A vehicle safety management method and system based on cloud computing

Abstract

The present invention relates to the technical field of vehicle safety management, and particularly relates to a vehicle safety management method and system based on cloud computing. The method includes: obtaining the state data of the currently traveling vehicle in real time; respectively calculating the Euclidean distances between the state data and multiple clustering clusters; if the minimum Euclidean distance is greater than a set threshold, the state data is abnormal and an alarm is given; that is, the solution of the present invention can manage the vehicle in a timely and safe manner.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle safety management. More specifically, the present invention relates to a vehicle safety management method and system based on cloud computing. Background Art

[0002] With the rapid development of intelligent transportation systems and vehicle electronics technologies, vehicle safety management has increasingly become an important topic in traffic management and road safety.

[0003] In this context, by introducing a vehicle safety management system based on cloud computing, enterprises can obtain and analyze comprehensive information about vehicles in real time, improve the immediacy of vehicle maintenance, thereby significantly reducing the accident rate and enhancing road safety. This transformation not only conforms to the trend of the development of intelligent transportation but also lays a solid foundation for future driverless and intelligent traffic management.

[0004] In the prior art, for example, a patent document with the authorization announcement number CN104732353B discloses a vehicle integrated management system based on cloud computing. In it, the safety management analyzes the driving safety level of the current vehicle based on the prediction result of the vehicle driving state, the detection result of surrounding targets, and the positioning information of the remaining vehicles in the cloud network, and gives the result with reference to the pre-set level and issues it to the vehicle safety management unit.

[0005] The above solution only determines the safety of the vehicle according to the current driving state and the pre-set level, and the pre-set level is generally set by experience. For different vehicles, due to differences in the number of trips, mileage, etc., the judgment criteria will be different. Therefore, using a unified standard to judge different vehicles may have inaccurate problems, which may lead to inaccurate analysis of vehicle safety data. Summary of the Invention

[0006] The object of the present invention is to provide a vehicle safety management method and system based on cloud computing to solve the problem in the prior art that the vehicle safety data cannot be accurately analyzed, resulting in the inability to accurately judge whether the vehicle is safe. For this purpose, the present invention provides solutions in the following two aspects.

[0007] In the first aspect, a vehicle safety management method based on cloud computing provided by the present invention includes:

[0008] Obtain the state data of the currently traveling vehicle and the normalized new state data in real time;

[0009] Calculate the Euclidean distances between the new state data and multiple clustering clusters respectively; if the minimum Euclidean distance is greater than the set threshold, the state data is abnormal and an alarm is given;

[0010] The clustering clusters are obtained by clustering the normalized historical state data of the vehicle under different driving times using the K-means algorithm; each historical state data includes multiple items of data; the same item data under all times forms a sequence from small to large;

[0011] Among them, the distance in the clustering process is positively correlated with the difference between any item of data in any two historical state data and the corresponding weight; the weight is the product of the influence sensitivity and the importance degree of the corresponding item sequence; the importance degree is positively correlated with the fluctuations of each data in the corresponding sequence and the fluctuations of the safety indices corresponding to each data; the safety index characterizes the speed fluctuation degree of the vehicle during each driving;

[0012] Influence sensitivity is: ; is the range of the safety index, , are respectively the range of the th item sequence, the number of data, , are respectively the th item sequence, the th, the th data, , , are respectively the safety index, the abnormality degree of the th item sequence, the th data, the safety index of the th data;

[0013] The set threshold is: ; is the set multiple, is the average value of the Euclidean distances between all historical state data and the clustering centers in their corresponding clustering clusters.

[0014] In the above solution, by comprehensively considering the correlation degree of each item of data with the degree of speed change during vehicle driving and the influence sensitivity of each item of data on the degree of speed change during driving, the weight of each item of data is determined, and the difference of each item of data is weighted to finally obtain the distance between the historical state data of any two drives and perform clustering, so that the clustering algorithm can better consider the influence of each item of data on the importance of vehicle safety, thereby making the classification of the vehicle's historical data more accurate; furthermore, by comparing the current state data with the classified historical data of the vehicle, the present invention can make the analysis of the vehicle's safety data more accurate and improve the accuracy of the judgment of the vehicle data during the current driving; that is, compared with the conventional K-means clustering algorithm (when clustering, directly measuring the distance according to the Euclidean distance between data points without considering the importance and influence sensitivity of each item of data on vehicle safety), the solution of the present invention improves the accuracy of the analysis of the vehicle state data obtained subsequently in real time.

[0015] Meanwhile, by using the mean value of the distances from the data in multiple clustering clusters in the vehicle's historical data to the clustering center to determine the set threshold, the accuracy of the vehicle state data judgment can be improved.

[0016] Optionally, the state data includes but is not limited to vehicle fuel quantity, average tire pressure, average vehicle driving speed, total vehicle usage years, brake usage frequency, and whether an accident occurs.

[0017] In the above solution, by comprehensively considering the data from multiple angles during vehicle driving, it can more comprehensively provide data support for subsequent vehicle safety analysis.

[0018] Optionally, the importance degree is: ; is the th item of data corresponding to the th drive, , are respectively the average value and the range of the th sequence, is the safety index during the th drive, , are the range and the average value of the safety index, and n is the number of historical drives of the vehicle.

[0019] In the above solution, by combining the fluctuations of each item of data and the fluctuations of the safety indexes corresponding to each item of data, the importance of the corresponding item of data can be determined.

[0020] Optionally, the distance is:

[0021] ;

[0022] Wherein, N is the total number of items of historical status data, is the th item of data corresponding to the th driving, is the th item of data corresponding to the th driving, is the weight of the th item of data.

[0023] Since the impact of each item of data on the vehicle safety index is different, it is necessary to determine the weights of different items of data to obtain a more accurate distance and perform more accurate clustering on the historical status data of the vehicle.

[0024] Optionally, the safety index is: ;

[0025] Wherein, represents the range of the speed sequence during the th driving, represents the th speed value in the speed sequence during the th driving, represents the average value of the speed sequence during the th driving, is an exponential function with the natural constant e as the base, is the th total number of speeds in the speed sequence during the

[0026] Optionally, the degree of abnormality is obtained by performing anomaly detection on the safety index data using the LOF algorithm; the safety index data includes a plurality of safety indicators composed of each item of data and the corresponding safety index.

[0027] Optionally, the method for normalizing the status data includes maximum-minimum normalization.

[0028] In a second aspect, a vehicle safety management system based on cloud computing includes:

[0029] A processor;

[0030] A memory storing computer instructions for vehicle safety management based on cloud computing, which when run by the processor cause the system to execute the above-mentioned vehicle safety management method based on cloud computing.

[0031] The beneficial effects of the present invention are:

[0032] The solution of the present invention determines the empirical data of the vehicle during historical operation by analyzing the importance of various data in the historical state data of the vehicle, and compares the current state data with the vehicle's own empirical data, which can determine whether the data of the current vehicle driving is abnormal and improve the accuracy of the analysis of the vehicle state data. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] By referring to the accompanying drawings and reading the following detailed description, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understandable. In the drawings, several embodiments of the present invention are shown by way of illustration and not limitation, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0034] Figure 1 Schematically shows the flowchart of the steps of a vehicle safety management method based on cloud computing in this embodiment;

[0035] Figure 2 Schematically shows the block diagram of the structure of a vehicle safety management system based on cloud computing in this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the protection scope of the present invention.

[0037] Specifically, taking a certain vehicle as an example, a vehicle safety management method based on cloud computing in this embodiment will be introduced, as Figure 1 shown, including the following steps:

[0038] Step S1, obtaining the state data of the currently traveling vehicle in real time.

[0039] The safety management data of the current vehicle collected in this embodiment includes but is not limited to state data, where the state data includes multiple items of data, specifically data such as vehicle fuel quantity, average tire pressure, average vehicle driving speed, total years of vehicle use, brake use frequency, and whether an accident has occurred.

[0040] Specifically, during the vehicle driving process, data such as vehicle fuel quantity (electric quantity), tire pressure, and driving speed are collected through corresponding sensors; the total years of vehicle use are obtained according to the vehicle purchase date; the brake use frequency is monitored according to the brake system.

[0041] It should be noted that the above-mentioned average tire pressure and average vehicle driving speed are used to characterize the state data of the entire vehicle driving process; for example, after the vehicle has been driving for two hours, the driving speed and tire pressure are collected through a set collection frequency. Then, from the start of driving to the current moment, in this embodiment, the average tire pressure and average driving speed are calculated to obtain partial state data of the vehicle driving.

[0042] In this embodiment, the state data is also normalized to obtain new normalized state data. The normalization method is the maximum-minimum value method.

[0043] Step S2: Calculate the distances between the current state data and multiple clustering clusters respectively.

[0044] In this embodiment, multiple clustering clusters are obtained by clustering the normalized historical state data of the vehicle under different driving times by using the K-means algorithm.

[0045] Specifically, the process of clustering the historical state data of the vehicle under different driving times by using the K-means algorithm is as follows:

[0046] First, obtain multiple historical state data of the current vehicle, and perform normalization processing on them respectively to obtain normalized historical state data, and set the K value during clustering.

[0047] Specifically, according to big data, each item of data in the historical state data from the start to the end of each vehicle operation is obtained, specifically including data such as vehicle fuel (electricity), average tire pressure, average driving speed, brake usage frequency, total vehicle usage years, etc.

[0048] In this embodiment, the th item of data in the state data when the vehicle is driving for the th time is denoted as , the total number of historical driving times of the vehicle obtained is denoted as , and the total number of items of data in each historical state data is denoted as .

[0049] Among them, the same item data under all driving times constitutes a collection sequence, and the data in the collection sequence is sorted from small to large to obtain a sorted sequence. It should be noted that the number of the collection sequence and the sorted sequence is equal, only the data arrangement methods are different.

[0050] Secondly, randomly select K historical state data as the initial centers, calculate the distances from each historical state data to each initial center, and assign each data point to the clustering cluster corresponding to the centroid with the smallest distance to it.

[0051] Then, update the cluster centers of the clusters. Specifically, calculate the cluster centers of each cluster based on all historical state data in the cluster to update the initial centers.

[0052] Finally, calculate the distances from all historical state data to each cluster center again, and reassign the clusters according to the distances between the historical state data and the cluster centers. Iterate in this way until the cluster centers no longer change, and complete the clustering of the historical state data.

[0053] The process of obtaining the above distances includes steps S21 - S24, specifically:

[0054] Step S21, calculate the safety index of the vehicle during each driving and the importance of each item of data to vehicle safety. Whether the vehicle is safe during driving is mainly reflected in the change of speed. If the speed is relatively stable during a vehicle operation, it means that the vehicle operation is relatively safe during this time and the overall operation state has little fluctuation; on the contrary, if the speed changes violently during the vehicle operation, it means that the driving is unstable and there may be safety problems. Therefore, in this embodiment, the safety index of the vehicle is determined by the speed change of the vehicle during each driving.

[0055] Exemplarily, for each vehicle operation from start to end, the preset data acquisition frequency is 5s, and the vehicle speed data is acquired every 5s. Obtain the speed sequence of the vehicle from the start to the end of the th driving, denoted as , and denote the total number of speeds included therein as , denote the th speed value as , and calculate the safety index of the vehicle during the th driving.

[0056] In one embodiment, the safety index is:

[0057] ;

[0058] where represents the safety index during the th driving, represents the range of the speed sequence during the th driving, represents the th speed value in the speed sequence during the th driving, represents the average value of the speed sequence of the th vehicle operation, is the exponential function with the natural constant e as the base.

[0059] Wherein the range is the difference between the maximum speed value and the minimum speed value in the speed sequence.

[0060] It should be noted that for any single instance of the vehicle's historical driving, it is necessary to obtain its safety index based on its speed changes, that is, each time the vehicle drives, it corresponds to a safety index, and when driving n times, it corresponds to n safety indices.

[0061] In one embodiment, according to the th item of data and the safety index, obtain the importance degree of the th item of data: ;

[0062] Wherein, represents the importance degree of the th item of data, represents the th item of data corresponding to the th driving, represents the average value of the th acquisition sequence (or sorted sequence), represents the range of the th acquisition sequence (or sorted sequence), represents the safety index at the th driving, represents the average value of all safety index data, represents the range of the safety index, and n is the number of historical driving times of the vehicle.

[0063] For the importance degree formula, among them, represents the difference between the th item of data of the th driving and the average value; in the formula, represents the difference between the safety index of the th driving and the average value of the safety index. Multiplying the two differences can better represent the correlation degree between the th item of data and the safety index, that is, the importance degree of the th item of data to the safety index.

[0064] The reason for obtaining the importance degree of each item of data in the historical state data for vehicle safety is that the impacts of different items of data on vehicle safety are different. By obtaining the correlation between each item of data and vehicle safety, the importance of different items of data can be determined. That is, if the correlation between any item of data and vehicle safety is higher, the weight of this item of data should be appropriately increased during the clustering process through distance measurement, and vice versa, the weight of this item of data should be decreased.

[0065] Step S22: Calculate the impact sensitivity of each piece of data on vehicle safety.

[0066] Among them, for any piece of data in the historical status data, the greater the impact of a certain degree of fluctuation on the safety index, the more obvious the impact of this piece of data on the change of the safety index. Therefore, the impact sensitivity of each piece of data is obtained according to the relative change situation between the impacts of each piece of data on the safety index.

[0067] Specifically, arrange the th piece of data in ascending order to obtain the sorted sequence of the th piece of data . Obtain the impact sensitivity of each piece of data according to the corresponding relationship between the fluctuation of each piece of data and the safety index as:

[0068] ;

[0069] Among them, is the range of the safety index, , are respectively the range of the th sequence and the number of data, , are respectively the th data and the th data in the th sequence, , , are respectively the safety index, the degree of abnormality, and the safety index of the th data and the th data in the th sequence.

[0070] Among them, the number of data in the th sequence is equal to the historical driving times n of the vehicle.

[0071] In the above, the degree of abnormality is obtained by performing outlier detection on all the data in the th sequence through the LOF algorithm. Since the LOF (Local Outlier Factor) algorithm is a prior art, it will not be elaborated here.

[0072] As the most preferred implementation manner, for any piece of data, due to the existence of special circumstances, the reliability of the data itself may be low. Therefore, it is necessary to perform outlier detection on the safety index data composed of the data and the safety index to obtain the degree of abnormality and improve the reliability of the calculation. Specifically, all the data in the th sequence and the corresponding safety index form the The safety index data of the th item is subjected to anomaly detection through the LOF algorithm to obtain the anomaly degree of the th data in the th sequence.

[0073] It should be noted that the above th item of safety index data includes multiple safety indicators composed of each item of data and the corresponding safety index.

[0074] Among them, represents the ratio of the range of the safety index to the range of the th item of data. The larger this ratio is, the larger the value of the numerator is relative to the data of the denominator; in the formula, is the ratio of the difference between two safety indexes corresponding to two adjacent data in the sorted sequence to the difference between the corresponding two adjacent data. Through this ratio, it can be reflected that the difference between two adjacent data in the th sequence changes the safety index data. That is, when the difference between two safety indexes corresponding to two adjacent data is fixed, the smaller the difference between two adjacent data is, the larger the value of is. At this time, the influence sensitivity of the

[0075] th item of data on the safety index data is greater.

[0076] The above influence sensitivity characterizes the magnitude of the influence of the change of each item of data on the safety index obtained according to the change of the safety index corresponding to the change of each item of data. can reduce the reference degree of data with low reliability to the influence sensitivity.

[0077] It should be noted that the importance represents the correlation between each item of data and the safety index data, and the influence sensitivity analyzes the degree of change that the change of each item of data will cause to the safety index data. Therefore, in the process of weighting the distance metric of each item of data, it is necessary to comprehensively consider the importance and influence sensitivity of each item of data.

[0078] Step S23, calculate the weight of each item of data.

[0079] Specifically, the product of the importance and influence sensitivity of each item of data is used as the weight of this item of data, that is, the weight of the th item of data is: ,

[0080] Among them, is the weight of the th item of data, represents the influence sensitivity of the th item of data, Indicates the importance level of the th item of data. In step S24, the difference between the corresponding item data in any two historical state data is weighted by the weight to obtain the distance.

[0081] Exemplarily, when the historical state data corresponding to the th driving is the initial center or the clustering center, the distance from the historical state data at the th driving to the initial center (or clustering center) is:

[0082] ;

[0083] where N is the total number of items of historical state data, is the th item of data corresponding to the th driving, is the th item of data corresponding to the th driving.

[0084] In the process of clustering vehicle safety data through the K-Means clustering algorithm as described above, since the impact of each item of data on the vehicle safety index is different, it is necessary to weight the distance metric between the historical state data of any two drives accordingly to obtain a more accurate distance metric for more accurate clustering of the historical state data of the vehicle.

[0085] After obtaining multiple clustering clusters, in this embodiment, it is also necessary to calculate the Euclidean distance between the state data of the currently driving vehicle and the clustering centers of the multiple clustering clusters respectively.

[0086] In step S3, when the minimum distance is greater than the set threshold, it is considered that the vehicle data is abnormal and a warning is issued.

[0087] In this embodiment, the minimum value of the Euclidean distance between the current state data and the clustering centers of each clustering cluster is selected as the evaluation value of the current state data.

[0088] When the minimum distance is greater than the set threshold, it proves that the current state data may be abnormal.

[0089] In one embodiment, the size of the set threshold can be determined according to the actual situation.

[0090] Of course, as other embodiments, the size of the set threshold is:

[0091] ; where is the set multiple, To obtain the average of the Euclidean distances between all data points and the cluster centers in their corresponding clusters.

[0092] Where the value of the set multiple is 1.3 times. Of course, it can also be adjusted according to the actual situation.

[0093] In this embodiment, when the distance between the minimum distance and the nearest cluster center is greater than the set threshold, it is considered that the current vehicle safety data is abnormal, and timely warning is required, and the vehicle needs to be repaired or maintained in time.

[0094] At the same time, as the historical state data of the vehicle increases, the set threshold may also change accordingly.

[0095] The solution of the present invention improves the accuracy of the analysis of the vehicle state data obtained in real time subsequently by analyzing the importance of each item of data in the vehicle state data.

[0096] The present invention also provides a vehicle safety management system based on cloud computing. As Figure 2 shown, the system includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a vehicle safety management method based on cloud computing as described above according to the present invention is implemented.

[0097] The management system also includes other components well known to those skilled in the art such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be described in detail here.

[0098] In the present invention, the foregoing memory may be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory (RRAM), a dynamic random access memory (DRAM), a static random access memory (SRAM), an enhanced dynamic random access memory (EDRAM), a high-bandwidth memory (HBM), a hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that can be stored or otherwise maintained by such a computer-readable medium.

[0099] In the description of this specification, "a plurality of" means at least two, such as two, three, or more, etc., unless otherwise specifically defined.

[0100] Although this specification has shown and described multiple embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and scope of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in the practice of the present invention.

Claims

1. A vehicle safety management method based on cloud computing, characterized in that, Including: Obtain the status data of the currently moving vehicle and the normalized new status data in real time; Calculate the Euclidean distances between the new status data and multiple clustering clusters respectively; if the minimum Euclidean distance is greater than the set threshold, the status data is abnormal and an alarm is given; The clustering clusters are obtained by clustering the normalized historical status data of the vehicle obtained under different driving times by using the K-means algorithm; each historical status data includes multiple items of data; the same item data under all times forms a sequence from small to large; Among them, the distance in the clustering process is positively correlated with the difference between any two items of historical state data and the corresponding weights; the weight is the product of the influence sensitivity and the importance degree of the corresponding item sequence; the importance degree is: ; is the th item of data corresponding to the th driving, , are respectively the average value and the range of the th item sequence, is the safety index at the th driving, , are the range and the average value of the safety index, n is the number of historical drives of the vehicle; the safety index characterizes the speed fluctuation degree of the vehicle during each drive, and the safety index is: ; represents the range of the speed sequence at the th driving, represents the th speed value in the speed sequence at the th driving, represents the average value of the speed sequence at the th driving, is the exponential function with the natural constant e as the base, is the total number of speeds in the speed sequence at the th driving; the influence sensitivity is: ; is the number of data in the th item sequence, , are respectively the th and the th data in the th item sequence, , , are respectively the safety index, the abnormality degree of the th data and the safety index of the th data in the th item sequence; Set the threshold is: ; is the set multiple, is the average of the Euclidean distances between all historical status data and the cluster centers in their corresponding clusters.

2. The vehicle safety management method based on cloud computing according to claim 1, characterized in that, The status data includes vehicle fuel quantity, average tire pressure, average vehicle driving speed, and brake usage frequency.

3. A vehicle safety management method based on cloud computing according to claim 1, characterized in that, The said distance is: ； where N is the total number of historical status data, is the th item of data corresponding to the th driving, is the th item of data corresponding to the th driving, is the weight of the th item of data.

4. The vehicle safety management method based on cloud computing according to claim 1, wherein The degree of abnormality is obtained by performing anomaly detection on the safety index data by using the LOF algorithm; the safety index data includes multiple safety indicators composed of each item of data and the corresponding safety index.

5. The vehicle safety management method based on cloud computing according to claim 1, characterized in that, The method for normalizing the status data includes maximum-minimum normalization.

6. A vehicle safety management system based on cloud computing, characterized in that, Including: A processor; A memory that stores computer instructions for vehicle safety management based on cloud computing. When the computer instructions are run by the processor, the system executes a vehicle safety management method based on cloud computing according to any one of claims 1-5.

Images