sensors. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies. To better address the issues of inconsistent data quality from sensors due to their varying locations within roads, inevitable sensor malfunctions during use leading to significant errors in vehicle detection data based on damaged sensors, and the drift of sensor parameters over time, making fixed detection parameters unsuitable for changing sensors, this invention provides an adaptive vehicle dynamic response detection method and system based on asphalt pavement sensors. This method selects data from multiple sensors with the highest quality, avoiding the negative impact of low-quality sensor data. It then uses the detection data from multiple sensors for joint decision-making and dynamically adjusts the parameters of each sensor involved in waveform detection based on the decision results, preventing detection errors caused by sensor parameter drift. Subsequently, if a sensor fails to detect continuously and cannot be recovered by adjusting parameters, the system removes it from the decision list and updates the decision sensor list, preventing system detection failures due to sensor malfunctions. As long as the number of normally functioning sensors exceeds the required number in the decision list, the system can operate normally, enhancing the robustness of the entire vehicle detection system.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] An adaptive vehicle dynamic response detection method based on asphalt pavement sensors includes the following steps:

[0007] Step (A): Collect all sensor data within a set time period and obtain the collected data. Then, obtain the N sensors with the best data quality ranking in the collected data, S = [S1, S2, ..., SN], and form a sensor decision list. The weight coefficient W = [W1, W2, ..., WN] of each sensor in the sensor decision list is used.

[0008] Step (B): Based on the sensor decision list, use a set time window length to find abrupt waveforms in the collected data, and output a list of waveform detection results R = [0, 1…1] within the time window;

[0009] Step (C) involves weighting the waveform detection result list R using the sensor's weighting coefficient W to obtain the waveform detection value H;

[0010] Step (D) compares the obtained waveform detection value H with the preset threshold T2. If the waveform detection value H is less than the threshold T2, it means that no vehicle has passed. If the waveform detection value H is greater than the preset threshold T2, there is waveform data in the time window and it means that a vehicle has passed.

[0011] Step (E): Based on the waveform detection result list R, count the sensors with an output of 0, and increment the number of decision errors for each sensor with an output of 0 by 1. Then, correct the sensors with too many errors.

[0012] Step (F): Set the step size and repeat step (BE) until all the sensor data is parsed and the vehicle dynamic response detection data is output.

[0013] The aforementioned adaptive vehicle dynamic response detection method based on asphalt pavement sensors includes step (A): collecting all sensor data within a set time period and obtaining the collected data; then, selecting the N sensors with the best data quality ranking from the collected data, S = [S1, S2, ..., SN], and forming a sensor decision list, where each sensor in the sensor decision list has a weight coefficient W = [W1, W2, ..., WN]. The sensors in this list are stress-strain sensors. The specific steps for determining the data quality of the collected data are as follows.

[0014] Step (A1) involves filtering the collected data X = [X1, X2, ... XN] to obtain the filtered data XF = [XF1, XF2, ... XFN].

[0015] Step (A2) calculate the average value XM of the filtered data XF = [XF1, XF2, ..., XFN], as shown in formula (1).

[0016]

[0017] Where Xi represents the i-th element in the collected data X;

[0018] Step (A3) calculates the change magnitude (Varaince1) of the filtered data compared to the original data, as shown in formula (2).

[0019]

[0020] Where XFi represents the i-th element in the filtered data XF;

[0021] Step (A4): Divide the filtered data XF into M parts and calculate the variance Varaince2 = [V1, V2, ... VM] for each of the M parts. Then set an independent threshold T for each sensor. Then iterate through each element VX in the variance Varaince2. If VX > T, then a vehicle appears at that moment and moves out of that element.

[0022] Step (A5) Calculate the mean of variance Varaincce2, Varaincce3, where the mean of variance Varaincce3 represents the degree of fluctuation in the filtered data.

[0023] Step (A6) involves taking a weighted average of the variation amplitude Varaince1 and the mean variance Varaince3 to obtain the data quality value V of the sensor.

[0024] Step (A7) sorts all the data quality values ​​V of the sensors from low to high, and selects the N sensors with the smallest data quality values ​​V to form a decision list.

[0025] The aforementioned adaptive vehicle dynamic response detection method based on asphalt pavement sensors, in step (B), uses a set time window length to find abrupt waveforms in the collected data based on the sensor decision list, and outputs a waveform detection result list R = [0, 1…1] within the time window, where Ri is the data of the i-th position in the waveform detection result list R, and each Ri in the waveform detection result list R represents the judgment of the i-th sensor in the sensor decision list on whether there is a waveform in the current time window, 0 indicates no waveform, and 1 indicates a waveform.

[0026] The aforementioned adaptive vehicle dynamic response detection method based on asphalt pavement sensors, in step (C), uses the sensor's weighting coefficient W to weight the waveform detection result list R and obtains the waveform detection value H, wherein the calculation process of the waveform detection value H is shown in formula (3).

[0027]

[0028] Where Wi represents the i-th element in the weight coefficient W.

[0029] In the aforementioned adaptive vehicle dynamic response detection method based on asphalt pavement sensor, step (D) involves comparing the obtained waveform detection value H with a preset threshold T2. If the waveform detection value H is less than the threshold T2, it indicates that no vehicle has passed. If the waveform detection value H is greater than the preset threshold T2, there is waveform data within the time window, indicating that a vehicle has passed. If the comparison result is that no vehicle has passed, the detection of the sensor ends. If the comparison result is that a vehicle has passed, the process continues to step (E).

[0030] The aforementioned adaptive vehicle dynamic response detection method based on asphalt pavement sensors includes step (E), which involves statistically analyzing the waveform detection result list R to identify sensors with an output of 0, incrementing the decision error count for each sensor with an output of 0 by 1, and then correcting sensors with excessive errors. The specific correction steps are as follows.

[0031] Step (E1): If the number of consecutive decision errors of the sensor reaches the set threshold T3, the sensor parameter correction program is started. The goal of the correction is that the number of consecutive decisions of 0 in the most recent T3 is less than T3 / 2. If the correction is successful, proceed to step (F). If the correction is unsuccessful, it is determined that the sensor is damaged and the decision sensor list is reselected. Damaged sensors must be removed during the selection.

[0032] In step (E2), if the number of sensors in the newly selected decision sensor list is less than the requirement N of the decision list, it indicates that there are too many damaged sensors and a warning will be displayed.

[0033] An adaptive vehicle dynamic response detection system based on asphalt pavement sensors includes a data acquisition module, a sensor decision list construction module, a sudden waveform detection module, a waveform detection value output module, a waveform detection value comparison module, a sensor calibration module, and a vehicle dynamic response detection data output module. The data acquisition module is used to collect all sensor data within a set time period and obtain the collected data.

[0034] The sensor decision list building module is used to obtain the N sensors S = [S1, S2, ..., SN] with the best data quality ranking in the collected data and form a sensor decision list, and the weight coefficient W = [W1, W2, ..., WN] of each sensor in the sensor decision list;

[0035] The abrupt change waveform detection module is used to find abrupt change waveforms in the collected data based on the sensor decision list and using a set time window length, and outputs a list of waveform detection results R = [0, 1… 1] within the time window;

[0036] The waveform detection value output module is used to weight the waveform detection result list R using the sensor's weighting coefficient W, and obtain the waveform detection value H;

[0037] The waveform detection value comparison module is used to compare the obtained waveform detection value H with a preset threshold T2. If the waveform detection value H is less than the threshold T2, it means that no vehicle has passed. If the waveform detection value H is greater than the preset threshold T2, there is waveform data in the time window and it means that a vehicle has passed.

[0038] The sensor calibration module is used to count the sensors with an output of 0 based on the waveform detection result list R, and to increment the number of decision errors of the sensors with an output of 0 by 1, and then calibrate the sensors with too many errors.

[0039] The vehicle dynamic response detection data output module is used to set the operation of the step-size repetitive abrupt change waveform detection module, waveform detection value output module, waveform detection value comparison module and sensor calibration module until all the sensor's collected data is parsed and the vehicle dynamic response detection data is output.

[0040] The beneficial effects of this invention are as follows: The adaptive vehicle dynamic response detection method and system based on asphalt pavement sensors of this invention avoids the negative impact of low sensor data quality by selecting data collected from multiple sensors with the highest data quality. Then, it uses the detection data from multiple sensors for joint decision-making and dynamically adjusts the parameters used by each sensor participating in waveform detection based on the decision results, avoiding detection errors caused by sensor parameter drift. Subsequently, if a sensor fails to detect continuously and cannot be recovered by adjusting parameters, the system removes it from the decision list and updates the decision sensor list, avoiding system detection failures due to damage to certain sensors. As long as the number of normally functioning sensors exceeds the number required in the decision list, the system can operate normally, enhancing the robustness of the entire vehicle detection system. This contributes to improving the theoretical and standard systems of highway pavement construction and maintenance technology, providing service and support for the high-quality development of highway pavements. Attached Figure Description

[0041] Figure 1 This is a flowchart of an adaptive vehicle dynamic response detection method based on an asphalt pavement sensor according to the present invention;

[0042] Figure 2 This is a comparison chart of the detection results of an embodiment of the present invention. Detailed Implementation

[0043] The present invention will now be further described with reference to the accompanying drawings.

[0044] like Figure 1 As shown, the present invention provides an adaptive vehicle dynamic response detection method based on asphalt pavement sensors, comprising the following steps:

[0045] Step (A) involves collecting all sensor data within a set time period and obtaining the collected data. Then, the N sensors with the best data quality ranking from the collected data, S = [S1, S2, ..., SN], are selected and formed into a sensor decision list. Each sensor in the sensor decision list has a weight coefficient W = [W1, W2, ..., WN]. The sensors in this list are stress-strain sensors. The specific steps for determining the data quality of the collected data are as follows:

[0046] Step (A1) involves filtering the collected data X = [X1, X2, ... XN] to obtain the filtered data XF = [XF1, XF2, ... XFN].

[0047] Among them, the filtering algorithm can be smoothing filter and Kalman filter;

[0048] Step (A2) calculate the average value XM of the filtered data XF = [XF1, XF2, ..., XFN], as shown in formula (1).

[0049]

[0050] Where Xi represents the i-th element in the collected data X;

[0051] Step (A3) calculates the change magnitude (Varaince1) of the filtered data compared to the original data, as shown in formula (2).

[0052] The larger the variation amplitude Varaince1, the more obvious the noise of the sensor.

[0053]

[0054] Where XFi represents the i-th element in the filtered data XF;

[0055] Step (A4): Divide the filtered data XF into M parts and calculate the variance Varaince2 = [V1, V2, ... VM] for each of the M parts. Then set an independent threshold T for each sensor. Then iterate through each element VX in the variance Varaince2. If VX > T, then a vehicle appears at that moment and moves out of that element.

[0056] The threshold T is set according to the characteristics of the sensor.

[0057] Step (A5) Calculate the mean of variance Varaincce2, Varaincce3, where the mean of variance Varaincce3 represents the degree of fluctuation in the filtered data.

[0058] Among them, the smaller the mean variance (Varaince3), the better the stability of the sensor;

[0059] Step (A6) involves taking a weighted average of the variation amplitude Varaince1 and the mean variance Varaince3 to obtain the data quality value V of the sensor.

[0060] Step (A7) sorts all the data quality values ​​V of the sensors from low to high, and selects the N sensors with the smallest data quality values ​​V to form a decision list.

[0061] Step (B) involves using a set time window length to find abrupt waveforms in the collected data based on the sensor decision list, and outputting a waveform detection result list R = [0, 1…1] within the time window, where Ri is the data of the i-th position in the waveform detection result list R, and each Ri in the waveform detection result list R represents the judgment of the i-th sensor in the sensor decision list on whether there is a waveform in the current time window, with 0 indicating no waveform and 1 indicating a waveform.

[0062] Among them, finding abrupt waveforms takes advantage of the fact that the actual time it takes for a vehicle to pass through a certain point is not too long, and each sensor can run the set detection algorithm.

[0063] In step (C), the waveform detection result list R is weighted using the sensor's weighting coefficient W to obtain the waveform detection value H. The calculation process of the waveform detection value H is shown in formula (3).

[0064]

[0065] Where Wi represents the i-th element in the weight coefficient W.

[0066] In step (D), the obtained waveform detection value H is compared with the preset threshold T2. If the waveform detection value H is less than the threshold T2, it means that no vehicle has passed. If the waveform detection value H is greater than the preset threshold T2, there is waveform data in the time window and it means that a vehicle has passed. If the comparison result is that no vehicle has passed, the detection of the sensor ends. If the comparison result is that a vehicle has passed, the process continues to step (E).

[0067] Step (E) involves compiling a list of waveform detection results R, identifying sensors with an output of 0, and incrementing the number of decision errors for each sensor with an output of 0 by 1. Sensors with excessive errors are then corrected. The specific correction steps are as follows.

[0068] Step (E1): If the number of consecutive decision errors of the sensor reaches the set threshold T3, the sensor parameter correction program is started. The goal of the correction is that the number of consecutive decisions of 0 in the most recent T3 is less than T3 / 2. If the correction is successful, proceed to step (F). If the correction is unsuccessful, it is determined that the sensor is damaged and the decision sensor list is reselected. Damaged sensors must be removed during the selection.

[0069] In step (E2), if the number of sensors in the newly selected decision sensor list is less than the requirement N of the decision list, it indicates that there are too many damaged sensors and a warning will be displayed.

[0070] Step (F): Set the step size and repeat step (BE) until all the sensor data is parsed and the vehicle dynamic response detection data is output.

[0071] An adaptive vehicle dynamic response detection system based on asphalt pavement sensors includes a data acquisition module, a sensor decision list construction module, a sudden waveform detection module, a waveform detection value output module, a waveform detection value comparison module, a sensor calibration module, and a vehicle dynamic response detection data output module. The data acquisition module collects data from all sensors within a set time period and obtains the collected data. The sensor decision list construction module selects the N sensors with the best data quality ranking from the collected data, S = [S1, S2, ..., SN], and constructs a sensor decision list, where each sensor in the decision list has a weight coefficient W = [W1, W2, ..., WN]. The sudden waveform detection module uses the sensor decision list to find sudden waveforms in the collected data within a set time window and outputs a waveform detection result list R = [0, 1…1] within that time window. The measurement output module is used to weight the waveform detection result list R using the sensor's weighting coefficient W and obtain the waveform detection value H. The waveform detection value comparison module is used to compare the obtained waveform detection value H with a preset threshold T2. If the waveform detection value H is less than the threshold T2, it indicates that no vehicle has passed. If the waveform detection value H is greater than the preset threshold T2, it indicates that there is waveform data within the time window and that a vehicle has passed. The sensor calibration module is used to count the sensors with an output of 0 based on the waveform detection result list R, and increment the number of decision errors for each sensor with an output of 0 by 1. Then, it calibrates the sensors with too many errors. The vehicle dynamic response detection data output module is used to set the operation of the step-size repeating abrupt waveform detection module, waveform detection value output module, waveform detection value comparison module, and sensor calibration module until all the sensor's collected data is parsed and the vehicle dynamic response detection data is output.

[0072] To better illustrate the effects of the present invention, a specific embodiment of the present invention is described below.

[0073] The sensor data deployed in the asphalt pavement in this embodiment includes data from sensors processing soil pressure at the base course, soil pressure at the subgrade, longitudinal data at the bottom of the lower layer, longitudinal data at the bottom of the lower base course, longitudinal data at the bottom of the subbase course, transverse data at the bottom of the subbase course, and transverse data at the bottom of the lower base course. The sensor with the least noise and the largest amplitude is the soil pressure sensor at the base course. The results based on the soil pressure sensor at the base course and the combined adaptive detection results based on the five sensors are as follows: Figure 2 As shown.

[0074] Depend on Figure 2 It can be seen that the dynamic weighing system accurately counts the passing vehicles, and the comparison between single-sensor and multi-sensor data shows that during peak vehicle traffic (such as 8 or 9 am), when different models of vehicles pass through the same location, the detection rate of vehicles by a single sensor is much lower than the result of multi-sensor joint detection.

[0075] In summary, the adaptive vehicle dynamic response detection method and system based on asphalt pavement sensors of the present invention first collects all sensor data within a set time period and obtains the collected data. Then, it acquires the N sensors S with the best data quality ranking from the collected data and forms a sensor decision list, with a weight coefficient W for each sensor in the sensor decision list. Next, based on the sensor decision list, it searches for abrupt waveforms in the collected data using a set time window length and outputs a waveform detection result list R within that time window. Then, it uses the sensor weight coefficient W to weight the waveform detection result list R and obtains the waveform detection value H. Subsequently, it compares the obtained waveform detection value H with a preset threshold T2. If the waveform detection value H is less than the threshold T2, it indicates that no vehicle has passed; if the waveform detection value H is greater than the preset threshold T2, it indicates that there is waveform data within that time window and that a vehicle has passed. Finally, it statistically analyzes the sensors with an output of 0 in the waveform detection result list R. The system increments the number of decision errors for sensors with an output of 0 by 1, then corrects sensors with excessive errors. Finally, it repeats this process with a set step size until all sensor data is parsed, outputting vehicle dynamic response detection data. This invention avoids the negative impact of poor sensor data quality by selecting data from multiple sensors with the highest quality. It then uses the detection data from multiple sensors for joint decision-making and dynamically adjusts the parameters used by each sensor participating in waveform detection based on the decision results, avoiding detection errors caused by sensor parameter drift. Subsequently, if a sensor fails to detect continuously and cannot be recovered by adjusting parameters, the system removes it from the decision list and updates the decision sensor list, preventing system detection failures due to sensor malfunction. As long as the number of normally functioning sensors exceeds the number required in the decision list, the system can operate normally, enhancing the robustness of the entire vehicle detection system.

[0076] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An adaptive vehicle dynamic response detection method based on asphalt pavement sensors, characterized in that: Includes the following steps, Step (A) involves collecting all sensor data within a set time period and obtaining the collected data. Then, the N sensors with the best data quality ranking from the collected data, S=[S1,S2,...,SN], are selected and formed into a sensor decision list. Each sensor in the sensor decision list has a weight coefficient W=[W1,W2,...WN]. The sensors in this list are stress-strain sensors. The specific steps for determining the data quality of the collected data are as follows: Step (A1): Filter the collected data X=[X1,X2,…XN] and obtain the filtered data XF=[XF1,XF2,…XFN]; Step (A2): Calculate the average value XM of the filtered data XF=[XF1,XF2,…XFN], as shown in formula (1). （1） in, This represents the i-th element in the collected data X; Step (A3): Calculate the magnitude of change in the filtered data compared to the original data. As shown in formula (2), （2） in, This represents the i-th element in the filtered data XF; Step (A4): Divide the filtered data XF into M parts, and calculate the variance of each of the M parts. Then, an independent threshold T is set for each sensor, and then the variance is iterated. For each element VX in the array, if VX > T, then a vehicle appears at that moment and is removed from that element; Step (A5), calculate the variance mean And the mean of variance Indicates the degree of fluctuation in the filtered data; Step (A6), regarding the range of change and variance mean The data quality value V of the sensor is obtained by performing a weighted average. Step (A7): Sort all the data quality values ​​V of the sensors from low to high, and select the N sensors with the smallest data quality values ​​V to form a decision list; Step (B): Based on the sensor decision list, use a set time window length to find abrupt waveforms in the collected data, and output the waveform detection result list R=[0,1…1] within the time window; Step (C): The waveform detection result list R is weighted using the sensor's weighting coefficient W, and the waveform detection value H is obtained. Step (D): Compare the obtained waveform detection value H with the preset threshold T2. If the waveform detection value H is less than the threshold T2, it means that no vehicle has passed. If the waveform detection value H is greater than the preset threshold T2, there is waveform data in the time window and it means that a vehicle has passed. Step (E): Based on the waveform detection result list R, count the sensors with an output of 0, and increment the number of decision errors for each sensor with an output of 0 by 1. Then, correct the sensors with too many errors. Step (F): Set the step size and repeat step (BE) until all the sensor data is parsed and the vehicle dynamic response detection data is output.

2. The adaptive vehicle dynamic response detection method based on asphalt pavement sensors according to claim 1, characterized in that: Step (B) involves using a set time window length to find abrupt waveforms in the collected data based on the sensor decision list, and outputting a list of waveform detection results R=[0,1…1] within that time window. Let i be the data at the i-th position in the waveform detection result list R, and let each of the waveform detection result lists R be... Both represent the judgment of the i-th sensor in the sensor decision list on whether there is a waveform in the current time window, where 0 indicates no waveform and 1 indicates a waveform.

3. The adaptive vehicle dynamic response detection method based on asphalt pavement sensors according to claim 2, characterized in that: Step (C) involves weighting the waveform detection result list R using the sensor's weighting coefficient W to obtain the waveform detection value H. The calculation process for the waveform detection value H is shown in formula (3). （3） in, This represents the i-th element in the weight coefficient W.

4. The adaptive vehicle dynamic response detection method based on asphalt pavement sensors according to claim 3, characterized in that: Step (D) compares the obtained waveform detection value H with the preset threshold T2. If the waveform detection value H is less than the threshold T2, it means that no vehicle has passed. If the waveform detection value H is greater than the preset threshold T2, there is waveform data in the time window and it means that a vehicle has passed. If the comparison result is that no vehicle has passed, the detection of the sensor ends. If the comparison result is that a vehicle has passed, the process continues to step (E).

5. The adaptive vehicle dynamic response detection method based on asphalt pavement sensors according to claim 4, characterized in that: Step (E): Based on the waveform detection result list R, count the sensors with an output of 0, and increment the number of decision errors for each sensor with an output of 0 by 1. Then, correct the sensors with too many errors. The specific correction steps are as follows. Step (E1): If the number of consecutive decision errors of the sensor reaches the set threshold T3, the sensor parameter correction program is started. The goal of the correction is that the number of consecutive decisions of 0 in the most recent T3 times is less than T3 / 2. If the correction is successful, proceed to step (F). If the correction is unsuccessful, it is determined that the sensor is damaged and the decision sensor list is reselected. Damaged sensors must be removed during the selection. In step (E2), if the number of sensors in the newly selected decision sensor list is less than the requirement N of the decision list, it indicates that there are too many damaged sensors and a warning will be displayed.

6. An adaptive vehicle dynamic response detection system based on asphalt pavement sensors, wherein the operation of the detection system is based on the detection method according to any one of claims 1-5, characterized in that: It includes a data acquisition module, a sensor decision list construction module, a sudden waveform detection module, a waveform detection value output module, a waveform detection value comparison module, a sensor calibration module, and a vehicle dynamic response detection data output module. The data acquisition module is used to collect all sensor data within a set time period and obtain the collected data. The sensor decision list building module is used to obtain the N sensors with the best data quality ranking in the collected data, S=[S1,S2,...,SN], and form a sensor decision list, and the weight coefficient W=[W1,W2,...WN] of each sensor in the sensor decision list; The abrupt change waveform detection module is used to find abrupt change waveforms in the collected data based on the sensor decision list using a set time window length, and output a list of waveform detection results R=[0,1…1] within the time window; The waveform detection value output module is used to weight the waveform detection result list R using the sensor's weighting coefficient W, and obtain the waveform detection value H; The waveform detection value comparison module is used to compare the obtained waveform detection value H with a preset threshold T2. If the waveform detection value H is less than the threshold T2, it means that no vehicle has passed. If the waveform detection value H is greater than the preset threshold T2, there is waveform data in the time window and it means that a vehicle has passed. The sensor calibration module is used to count the sensors with an output of 0 based on the waveform detection result list R, and to increment the number of decision errors of the sensors with an output of 0 by 1, and then calibrate the sensors with too many errors. The vehicle dynamic response detection data output module is used to set the operation of the step-size repetitive abrupt change waveform detection module, waveform detection value output module, waveform detection value comparison module and sensor calibration module until all the sensor's collected data is parsed and the vehicle dynamic response detection data is output.

Images