A multi-level vehicle data monitoring system

Abstract

The present application relates to the field of data monitoring, especially to a multi-level vehicle data monitoring system, comprising: a processing unit for caching and transmitting subway data; a node analysis unit for determining the node category corresponding to each processing node according to the overlapping coefficient and the number of effective lines corresponding to each processing node; a work analysis unit for determining the working state of the node according to the channel utilization rate and the number of related nodes corresponding to the independent node, and setting the information retention mode as dynamic retention or fixed time retention according to the working state; in dynamic retention, determining the preset adjustment time or selecting the compensation node for information retention according to the characteristic abnormal reference value and the conflict influence reference value; a task adjustment unit for determining the node call state according to the call frequency characteristic value of each processing node, and determining the task adjustment mode of the processing node according to the node call state; improve the transmission efficiency of vehicle monitoring data.

Description

Technical Field

[0001] This invention relates to the field of data monitoring, and more particularly to a multi-level vehicle data monitoring system. Background Technology

[0002] With the expansion of urban rail transit and the continuous improvement of intelligent requirements, data monitoring systems are increasingly being applied in areas such as train safety management, fault early warning, and operation optimization. Traditional monitoring systems typically employ decentralized and independent deployments, leading to significant problems such as poor data sharing efficiency and long data backup cycles among different departments and levels within the subway system. Therefore, multi-level collaborative monitoring systems are gradually being applied in the subway field, linking data monitoring terminals with multiple levels to enable each level to quickly and timely obtain the required monitoring data and perform corresponding tasks, thereby improving the efficiency of subway monitoring and management, as well as information sharing and work efficiency among relevant parties. However, with the increase in levels and the uncertainty of information needs, the existing single data transmission and storage methods are insufficient to meet dynamic data requirements. Therefore, how to achieve effective subway data monitoring and storage is a technical problem that urgently needs to be solved by those skilled in the art.

[0003] Chinese Patent Publication No. CN110958313A discloses a method for vehicle-to-ground data transmission of fully automated metro vehicles, including: the train network control system is an MVB bus, and the central server is in Ethernet form; the MVB network protocol on the train is converted into the Ethernet protocol for transmission through a gateway (GW) device; the topology of the vehicle-to-ground transmission channel is redundant, so even if one channel is interrupted, it will not affect the vehicle-to-ground data transmission. The ground server is equipped with decision logic to judge the validity of data, and monitors the life signal of data packets transmitted from the wireless channel in real time. When the life signal of the trusted channel is abnormal, it switches to monitoring the data of the redundant channel. The two network redundancy forms set in this scheme do not affect the real-time transmission of data even if one channel is completely interrupted, and the Ethernet transmission protocol meets the bandwidth requirements of the data, and all vehicle data can be transmitted to the center. It can be seen that the above technical solution has the following problems: it can only switch the data transmission channel, but does not consider that the sudden increase in channel load caused by the instability of data volume can make it difficult for a single transmission mode to meet the dynamic scene transmission requirements, resulting in poor monitoring data transmission efficiency. Summary of the Invention

[0004] To address this issue, the present invention provides a multi-level vehicle data monitoring system to overcome the problem that the single transmission method in the prior art is insufficient to meet the dynamic scene transmission requirements, resulting in poor monitoring data transmission efficiency.

[0005] To achieve the above objectives, the present invention provides a multi-level vehicle data monitoring system, comprising:

[0006] The processing unit includes several processing nodes for caching and transmitting subway data;

[0007] The node analysis unit, which is connected to the processing unit, is used to determine whether each processing node is a compensation node or an independent node based on the overlap coefficient and the number of effective lines.

[0008] The work analysis unit is connected to the processing unit and the node analysis unit respectively. It is used to determine the working status of the node based on the channel utilization rate of the independent node and the number of related nodes, and to set the information retention method as dynamic retention or fixed duration retention according to the working status of the node.

[0009] During dynamic retention, the preset adjustment duration or the selection of compensation nodes is determined based on the reference values ​​of feature anomalies and conflict impacts.

[0010] The task adjustment unit, connected to the processing unit, node analysis unit, and job analysis unit, determines the node retrieval status based on the retrieval frequency representation value of each processing node, and determines whether to adjust the cache for the processing node based on the node retrieval status. Furthermore, the cache adjustment method is determined based on the retrieval level diversity and data retrieval dispersion of each dataset.

[0011] Furthermore, the node analysis unit determines the node category corresponding to each processing node based on the overlap coefficient and the number of effective lines corresponding to each processing node.

[0012] If the overlap coefficient is greater than the preset overlap coefficient and the number of effective lines is less than or equal to the preset number of effective lines, then the node category is a compensation node.

[0013] If the overlap coefficient is less than or equal to the preset overlap coefficient or the number of valid lines is greater than the preset number of valid lines, then the node category is an independent node.

[0014] Furthermore, the overlap coefficient is determined based on the overlap area corresponding to the processing node and the number of overlapping lines;

[0015] The overlap coefficient is positively correlated with the overlap area corresponding to the processing node and the number of overlapping lines.

[0016] Furthermore, the work analysis unit determines the working status of a node based on the channel utilization rate of the independent node and the number of related nodes, and sets the information retention method according to the working status.

[0017] If the node's working status is that the channel utilization rate is greater than the preset channel utilization rate or the number of related nodes is greater than the preset number of related nodes, the information retention method is dynamic retention;

[0018] If a node's working status is such that its channel utilization is less than or equal to the preset channel utilization and the number of related nodes is less than or equal to the preset number of related nodes, the information retention method is fixed-duration retention.

[0019] Furthermore, the work analysis unit responds to the dynamic retention conditions and obtains the characteristic anomaly reference value and conflict impact reference value corresponding to the independent node;

[0020] If the abnormal feature reference value is less than the preset abnormal feature reference value or the conflict impact reference value is greater than or equal to the preset conflict impact reference value, the preset adjustment time is determined based on the feature value.

[0021] If the feature anomaly reference value is greater than or equal to the preset feature anomaly reference value and the conflict impact reference value is less than the preset conflict impact reference value, it is determined to select a compensation node to retain information.

[0022] The dynamic retention condition is to determine that the information retention method is dynamic retention.

[0023] Furthermore, the work analysis unit responds to the preset compensation conditions, determines the compensation reference value based on the stability coefficient and radiation coefficient of the compensation node, and selects the compensation node whose corresponding compensation reference value is greater than the preset compensation reference value for information retention.

[0024] The compensation reference value is positively correlated with both the stability coefficient and the radiation coefficient.

[0025] The preset compensation conditions are used by the work analysis unit to determine and select compensation nodes for information retention.

[0026] Furthermore, the task adjustment unit determines the node retrieval status based on the retrieval frequency representation value of each processing node, and determines whether to perform cache adjustment based on the node retrieval status;

[0027] If the node retrieval status of the processing node is such that the retrieval frequency characterization value is greater than the preset retrieval frequency characterization value, the task adjustment unit determines to adjust the cache.

[0028] Furthermore, the task adjustment unit responds to the first preset adjustment condition and performs the task adjustment method of cache adjustment, determining the cache adjustment method based on the retrieval level diversity and data retrieval dispersion of each dataset.

[0029] If the retrieval level diversity is greater than the preset retrieval level diversity and the data retrieval dispersion is less than or equal to the preset data retrieval dispersion, the task adjustment unit determines that the cache adjustment method is to be handled with attention.

[0030] If the retrieval level diversity is less than or equal to the preset retrieval level diversity or the data retrieval dispersion is greater than the preset data retrieval dispersion, the task adjustment unit determines that the cache adjustment method is cyclic processing.

[0031] The first preset adjustment condition is that the task adjustment unit determines the task adjustment method to be cache adjustment.

[0032] Furthermore, the task adjustment unit executes a cache adjustment method for loop processing, and determines the loop iteration duration corresponding to the dataset based on the effective difference;

[0033] The cycle duration and the effective difference are negatively correlated.

[0034] Furthermore, the task adjustment unit implements a cache adjustment method for handling concerns, caching datasets whose task retrieval frequency is greater than the average task retrieval frequency.

[0035] Compared with the prior art, the beneficial effect of the present invention is that, in the technical solution of the present invention, the node analysis unit determines the node category corresponding to each processing node based on the overlap coefficient and the number of effective lines corresponding to each processing node. The overlap coefficient reflects the degree of overlap between processing nodes, and the number of effective lines reflects the distribution of operating lines within the coverage area of ​​the processing node, thereby screening out compensation nodes with certain compensation capabilities, preparing for the setting of subsequent information retention methods, and thus improving data processing efficiency.

[0036] Furthermore, in the technical solution of the present invention, the working analysis unit determines the working status of the node based on the channel utilization rate of the independent node and the number of related nodes. The working status reflects the information processing capability and the probability of being compensated of the independent node, and different information retention methods are set accordingly. The setting of information retention methods is conducive to the data load balancing of each processing node, which in turn is beneficial to the subsequent dynamic adjustment of the data load of the processing node, and further improves the data monitoring and transmission efficiency of the present invention.

[0037] Furthermore, in the technical solution of this invention, the work analysis unit determines and selects a compensation node for information retention based on the characteristic anomaly value and the conflict impact reference value corresponding to the independent node, or sets a preset adjustment time for data at the independent node based on the characteristic value. The characteristic anomaly value reflects the degree of anomaly of the corresponding subway data, and the conflict impact reference value reflects the compensation capability of the compensation node. This avoids the difficulty of adapting to the actual situation with a single information retention method, improves the scenario adaptability of the information retention method, and thus improves the efficiency of monitoring data transmission and management.

[0038] Furthermore, in the technical solution of the present invention, the working analysis unit determines the compensation reference value based on the stability coefficient and radiation coefficient of the compensation node, and selects the compensation node with the largest corresponding compensation reference value for information retention; the stability coefficient reflects the data processing capability of the compensation node, and the radiation coefficient reflects the degree of influence of changes in dynamic data around the compensation node, making the determination of the compensation reference value more accurate, thereby improving the selection accuracy of the compensation node.

[0039] Furthermore, in the technical solution of the present invention, the task adjustment unit determines the node retrieval status based on the retrieval frequency characterization value of each processing node, and determines whether to perform cache adjustment based on the node retrieval status;

[0040] By using the retrieval frequency characterization value to reflect the frequency of retrieval of information data corresponding to each processing node, and determining whether to adjust the cache based on the node retrieval status, the data caching effect can be improved, resulting in better utilization of cached data. Attached Figure Description

[0041] Figure 1 This is a unit connection diagram of a multi-level vehicle data monitoring system according to the present invention;

[0042] Figure 2 This is a flowchart illustrating how the present invention determines the corresponding node category based on the overlap coefficient and the number of effective lines for each processing node.

[0043] Figure 3 This is a flowchart illustrating how the present invention determines the method for retaining setting information based on the working status of a node. Detailed Implementation

[0044] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0045] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0046] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0047] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0048] Please see Figures 1 to 3 As shown, the present invention provides a multi-level vehicle data monitoring system, comprising:

[0049] The processing unit includes several processing nodes for caching and transmitting subway data;

[0050] The node analysis unit, which is connected to the processing unit, is used to determine whether each processing node is a compensation node or an independent node based on the overlap coefficient and the number of effective lines.

[0051] The work analysis unit is connected to the processing unit and the node analysis unit respectively. It is used to determine the working status of the node based on the channel utilization rate of the independent node and the number of related nodes, and to set the information retention method as dynamic retention or fixed duration retention according to the working status of the node.

[0052] During dynamic retention, the preset adjustment duration or the selection of compensation nodes is determined based on the reference values ​​of feature anomalies and conflict impacts.

[0053] The task adjustment unit, connected to the processing unit, node analysis unit, and job analysis unit, determines the node retrieval status based on the retrieval frequency representation value of each processing node, and determines whether to adjust the cache for the processing node based on the node retrieval status. Furthermore, the cache adjustment method is determined based on the retrieval level diversity and data retrieval dispersion of each dataset.

[0054] In this invention, the system is applied to a subway management platform. The system is connected to both a retrieval terminal and a database. A new task is a retrieval task requested by the retrieval terminal to retrieve a dataset from the database. This dataset is a collection of data generated during subway operation, including but not limited to subway monitoring data, subway fault data, subway operation data, and subway staff data. Subway monitoring data includes subway surveillance images, subway station surveillance videos, and turnstile data. Subway fault data includes fault time and fault type. Subway operation data includes subway operating speed, traction level, bus voltage, bus current, actual electric braking force traction energy consumption, total energy consumption, total air pressure, and air compressor operating time. Subway staff data includes staff duty rosters. It is understood that the retrieval terminal includes several departments, such as a platform application development service department, an operation and maintenance management department, and an operations department, to retrieve the dataset for model development, operation and maintenance data analysis, and operation management. This is readily understood by those skilled in the art and will not be elaborated upon here.

[0055] The processing node is an edge node or base station with data processing and data transmission functions. New tasks are transmitted and cached by the processing node. The transmission also includes the processing node transmitting the subway data within its coverage area to the database and management terminal. In the management terminal, subway management personnel can process the data and identify whether the data has any risks. This is something that is easy for those skilled in the art to understand and will not be elaborated here.

[0056] Specifically, the node analysis unit determines the node category of each processing node based on the overlap coefficient and the number of effective lines corresponding to each processing node.

[0057] If the overlap coefficient is greater than the preset overlap coefficient and the number of effective lines is less than or equal to the preset number of effective lines, then the node category is a compensation node.

[0058] If the overlap coefficient is less than or equal to the preset overlap coefficient or the number of valid lines is greater than the preset number of valid lines, then the node category is an independent node.

[0059] For a single processing node, the number of effective lines is the number of subway lines within the coverage area of ​​the processing node whose line length is greater than a preset line length. Here, the line length is the route length of the subway line within the coverage area of ​​the processing node, and the coverage area is the maximum actual area where the processing node can transmit information.

[0060] This invention characterizes the processing capacity of processing nodes and the degree of correlation between processing nodes through the overlap coefficient and the number of effective lines. The greater the user's demand for the compensation capacity of the compensation nodes, the smaller the preset number of effective lines and the larger the preset overlap coefficient. A method for determining these values ​​is provided: the overlap coefficient and the number of effective lines in historical records that meet user needs are detected, and outliers are removed. The average of the overlap coefficient and the number of effective lines after removing outliers is recorded as the preset overlap coefficient and the preset number of effective lines, respectively. Methods for removing outliers include, but are not limited to, the 3σ criterion or the IQR method. The users of this invention are subway management personnel using this system.

[0061] This invention utilizes historical records, each of which includes at least the overlap coefficient, number of effective lines, overlap area, number of overlapping lines, channel utilization, number of related nodes, characteristic anomaly reference value, collision impact reference value, retrieval frequency characterization value, retrieval level diversity, and data retrieval dispersion. Furthermore, each historical record is associated with a qualification flag, which indicates whether the historical record meets user requirements. Whether a historical record meets user requirements can be determined, but is not limited to, based on the processing efficiency of the processing node. How to determine whether a historical record meets user requirements based on user-defined processing efficiency indicators (e.g., average task processing speed) is a topic already understood by those skilled in the art and is not limited here.

[0062] Specifically, the overlap coefficient is determined based on the overlap area corresponding to the processing node and the number of overlapping lines;

[0063] The overlap coefficient is positively correlated with the overlap area corresponding to the processing node and the number of overlapping lines.

[0064] For a single processing node, the corresponding overlap coefficient = the sum of the overlapping areas of the coverage areas of the processing node and other processing nodes / preset overlapping area + the sum of the number of overlapping lines of the processing node and other processing nodes / preset number of overlapping lines. Here, the area of ​​the overlapping area of ​​the corresponding coverage areas between two processing nodes is recorded as the overlapping area, and the number of subway lines within the overlapping area is recorded as the number of overlapping lines.

[0065] Users can adaptively set the preset overlap area and preset number of overlapping lines according to the actual application scenario. It is understood that the greater the user's demand for the compensation capability of the compensation node, the larger the preset overlap area and the larger the preset number of overlapping lines. This invention provides a method for setting these values: extract the corresponding overlap area and number of overlapping lines from the historical records that meet the user's needs, filter out outliers in the overlap area and number of overlapping lines respectively, and record the average values ​​of the overlap area and number of overlapping lines after removing outliers as the preset overlap area and preset number of overlapping lines respectively.

[0066] Specifically, the work analysis unit determines the working status of a node based on the channel utilization rate of the independent node and the number of related nodes, and sets the information retention method according to the working status.

[0067] If the node's working status is that the channel utilization rate is greater than the preset channel utilization rate or the number of related nodes is greater than the preset number of related nodes, the information retention method is dynamic retention;

[0068] If a node's working status is such that its channel utilization is less than or equal to the preset channel utilization and the number of related nodes is less than or equal to the preset number of related nodes, the information retention method is fixed-duration retention.

[0069] Channel utilization rate for an independent node = Current occupied bandwidth of the independent node / Total available bandwidth.

[0070] For a single independent node, the number of its corresponding compensation nodes is recorded as the number of related nodes for that independent node.

[0071] When the information retention method is fixed-duration retention, the dataset corresponding to each new task is deleted after being cached on an independent node for a fixed duration. The value of the fixed duration can be set by the user according to the actual application requirements. It can be understood that the fixed duration is shorter for data with higher security requirements and longer for data with lower security requirements. In addition, it can be combined with channel utilization. The lower the channel utilization, the longer the fixed duration. Security requirements and channel utilization are easy to understand for those skilled in the art and will not be elaborated here.

[0072] Users can adaptively set the preset channel utilization rate and preset number of related nodes according to the actual application scenario. It is understood that the higher the user's requirements for the processing efficiency and stability of independent nodes, the lower the preset channel utilization rate and the higher the preset number of related nodes. This invention provides a method for setting the preset channel utilization rate and preset number of related nodes, which extracts the channel utilization rate and the number of related nodes from the historical records that meet the user's needs, filters out outliers, and records the average value of the channel utilization rate and the number of related nodes after removing outliers as the preset channel utilization rate and the preset number of related nodes, respectively.

[0073] Specifically, the work analysis unit responds to the dynamic retention conditions and obtains the feature anomaly reference value and conflict impact reference value corresponding to the independent node;

[0074] If the abnormal feature reference value is less than the preset abnormal feature reference value or the conflict impact reference value is greater than or equal to the preset conflict impact reference value, the preset adjustment time is determined based on the feature value.

[0075] If the feature anomaly reference value is greater than or equal to the preset feature anomaly reference value and the conflict impact reference value is less than the preset conflict impact reference value, it is determined to select a compensation node to retain information.

[0076] The dynamic retention condition is to determine that the information retention method is dynamic retention.

[0077] When setting the preset adjustment duration based on the feature value, the dataset corresponding to the newly added task is cached by an independent node, and the cache duration = fixed duration - preset adjustment duration. The preset adjustment duration is positively correlated with the feature value, and the feature value is the maximum channel utilization corresponding to the preset monitoring duration before the current time.

[0078] The method for confirming the characteristic anomaly reference value is as follows: obtain the channel utilization curve corresponding to the preset monitoring duration before the current time. The channel utilization curve is presented in the form of a two-dimensional coordinate system, with time on the horizontal axis and the channel utilization rate corresponding to an independent node on the vertical axis. Randomly select several points on the channel utilization curve, obtain the vertical coordinate corresponding to each point, and calculate the characteristic anomaly reference value S. The formula for calculating S is:

[0079]

[0080] Where Si is the ordinate of the i-th point, S0 is the average ordinate of all points, i = 1, 2, 3, ..., n, and n is the total number of points obtained. The order of the points is randomly set and does not affect the calculation results. It can be understood that the greater the user's requirement for the accuracy of the feature anomaly reference value, the longer the preset monitoring time and the larger n will be.

[0081] Conflict impact reference value = total number of invalid inflection points / total number of inflection points of independent nodes;

[0082] The process involves acquiring the channel utilization change curves of independent nodes and their related nodes corresponding to a preset monitoring duration prior to the current time. Peak points are extracted from each channel utilization change curve. A peak point is considered an inflection point, and the ordinates of any two adjacent points are smaller than the ordinate of the inflection point itself. Conflict analysis is performed on each inflection point corresponding to an independent node. When performing conflict analysis on a single inflection point, this inflection point is designated as the target inflection point. The inflection point closest to the target inflection point for each related node is detected. If the distance between the inflection point and the target inflection point is less than a preset distance, it is recorded as a conflict inflection point. The number of conflict inflection points corresponding to the target inflection point is calculated. If the quantity is less than the preset number of inflection points, the target inflection point is a valid inflection point. If the number of conflicting inflection points corresponding to the target inflection point is greater than or equal to the preset number of inflection points, the target inflection point is an invalid inflection point. The distance between inflection points is the absolute value of the difference between the x-coordinates of the two inflection points. The preset number of inflection points and the preset distance can be set by the user according to the actual scenario. The greater the user's demand for the compensation capability of the compensation node and the greater the demand for the processing efficiency of the independent node, the smaller the preset number of inflection points and the larger the preset distance. One set of values ​​is provided: the preset number of inflection points is 50% of the number of related nodes, and the preset distance is 1 minute.

[0083] The preset reference values ​​for feature anomalies and the preset reference values ​​for conflict impacts can be adaptively set by the user according to the actual application scenario. It can be understood that the greater the user's demand for the processing efficiency of independent nodes, the smaller the preset reference value for feature anomalies; the greater the user's demand for the compensation effect and compensation capability of compensation nodes, the smaller the preset reference value for conflict impacts. This invention provides a method for setting the preset reference values ​​for feature anomalies and the preset reference values ​​for conflict impacts. The method extracts the reference values ​​for feature anomalies and the reference values ​​for conflict impacts from the historical records that meet the user's needs, filters out the outliers, and records the average values ​​of the reference values ​​for feature anomalies and the reference values ​​for conflict impacts after removing the outliers as the preset reference values ​​for feature anomalies and the preset reference values ​​for conflict impacts, respectively.

[0084] Specifically, the work analysis unit responds to the preset compensation conditions, determines the compensation reference value based on the stability coefficient and radiation coefficient of the compensation node, and selects the compensation node whose corresponding compensation reference value is greater than the preset compensation reference value for information retention.

[0085] The compensation reference value is positively correlated with both the stability coefficient and the radiation coefficient.

[0086] The preset compensation conditions are used by the work analysis unit to determine and select compensation nodes for information retention.

[0087] For a single compensation node, the stability coefficient is determined by obtaining the CPU usage difference of the compensation node corresponding to the preset monitoring time before the current time. The CPU usage difference = maximum CPU usage value - minimum CPU usage value, and the stability coefficient = 1 / CPU usage difference.

[0088] For a single compensation node, its corresponding radiation coefficient is the number of independent nodes whose overlapping area is greater than the preset overlapping area.

[0089] The compensation reference value = A1 × stability coefficient + A2 × radiation coefficient; where A1 is the first weighting coefficient, A2 is the second weighting coefficient, and A1 + A2 = 1. The values ​​of A1 and A2 can be adaptively set by the user according to the actual application scenario. It can be understood that the greater the contribution of the stability coefficient to the compensation reference value, the larger A1 will be, and the same applies to the value of A2. This invention provides a value of A1 and A2 of A1 = 0.6 and A2 = 0.4.

[0090] The compensation node with a corresponding compensation reference value greater than the preset compensation reference value is selected for information retention. That is, the dataset of the new task corresponding to the independent node is sent to the compensation node for caching. The caching time is a fixed time. If there is a new task that needs to retrieve the dataset later, the new task will be directly assigned to the compensation node.

[0091] Specifically, the task adjustment unit determines the node retrieval status based on the retrieval frequency representation value of each processing node, and determines whether to perform cache adjustment based on the node retrieval status.

[0092] If the node retrieval status of the processing node is such that the retrieval frequency characterization value is greater than the preset retrieval frequency characterization value, the task adjustment unit determines to adjust the cache.

[0093] If the node retrieval status of the processing node is such that the retrieval frequency representation value is less than or equal to the preset retrieval frequency representation value, the task adjustment unit determines that no cache adjustment is required.

[0094] The retrieval frequency characterization value is the number of times information is retrieved through this processing node within a preset monitoring period before the current time. The user can adaptively set the value of the preset retrieval frequency characterization value according to the actual application scenario. The greater the user's need for the stability of the processing node, the smaller the value of the preset retrieval frequency characterization value. This invention provides a method for setting the preset retrieval frequency characterization value, which extracts the retrieval frequency characterization value from the historical records that meet the user's needs, filters out outliers, and records the average value of the retrieval frequency characterization value after removing outliers as the preset retrieval frequency characterization value.

[0095] Specifically, the task adjustment unit responds to the first preset adjustment condition and performs the task adjustment method of cache adjustment, determining the cache adjustment method based on the retrieval level diversity and data retrieval dispersion of each dataset.

[0096] If the retrieval level diversity is greater than the preset retrieval level diversity and the data retrieval dispersion is less than or equal to the preset data retrieval dispersion, the task adjustment unit determines that the cache adjustment method is to be handled with attention.

[0097] If the retrieval level diversity is less than or equal to the preset retrieval level diversity or the data retrieval dispersion is greater than the preset data retrieval dispersion, the task adjustment unit determines that the cache adjustment method is cyclic processing.

[0098] The first preset adjustment condition is that the task adjustment unit determines the task adjustment method to be cache adjustment.

[0099] The retrieval hierarchy diversity is the number of departments that retrieve information through this processing node within the preset monitoring period before the current time. It can be understood that the data retrieval dispersion is equal to the number of different datasets retrieved through this processing node within the preset monitoring period before the current time. For example, if two retrievals correspond to the same dataset, the number is 1.

[0100] The preset values ​​for retrieval level diversity and preset data retrieval dispersion can be adaptively set by the user according to the actual application scenario. It can be understood that the greater the retrieval level diversity and the smaller the data retrieval dispersion, the more concentrated the types of data retrieved. Correspondingly, the caching adjustment method with higher attention should be selected to avoid cache redundancy caused by caching uninteresting datasets. This invention provides a method for setting the preset values ​​for retrieval level diversity and preset data retrieval dispersion. The retrieval level diversity and data retrieval dispersion that meet the user's needs are extracted from the historical records, outliers are filtered out, and the average values ​​of the retrieval level diversity and data retrieval dispersion after removing outliers are recorded as the preset retrieval level diversity and preset data retrieval dispersion, respectively.

[0101] Specifically, the task adjustment unit executes a cache adjustment method that performs loop processing, and determines the loop iteration duration corresponding to the dataset based on the effective difference;

[0102] The cycle duration and the effective difference are negatively correlated.

[0103] The effective difference is the larger of the retrieval level diversity difference and the data retrieval dispersion difference; where, retrieval level diversity difference = preset retrieval level diversity difference - retrieval level diversity difference, and data retrieval dispersion difference = preset data retrieval dispersion - data retrieval dispersion.

[0104] Cycle iteration duration = base iteration duration - effective difference × K, where the cycle iteration duration is an integer rounded up, and K is a conversion coefficient. The value of K can be set by the user according to their needs. The larger the value of K, the greater the importance attached to the diversity of retrieval levels and the dispersion of data retrieval. One possible value for K is K = 0.2. The result of effective difference × K is assigned to the unit min. The base iteration duration is set by the user and is not specifically limited. It can be set according to the dispersion of data retrieval. The larger the dispersion of data retrieval, the smaller the base iteration duration. In this invention, the base iteration duration is 5 min. The result of effective difference × K has a corresponding maximum value, which is 2 min. If the result of effective difference × K exceeds the maximum value, it is directly recorded as the maximum value.

[0105] In the cyclic processing cache adjustment method, for a single processing node, the dataset corresponding to the retrieval task within the preset monitoring time before the current time of the processing node is cached in a random order on the processing node. The cache duration is the cycle iteration duration. The dataset of the retrieval task can be cached simultaneously. This is something that is easy for those skilled in the art to understand and will not be elaborated here.

[0106] Specifically, the task adjustment unit implements a cache adjustment method for handling concerns, which caches datasets whose task retrieval frequency is greater than the average task retrieval frequency, and does not cache datasets whose task retrieval frequency is less than or equal to the preset average task retrieval frequency.

[0107] The task retrieval frequency of the dataset is the number of times the dataset is retrieved within the preset monitoring period before the current time. For a single processing node, the average task retrieval frequency is the average number of times each dataset is retrieved within the preset monitoring period before the current time.

[0108] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A multi-level vehicle data monitoring system, characterized in that, include: The processing unit includes several processing nodes for caching and transmitting subway data; The node analysis unit, which is connected to the processing unit, is used to determine whether each processing node is a compensation node or an independent node based on the overlap coefficient and the number of effective lines. The work analysis unit is connected to the processing unit and the node analysis unit respectively. It is used to determine the working status of the node based on the channel utilization rate of the independent node and the number of related nodes, and to set the information retention method as dynamic retention or fixed duration retention according to the working status of the node. During dynamic retention, the preset adjustment duration or the selection of compensation nodes is determined based on the reference values ​​of feature anomalies and conflict impacts. The task adjustment unit is connected to the processing unit, the node analysis unit, and the work analysis unit respectively. It is used to determine the node retrieval status based on the retrieval frequency characterization value of each processing node, and to determine whether to perform cache adjustment for the processing node based on the node retrieval status. In the cache adjustment, the cache adjustment method is determined based on the retrieval level diversity and data retrieval dispersion of each dataset. For a single processing node, the corresponding overlap coefficient = the sum of the overlapping areas of the coverage areas of the processing node and other processing nodes / preset overlapping area + the sum of the number of overlapping lines of the processing node and other processing nodes / preset number of overlapping lines. Here, the area of ​​the overlapping area of ​​the corresponding coverage areas between two processing nodes is recorded as the overlapping area, and the number of subway lines within the overlapping area is recorded as the number of overlapping lines. For a single processing node, the number of effective lines is the number of subway lines within the coverage area of ​​the processing node whose line length is greater than the preset line length. The line length is the route length of the subway line within the coverage area of ​​the processing node, and the coverage area is the maximum actual area where the processing node can transmit information. Channel utilization rate for an independent node = Current occupied bandwidth of the independent node / Total available bandwidth; For a single independent node, the number of its corresponding compensation nodes is recorded as the number of related nodes of that independent node; The method for confirming the characteristic anomaly reference value is as follows: obtain the channel utilization curve corresponding to the preset monitoring duration before the current time. The channel utilization curve is presented in the form of a two-dimensional coordinate system, with time on the horizontal axis and the channel utilization rate corresponding to an independent node on the vertical axis. Randomly select several points on the channel utilization curve, obtain the vertical coordinate corresponding to each point, and calculate the characteristic anomaly reference value S. The formula for calculating S is: ； Where Si is the ordinate of the i-th point, S0 is the average ordinate of each point, i = 1, 2, 3, ..., n, where n is the total number of points obtained, and the conflict impact reference value = the total number of invalid inflection points / the total number of inflection points of independent nodes. The retrieval frequency is characterized by the number of times information is retrieved through this processing node within the preset monitoring period before the current time. The retrieval hierarchy diversity is the number of departments that retrieve information through this processing node within the preset monitoring period before the current time. The data retrieval dispersion is the number of different datasets retrieved through this processing node within the preset monitoring period before the current time.

2. The multi-level vehicle data monitoring system according to claim 1, characterized in that... The node analysis unit determines the node category of each processing node based on the overlap coefficient and the number of effective lines. If the overlap coefficient is greater than the preset overlap coefficient and the number of effective lines is less than or equal to the preset number of effective lines, then the node category is a compensation node. If the overlap coefficient is less than or equal to the preset overlap coefficient or the number of valid lines is greater than the preset number of valid lines, then the node category is an independent node.

3. The multi-level vehicle data monitoring system according to claim 2, characterized in that, The overlap coefficient is determined based on the overlap area corresponding to the processing node and the number of overlapping lines; The overlap coefficient is positively correlated with the overlap area corresponding to the processing node and the number of overlapping lines.

4. The multi-level vehicle data monitoring system according to claim 3, characterized in that, The work analysis unit determines the working status of a node based on the channel utilization rate of the independent node and the number of related nodes, and sets the information retention method according to the working status. If the node's working status is that the channel utilization rate is greater than the preset channel utilization rate or the number of related nodes is greater than the preset number of related nodes, the information retention method is dynamic retention; If a node's working status is such that its channel utilization is less than or equal to the preset channel utilization and the number of related nodes is less than or equal to the preset number of related nodes, the information retention method is fixed-duration retention.

5. The multi-level vehicle data monitoring system according to claim 4, characterized in that, The working analysis unit responds to dynamic retention conditions and obtains the feature anomaly reference value and conflict impact reference value corresponding to the independent node; If the abnormal feature reference value is less than the preset abnormal feature reference value or the conflict impact reference value is greater than or equal to the preset conflict impact reference value, the preset adjustment time is determined based on the feature value. If the feature anomaly reference value is greater than or equal to the preset feature anomaly reference value and the conflict impact reference value is less than the preset conflict impact reference value, it is determined to select a compensation node to retain information. The dynamic retention condition is that the information retention method is determined to be dynamic retention; The characteristic value is the maximum channel utilization rate corresponding to the preset monitoring duration before the current time.

6. The multi-level vehicle data monitoring system according to claim 5, characterized in that, The work analysis unit responds to the preset compensation conditions, determines the compensation reference value based on the stability coefficient and radiation coefficient of the compensation node, and selects the compensation node whose corresponding compensation reference value is greater than the preset compensation reference value for information retention. The compensation reference value is positively correlated with both the stability coefficient and the radiation coefficient. The preset compensation conditions are determined by the work analysis unit to select compensation nodes for information retention. For a single compensation node, the corresponding stability coefficient is determined by obtaining the CPU usage difference of the compensation node corresponding to the preset monitoring time before the current time. The CPU usage difference = maximum CPU usage value - minimum CPU usage value, and the stability coefficient = 1 / CPU usage difference. For a single compensation node, its corresponding radiation coefficient is the number of independent nodes whose overlapping area is greater than the preset overlapping area.

7. The multi-level vehicle data monitoring system according to claim 6, characterized in that, The task adjustment unit determines the node retrieval status based on the retrieval frequency representation value of each processing node, and determines whether to adjust the cache for the processing node based on the node retrieval status. If the node retrieval status of the processing node is such that the retrieval frequency characterization value is greater than the preset retrieval frequency characterization value, the task adjustment unit determines to adjust the cache.

8. The multi-level vehicle data monitoring system according to claim 7, characterized in that, The task adjustment unit responds to the first preset adjustment condition, performs cache adjustment, and determines the cache adjustment method based on the retrieval level diversity and data retrieval dispersion of each dataset. If the retrieval level diversity is greater than the preset retrieval level diversity and the data retrieval dispersion is less than or equal to the preset data retrieval dispersion, the task adjustment unit determines that the cache adjustment method is to be handled with attention. If the retrieval level diversity is less than or equal to the preset retrieval level diversity or the data retrieval dispersion is greater than the preset data retrieval dispersion, the task adjustment unit determines that the cache adjustment method is loop processing. The first preset adjustment condition is that the task adjustment unit determines that the task adjustment method is cache adjustment; The task adjustment unit executes a cache adjustment method for loop processing, and determines the loop iteration duration corresponding to the dataset based on the effective difference. The cycle iteration duration and the effective difference are negatively correlated; The task adjustment unit implements a cache adjustment method for handling concerns, caching datasets whose task retrieval frequency is greater than the average task retrieval frequency.

Images