A rail transit accident monitoring system and method based on wireless communication

Abstract

The present invention discloses a rail transit accident monitoring system and method based on wireless communication, and relates to the technical field of rail transit accident monitoring. The system of the present invention comprises: a data acquisition and processing module, a state evaluation module, a data analysis module, a real-time monitoring module and an accident judgment and response module; the data acquisition and processing module collects historical data of nodes in the monitoring area, obtains log data, and marks the historical data; the state evaluation module evaluates the operating status of the wireless sensor node and identifies the key nodes; the data analysis module analyzes the trend curve of the historical data of the key nodes, extracts the trend characteristics, and establishes a mapping relationship between the trend characteristics and the trend curve; the real-time monitoring module obtains the accident notification and real-time data of the key nodes in real time, generates a trend curve and extracts the real-time trend characteristics; the accident judgment and response module calculates the deviation degree index, judges the accuracy of the accident notification and outputs the relevant notification information.

Description

Technical Field

[0001] The present invention relates to the technical field of rail transit accident monitoring, and in particular to a rail transit accident monitoring system and method based on wireless communication. Background Art

[0002] With the rapid development of urban rail transit, rail transit has played an important role in improving urban travel efficiency and alleviating traffic pressure. However, the safety and reliability of rail transit systems are still urgent issues to be solved. Accident monitoring systems are a key component to ensure rail transit safety. Traditional rail transit accident monitoring systems usually rely on wired communication technologies such as fiber optic sensors and cables, which improves the stability of the system to a certain extent, but also brings many inconveniences.

[0003] The rise of wireless communication technology has provided new possibilities for upgrading rail transit accident monitoring systems. Through wireless sensor networks, real-time data collection, transmission and processing can be achieved, thereby improving the flexibility and response speed of accident monitoring. Wireless communication can not only reduce wiring costs, but also enable more efficient data transmission in some complex environments (such as tunnels, bridges, etc.). However, existing wireless communication technologies still face many challenges in practical applications, which affects their promotion in rail transit accident monitoring.

[0004] In urban rail transit, especially in subway systems, the surrounding environment is complex and electromagnetic interference is serious. Existing wireless communication technologies are susceptible to electromagnetic interference during transmission, resulting in data loss or misjudgment. For example, when a subway line is in operation, due to interference from nearby high-voltage power lines, the accuracy of the wireless sensor node data in the monitoring area cannot be guaranteed. Therefore, if the accuracy of the wireless sensor node data is not determined, and only the collected wireless sensor node data is compared with the preset standard data interval, misjudgment is likely to occur, thereby affecting the monitoring effect of rail transit accidents. Summary of the invention

[0005] The object of the present invention is to provide a rail transit accident monitoring system and method based on wireless communication to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] A rail traffic accident monitoring method based on wireless communication comprises the following steps:

[0008] Step S100. Obtain historical data corresponding to the wireless sensor nodes in the monitoring area, wherein the historical data includes track status data and environmental data; obtain log data corresponding to the rail traffic accident from the database, divide the historical data corresponding to the wireless sensor nodes according to the log data, and mark corresponding data tags according to the division results;

[0009] Step S200. Acquire the location data of the wireless sensor nodes in the monitoring area, and evaluate the operation status of the wireless sensor nodes in combination with the historical data of the wireless sensor nodes and the log data corresponding to the rail traffic accident; identify the key nodes in the monitoring area according to the evaluation results of the operation status of the wireless sensor nodes;

[0010] Step S300. According to the historical data corresponding to the key nodes in the monitoring area, the trend curve of the historical data corresponding to the key nodes is analyzed, and the trend characteristics of the historical data are extracted based on the trend curve of the historical data; according to the trend curve of the historical data and the trend characteristics of the corresponding historical data, a mapping relationship between the two is constructed;

[0011] Step S400. Obtain accident notifications and corresponding real-time data of key nodes in the monitoring area in real time, analyze the real-time data to obtain real-time data trend curves and corresponding real-time trend characteristics; judge the accuracy of the current accident notification based on the mapping relationship between the trend curves of historical data and the corresponding trend characteristics of historical data, and output the corresponding notification information.

[0012] Furthermore, step S100 includes:

[0013] S101. Obtain historical data corresponding to wireless sensor nodes in the monitoring area, and all wireless sensor nodes in the monitoring area correspond to unique numbers, and there is a unique correspondence between the numbers of the wireless sensor nodes and the corresponding historical data; obtain log data corresponding to rail transit accidents from a database, and the log data records information data of several rail transit events, including the numbers of wireless sensor nodes corresponding to rail transit events, as well as the time period and event type of the event; the event types are divided into real events and false events, wherein a real event refers to an accident notification and confirmation of the occurrence of an accident by relevant personnel, and a false event refers to an accident notification and confirmation of the occurrence of an accident by relevant personnel, or there is no accident notification but the occurrence of an accident is confirmed by relevant personnel;

[0014] S102. For each wireless sensor node in the monitoring area, extract the information data of the rail transit event corresponding to the number of the wireless sensor node from the log data corresponding to the rail transit accident, and extract the historical data of the corresponding time period according to the time period of the event in the information data; the time period of the event occurrence is a time period in which there is no missing historical data and the corresponding historical data is within a preset standard data interval; according to the event type in the information data, mark the historical data of each extracted time period, wherein the data label of the historical data corresponding to the real event is 1, and the data label of the historical data corresponding to the false event is 0.

[0015] Further, step S200 includes:

[0016] S201. For each wireless sensor node, obtain the location data of the wireless sensor node from the database, correspond the location data with the historical data corresponding to the selected time period T, so as to form a data unit; based on the data unit, and in combination with the log data corresponding to the rail traffic accident in the selected time period T where the data unit is located, calculate the state evaluation index Z of the corresponding wireless sensor node, and the specific calculation formula is:

[0017]

[0018] Among them, N_1 represents the number of real events corresponding to the current wireless sensor node in the selected time period T, N_0 represents the number of false events corresponding to the current wireless sensor node in the selected time period T, N represents i represents the total number of rail transit events recorded in the log data corresponding to the current wireless sensor node in the selected time period T, and N=N_1+N_0; i represents the data number of the time period corresponding to the rail transit events recorded in the log data corresponding to the current wireless sensor node in the selected time period T, ranging from 1 to N; mi represents the number of data points that do not belong to the preset standard data interval in the historical data of the i-th rail transit event corresponding to the current wireless sensor node in the selected time period T; Mi represents the number of historical data of the i-th rail transit event corresponding to the current wireless sensor node in the selected time period T; d represents the distance from the current wireless sensor node to the starting point of the monitoring area, and D represents the distance from the starting point to the end point of the monitoring area; wherein α and β represent different adjustment coefficients, and the values ​​are both greater than 0 and less than 1;

[0019] S202. Summarize the state evaluation indexes Z corresponding to all wireless sensor nodes within the selected time period T, and extract the state evaluation indexes of only real events in the log data according to the corresponding log data within the selected time period T, and obtain the state evaluation index set Q1 corresponding to the real events; based on the state evaluation index set Q1, calculate the corresponding average value μ and standard deviation σ, and obtain the corresponding state evaluation index threshold Z0 according to the average value and standard deviation, and the state evaluation index threshold Z0 = μ-k×σ, where k represents a preset coefficient; compare the state evaluation index Z corresponding to the wireless sensor node within the selected time period T with the state evaluation index threshold Z0, and take the wireless sensor nodes whose state evaluation index threshold Z0 is less than the state evaluation index threshold Z0 as key nodes.

[0020] Furthermore, step S300 includes:

[0021] S301. For each key node in the monitoring area, obtain historical data within the selected time period T, extract historical data with data label 0, so as to form a false event historical data set L, and L = {l1, l2, ..., lg}, where l1 represents the historical data of the first false event of the key node in the selected time period T, l2 represents the historical data of the second false event of the key node in the selected time period T, and so on, lg represents the historical data of the g-th false event of the key node in the selected time period T, and g represents the total number of false events of the key node in the selected time period T;

[0022] S302. Perform normalization on the historical data of each element in the false event historical data set L, and visualize the normalized historical data of false events through visualization tools, thereby obtaining g historical data trend curves, and each historical data trend curve corresponds one-to-one to an element in the corresponding false event historical data set L; for each historical data trend curve, extract the trend features of the corresponding historical data, and convert the extracted trend features into a trend feature vector V, and V = [v1, v2, ..., vh], where v1 represents the first trend feature value of the historical data trend curve, and v2 represents the historical data trend curve The second trend characteristic value of the historical data trend curve, and so on. vh represents the hth trend characteristic value of the historical data trend curve, h represents the number of the trend characteristic value of the historical data trend curve, and the maximum value of the number of trend characteristics in the g historical data trend curves is taken; and the sequential numbering of the trend characteristic values ​​of the trend characteristic vector V corresponding to each historical data trend curve is consistent. When the number of trend characteristics of a historical data trend curve is less than h, the trend characteristics of this historical data trend curve are intersected with the trend characteristics of the historical data trend curve with h trend characteristics, and the trend characteristic values ​​that do not exist in the historical data trend curve with less than h trend characteristics are assigned 0;

[0023] S303. According to the historical data trend curve and the corresponding historical data trend feature vector V, a mapping relationship between the two is constructed, and the corresponding mapping relationship expression is: Y = f(V) + φ, where Y represents the historical data trend curve, f(V) represents the functional relationship between the trend feature vector V and Y, and φ represents the random error term, which reflects the part that the model fails to capture or the noise in the data.

[0024] Furthermore, step S400 includes:

[0025] S401. Obtain accident notifications and corresponding real-time data of key nodes in the monitoring area in real time, visualize the real-time data, and obtain a real-time data trend curve, and extract corresponding real-time data trend features according to the real-time data trend curve, thereby obtaining a real-time trend feature vector V'; based on the mapping relationship between the trend curve of historical data and the corresponding trend features of historical data in S303, obtain the corresponding prediction trend curve according to the real-time trend feature vector V';

[0026] S402. Calculate the deviation index R between the prediction trend curve and the real-time data trend curve. The specific calculation formula is:

[0027]

[0028] Among them, u represents the number of data points, which is a positive integer from 1 to U; Y_u represents the value of the predicted trend curve at the u-th data point; P_u represents the value of the real-time data trend curve at the u-th data point; the calculated deviation degree index R is compared with the preset deviation degree threshold R0. If the deviation degree index R is greater than or equal to the preset deviation degree threshold R0, it means that the current accident notification is true, and the accident notification is output to the relevant personnel for processing; if the deviation degree index R is less than the preset deviation degree threshold R0, it means that the current accident notification is false, the accident notification is intercepted and the location data of the current key node is output to the relevant personnel, and the relevant personnel conduct corresponding investigation.

[0029] A rail transit accident monitoring system based on wireless communication, the system comprises: a data acquisition and processing module, a state evaluation module, a data analysis module, a real-time monitoring module and an accident judgment and response module;

[0030] The data acquisition and processing module collects historical data of wireless sensor nodes in the monitoring area, including track status and environmental data; obtains log data of rail transit accidents from the database, and matches and marks historical data based on real events and false events recorded in the log data;

[0031] The status assessment module obtains the location data of the wireless sensor nodes, and evaluates the operating status of each wireless sensor node in combination with the historical data of the nodes and the log data corresponding to the rail transit accidents; and identifies the key nodes in the monitoring area based on the assessment results;

[0032] The data analysis module analyzes the trend curves of the historical data of key nodes and extracts the trend features, thereby establishing a mapping relationship between the trend features of the historical data and the trend curves;

[0033] The real-time monitoring module obtains accident notifications and related real-time data of key nodes in real time, generates real-time data trend curves and extracts real-time trend features;

[0034] The accident judgment and response module obtains the predicted trend curve corresponding to the real-time data based on the mapping relationship between the trend characteristics of the historical data and the trend curve, calculates the deviation degree index between the predicted trend curve and the real-time data trend curve, judges the accuracy of the accident notification based on the deviation degree index, and outputs relevant notification information.

[0035] Further, the data acquisition and processing module includes a data acquisition unit and a data marking unit;

[0036] The data acquisition unit obtains the historical data of the wireless sensor nodes and the log data corresponding to the rail transit accidents, including the time period of the event, the event type and the related sensor node number; the data marking unit marks the historical data according to the log data and divides the historical data corresponding to the real event and the false event.

[0037] Further, the state assessment module includes a position data acquisition unit, a state assessment index calculation unit and a key node identification unit;

[0038] The location data acquisition unit acquires the location data of each wireless sensor node from the database; the state evaluation index calculation unit calculates the state evaluation index of the wireless sensor node based on the historical data of the wireless sensor node and the log data corresponding to the rail transit accident; the key node identification unit identifies the key node according to the state evaluation index of the wireless sensor node.

[0039] The data analysis module includes a data trend analysis unit and a mapping relationship establishment unit;

[0040] The data trend analysis unit analyzes the trend curve of the historical data of the identified key node, extracts the trend features, and forms a trend feature vector; the mapping relationship establishment unit constructs a mapping relationship between the historical data trend curve and its trend feature vector.

[0041] Furthermore, the real-time monitoring module includes an accident notification acquisition unit, a real-time data acquisition unit, and a real-time trend feature extraction unit;

[0042] The accident notification acquisition unit is used to receive accident notifications of key nodes in the monitoring area in real time; the real-time data acquisition unit is used to obtain corresponding real-time data and display it visually, thereby obtaining a real-time data trend curve; the real-time trend feature extraction unit extracts real-time trend features according to the real-time data trend curve, thereby obtaining a real-time trend feature vector;

[0043] The accident judgment and response module includes a prediction trend acquisition unit, a deviation degree calculation unit, and a prompt information output unit;

[0044] The prediction trend acquisition unit obtains the corresponding prediction trend curve according to the real-time trend feature vector based on the mapping relationship between the trend curve of historical data and the corresponding trend characteristics of the historical data; the deviation degree calculation unit is used to calculate the deviation degree index between the prediction trend curve and the real-time data trend curve; the prompt information output unit judges the authenticity of the accident notification according to the calculation result of the deviation degree calculation unit, and outputs corresponding prompt information according to the judgment result.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] By systematically analyzing the historical data of wireless sensor nodes and comparing them with accident logs, the present invention can effectively reduce data misjudgment caused by electromagnetic interference. By marking real and false events, it helps to ensure that only confirmed data is used for accident monitoring, thereby improving the reliability of the overall data. The present invention can identify key nodes in the monitoring area by calculating and comparing the status evaluation index. This process can reflect the operating status of each node in real time, help to quickly locate potential risk points, and optimize the allocation of monitoring resources. By extracting the trend features of historical data and establishing a mapping relationship, the present invention can not only predict the current accident trend, but also monitor the accuracy of accident notifications in real time. The comparison of the predicted trend and real-time data helps to quickly identify real accidents and false alarms, thereby improving the response efficiency of the monitoring system. Since the method comprehensively considers multiple factors (such as node status evaluation, historical data trend analysis, etc.), it can effectively reduce the probability of false alarms and missed alarms compared to traditional methods. An accurate accident notification judgment mechanism can avoid waste of resources and unnecessary emergency responses caused by false information. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0048] Figure 1 It is a schematic diagram of a module of a rail transit accident monitoring system based on wireless communication of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] See also Figure 1 , the present invention provides a technical solution:

[0051] A rail transit accident monitoring system based on wireless communication, the system comprises: a data acquisition and processing module, a state evaluation module, a data analysis module, a real-time monitoring module and an accident judgment and response module;

[0052] The data acquisition and processing module collects historical data of wireless sensor nodes in the monitoring area, including track status and environmental data; obtains log data of rail transit accidents from the database, and matches and marks historical data based on real events and false events recorded in the log data;

[0053] The status assessment module obtains the location data of the wireless sensor nodes, and evaluates the operating status of each wireless sensor node in combination with the historical data of the nodes and the log data corresponding to the rail transit accidents; and identifies the key nodes in the monitoring area based on the assessment results;

[0054] The data analysis module analyzes the trend curves of the historical data of key nodes and extracts the trend features, thereby establishing a mapping relationship between the trend features of the historical data and the trend curves;

[0055] The real-time monitoring module obtains accident notifications and related real-time data of key nodes in real time, generates real-time data trend curves and extracts real-time trend features;

[0056] The accident judgment and response module obtains the predicted trend curve corresponding to the real-time data based on the mapping relationship between the trend characteristics of the historical data and the trend curve, calculates the deviation degree index between the predicted trend curve and the real-time data trend curve, judges the accuracy of the accident notification based on the deviation degree index, and outputs relevant notification information.

[0057] The data acquisition and processing module includes a data acquisition unit and a data marking unit;

[0058] The data acquisition unit obtains the historical data of the wireless sensor nodes and the log data corresponding to the rail transit accidents, including the time period of the event, the event type and the related sensor node number; the data marking unit marks the historical data according to the log data and divides the historical data corresponding to the real event and the false event.

[0059] The state assessment module includes a position data acquisition unit, a state assessment index calculation unit and a key node identification unit;

[0060] The location data acquisition unit acquires the location data of each wireless sensor node from the database; the state evaluation index calculation unit calculates the state evaluation index of the wireless sensor node based on the historical data of the wireless sensor node and the log data corresponding to the rail transit accident; the key node identification unit identifies the key node according to the state evaluation index of the wireless sensor node.

[0061] The data analysis module includes a data trend analysis unit and a mapping relationship establishment unit;

[0062] The data trend analysis unit analyzes the trend curve of the historical data of the identified key node, extracts the trend features, and forms a trend feature vector; the mapping relationship establishment unit constructs a mapping relationship between the historical data trend curve and its trend feature vector.

[0063] The real-time monitoring module includes an accident notification acquisition unit, a real-time data acquisition unit, and a real-time trend feature extraction unit;

[0064] The accident notification acquisition unit is used to receive accident notifications of key nodes in the monitoring area in real time; the real-time data acquisition unit is used to obtain corresponding real-time data and display it visually, thereby obtaining a real-time data trend curve; the real-time trend feature extraction unit extracts real-time trend features according to the real-time data trend curve, thereby obtaining a real-time trend feature vector;

[0065] The accident judgment and response module includes a prediction trend acquisition unit, a deviation degree calculation unit, and a prompt information output unit;

[0066] The prediction trend acquisition unit obtains the corresponding prediction trend curve according to the real-time trend feature vector based on the mapping relationship between the trend curve of historical data and the corresponding trend characteristics of the historical data; the deviation degree calculation unit is used to calculate the deviation degree index between the prediction trend curve and the real-time data trend curve; the prompt information output unit judges the authenticity of the accident notification according to the calculation result of the deviation degree calculation unit, and outputs corresponding prompt information according to the judgment result.

[0067] A rail transit accident monitoring method based on wireless communication comprises the following steps:

[0068] Step S100. Obtain historical data corresponding to the wireless sensor nodes in the monitoring area, wherein the historical data includes track status data and environmental data; obtain log data corresponding to the rail traffic accident from the database, divide the historical data corresponding to the wireless sensor nodes according to the log data, and mark corresponding data tags according to the division results;

[0069] Step S200. Acquire the location data of the wireless sensor nodes in the monitoring area, and evaluate the operation status of the wireless sensor nodes in combination with the historical data of the wireless sensor nodes and the log data corresponding to the rail traffic accident; identify the key nodes in the monitoring area according to the evaluation results of the operation status of the wireless sensor nodes;

[0070] Step S300. According to the historical data corresponding to the key nodes in the monitoring area, the trend curve of the historical data corresponding to the key nodes is analyzed, and the trend characteristics of the historical data are extracted based on the trend curve of the historical data; according to the trend curve of the historical data and the trend characteristics of the corresponding historical data, a mapping relationship between the two is constructed;

[0071] Step S400. Obtain accident notifications and corresponding real-time data of key nodes in the monitoring area in real time, analyze the real-time data to obtain real-time data trend curves and corresponding real-time trend characteristics; judge the accuracy of the current accident notification based on the mapping relationship between the trend curves of historical data and the corresponding trend characteristics of historical data, and output the corresponding notification information.

[0072] Step S100 includes:

[0073] S101. Obtain historical data corresponding to wireless sensor nodes in the monitoring area, and all wireless sensor nodes in the monitoring area correspond to unique numbers, and there is a unique correspondence between the numbers of the wireless sensor nodes and the corresponding historical data; obtain log data corresponding to rail transit accidents from a database, and the log data records information data of several rail transit events, including the numbers of wireless sensor nodes corresponding to rail transit events, as well as the time period and event type of the event; the event types are divided into real events and false events, wherein a real event refers to an accident notification and confirmation of the occurrence of an accident by relevant personnel, and a false event refers to an accident notification and confirmation of the occurrence of an accident by relevant personnel, or there is no accident notification but the occurrence of an accident is confirmed by relevant personnel;

[0074] S102. For each wireless sensor node in the monitoring area, extract the information data of the rail transit event corresponding to the number of the wireless sensor node from the log data corresponding to the rail transit accident, and extract the historical data of the corresponding time period according to the time period of the event in the information data; the time period of the event occurrence is a time period in which there is no missing historical data and the corresponding historical data is within a preset standard data interval; according to the event type in the information data, mark the historical data of each extracted time period, wherein the data label of the historical data corresponding to the real event is 1, and the data label of the historical data corresponding to the false event is 0.

[0075] Step S200 includes:

[0076] S201. For each wireless sensor node, obtain the location data of the wireless sensor node from the database, correspond the location data with the historical data corresponding to the selected time period T, so as to form a data unit; based on the data unit, and in combination with the log data corresponding to the rail traffic accident in the selected time period T where the data unit is located, calculate the state evaluation index Z of the corresponding wireless sensor node, and the specific calculation formula is:

[0077]

[0078] Among them, N_1 represents the number of real events corresponding to the current wireless sensor node in the selected time period T, N_0 represents the number of false events corresponding to the current wireless sensor node in the selected time period T, N represents i represents the total number of rail transit events recorded in the corresponding log data of the current wireless sensor node in the selected time period T, and N=N_1+N_0; i represents the data number of the time period corresponding to the rail transit events recorded in the corresponding log data of the current wireless sensor node in the selected time period T, ranging from 1 to N; mi represents the number of data points that do not belong to the preset standard data interval in the historical data of the i-th rail transit event corresponding to the current wireless sensor node in the selected time period T; Mi represents the number of historical data of the i-th rail transit event corresponding to the current wireless sensor node in the selected time period T; d represents the distance from the current wireless sensor node to the starting point of the monitoring area, and D represents the distance from the starting point to the end point of the monitoring area; wherein α and β represent different adjustment coefficients, and the values ​​are both greater than 0 and less than 1;

[0079] S202. Summarize the state evaluation indexes Z corresponding to all wireless sensor nodes within the selected time period T, and extract the state evaluation indexes of only real events in the log data according to the corresponding log data within the selected time period T, and obtain the state evaluation index set Q1 corresponding to the real events; based on the state evaluation index set Q1, calculate the corresponding average value μ and standard deviation σ, and obtain the corresponding state evaluation index threshold Z0 according to the average value and standard deviation, and the state evaluation index threshold Z0 = μ-k×σ, where k represents a preset coefficient; compare the state evaluation index Z corresponding to the wireless sensor node within the selected time period T with the state evaluation index threshold Z0, and take the wireless sensor nodes whose state evaluation index threshold Z0 is less than the state evaluation index threshold Z0 as key nodes.

[0080] Step S300 includes:

[0081] S301. For each key node in the monitoring area, obtain historical data within the selected time period T, extract historical data with data label 0, so as to form a false event historical data set L, and L = {l1, l2, ..., lg}, where l1 represents the historical data of the first false event of the key node in the selected time period T, l2 represents the historical data of the second false event of the key node in the selected time period T, and so on, lg represents the historical data of the g-th false event of the key node in the selected time period T, and g represents the total number of false events of the key node in the selected time period T;

[0082] S302. Perform normalization on the historical data of each element in the false event historical data set L, and visualize the normalized historical data of false events through visualization tools, thereby obtaining g historical data trend curves, and each historical data trend curve corresponds one-to-one to an element in the corresponding false event historical data set L; for each historical data trend curve, extract the trend features of the corresponding historical data, and convert the extracted trend features into a trend feature vector V, and V = [v1, v2, ..., vh], where v1 represents the first trend feature value of the historical data trend curve, and v2 represents the historical data trend curve The second trend characteristic value of the historical data trend curve, and so on. vh represents the hth trend characteristic value of the historical data trend curve, h represents the number of the trend characteristic value of the historical data trend curve, and the maximum value of the number of trend characteristics in the g historical data trend curves is taken; and the sequential numbering of the trend characteristic values ​​of the trend characteristic vector V corresponding to each historical data trend curve is consistent. When the number of trend characteristics of a historical data trend curve is less than h, the trend characteristics of this historical data trend curve are intersected with the trend characteristics of the historical data trend curve with h trend characteristics, and the trend characteristic values ​​that do not exist in the historical data trend curve with less than h trend characteristics are assigned 0;

[0083] In this embodiment, assuming that h is 5, the corresponding trend feature vector V = [v1, v2, v3, v4, v5]; when the number of trend features extracted from a historical data trend curve is equal to 3, and the corresponding trend feature eigenvalues ​​are v2, v4, and v5 respectively, then the trend feature vector V = [0, v2, 0, v4, v5] of this historical data trend curve.

[0084] S303. According to the historical data trend curve and the corresponding historical data trend feature vector V, a mapping relationship between the two is constructed, and the corresponding mapping relationship expression is: Y = f(V) + φ, where Y represents the historical data trend curve, f(V) represents the functional relationship between the trend feature vector V and Y, and φ represents the random error term, which reflects the part that the model fails to capture or the noise in the data.

[0085] In this embodiment, it is assumed that the relationship between the trend feature vector V corresponding to f(V) and the output Y is linear, which can be expressed as: f(V) = a0+a1v1+a2v2+a3v3+a4v4+a5v5, where a0, a1, a2, a3, a4, and a5 are model parameters obtained by fitting historical data;

[0086] If the relationship between the data is more complex, nonlinear functions such as polynomial regression, support vector machine, neural network, etc. can be used; machine learning models (such as random forest, neural network) can also be used to fit the relationship of f(V). In this case, f(V) may not have a clear analytical expression, but is obtained through training data.

[0087] φ represents a random error term, which reflects the part that the model fails to capture or the noise in the data. It represents the difference between the actual output and the output predicted by the function f(V), and φ is usually assumed to be a random variable with zero mean.

[0088] Step S400 includes:

[0089] S401. Obtain accident notifications and corresponding real-time data of key nodes in the monitoring area in real time, visualize the real-time data, and obtain a real-time data trend curve, and extract corresponding real-time data trend features according to the real-time data trend curve, thereby obtaining a real-time trend feature vector V'; based on the mapping relationship between the trend curve of historical data and the corresponding trend features of historical data in S303, obtain the corresponding prediction trend curve according to the real-time trend feature vector V';

[0090] S402. Calculate the deviation index R between the prediction trend curve and the real-time data trend curve. The specific calculation formula is:

[0091]

[0092] Among them, u represents the number of data points, which is a positive integer from 1 to U; Y_u represents the value of the predicted trend curve at the u-th data point; P_u represents the value of the real-time data trend curve at the u-th data point; the calculated deviation degree index R is compared with the preset deviation degree threshold R0. If the deviation degree index R is greater than or equal to the preset deviation degree threshold R0, it means that the current accident notification is true, and the accident notification is output to the relevant personnel for processing; if the deviation degree index R is less than the preset deviation degree threshold R0, it means that the current accident notification is false, the accident notification is intercepted and the location data of the current key node is output to the relevant personnel, and the relevant personnel conduct corresponding investigation.

[0093] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0094] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A rail traffic accident monitoring method based on wireless communication, characterized in that: The method comprises the following steps: Step S100. Obtain historical data corresponding to the wireless sensor nodes in the monitoring area, wherein the historical data includes track status data and environmental data; obtain log data corresponding to the rail traffic accident from the database, divide the historical data corresponding to the wireless sensor nodes according to the log data, and mark corresponding data tags according to the division results; Step S200. Acquire the location data of the wireless sensor nodes in the monitoring area, and evaluate the operation status of the wireless sensor nodes in combination with the historical data of the wireless sensor nodes and the log data corresponding to the rail traffic accident; identify the key nodes in the monitoring area according to the evaluation results of the operation status of the wireless sensor nodes; Step S300. According to the historical data corresponding to the key nodes in the monitoring area, the trend curve of the historical data corresponding to the key nodes is analyzed, and the trend characteristics of the historical data are extracted based on the trend curve of the historical data; according to the trend curve of the historical data and the trend characteristics of the corresponding historical data, a mapping relationship between the two is constructed; Step S400. Obtain accident notifications and corresponding real-time data of key nodes in the monitoring area in real time, analyze the real-time data to obtain real-time data trend curves and corresponding real-time trend features; determine the accuracy of the current accident notification based on the mapping relationship between the trend curves of historical data and the corresponding trend features of historical data, and output the corresponding notification information; The step S100 includes: S101. Obtain historical data corresponding to wireless sensor nodes in the monitoring area, and all wireless sensor nodes in the monitoring area correspond to unique numbers, and there is a unique correspondence between the numbers of the wireless sensor nodes and the corresponding historical data; obtain log data corresponding to rail transit accidents from a database, and the log data records information data of several rail transit events, including the numbers of wireless sensor nodes corresponding to rail transit events, as well as the time period and event type of the event; the event types are divided into real events and false events, wherein a real event refers to an accident notification and confirmation of the occurrence of an accident by relevant personnel, and a false event refers to an accident notification and confirmation of the occurrence of an accident by relevant personnel, or there is no accident notification but the occurrence of an accident is confirmed by relevant personnel; S102. For each wireless sensor node in the monitoring area, extract the information data of the rail transit event corresponding to the number of the wireless sensor node from the log data corresponding to the rail transit accident, and extract the historical data of the corresponding time period according to the time period of the event in the information data; the time period of the event is a time period in which there is no missing historical data and the corresponding historical data is within a preset standard data interval; according to the event type in the information data, mark the historical data of each extracted time period, wherein the data label of the historical data corresponding to the real event is 1, and the data label of the historical data corresponding to the false event is 0; The step S200 includes: S201. For each wireless sensor node, obtain the location data of the wireless sensor node from the database, correspond the location data with the historical data corresponding to the selected time period T, so as to form a data unit; based on the data unit, and in combination with the log data corresponding to the rail traffic accident in the selected time period T where the data unit is located, calculate the state evaluation index Z of the corresponding wireless sensor node, and the specific calculation formula is: Among them, N_1 represents the number of real events corresponding to the current wireless sensor node in the selected time period T, N_0 represents the number of false events corresponding to the current wireless sensor node in the selected time period T, N represents i represents the total number of rail transit events recorded in the log data corresponding to the current wireless sensor node in the selected time period T, and N=N_1+N_0; i represents the data number of the time period corresponding to the rail transit events recorded in the log data corresponding to the current wireless sensor node in the selected time period T, ranging from 1 to N; mi represents the number of data points that do not belong to the preset standard data interval in the historical data of the i-th rail transit event corresponding to the current wireless sensor node in the selected time period T; Mi represents the number of historical data of the i-th rail transit event corresponding to the current wireless sensor node in the selected time period T; d represents the distance from the current wireless sensor node to the starting point of the monitoring area, and D represents the distance from the starting point to the end point of the monitoring area; wherein α and β represent different adjustment coefficients, and the values ​​are both greater than 0 and less than 1; S202. Summarize the state evaluation indexes Z corresponding to all wireless sensor nodes within the selected time period T, and extract the state evaluation indexes of only real events in the log data according to the corresponding log data within the selected time period T, and obtain the state evaluation index set Q1 corresponding to the real events; based on the state evaluation index set Q1, calculate the corresponding average value μ and standard deviation σ, and obtain the corresponding state evaluation index threshold Z0 according to the average value and standard deviation, and the state evaluation index threshold Z0 = μ-k×σ, where k represents a preset coefficient; compare the state evaluation index Z corresponding to the wireless sensor node within the selected time period T with the state evaluation index threshold Z0, and take the wireless sensor nodes whose state evaluation index threshold Z0 is less than the state evaluation index threshold Z0 as key nodes.

2. A rail transit accident monitoring and analysis method based on wireless communication according to claim 1, characterized in that: The step S300 includes: S301. For each key node in the monitoring area, obtain historical data within the selected time period T, extract historical data with data label 0, so as to form a false event historical data set L, and L = {l1, l2, ..., lg}, where l1 represents the historical data of the first false event of the key node in the selected time period T, l2 represents the historical data of the second false event of the key node in the selected time period T, and so on, lg represents the historical data of the g-th false event of the key node in the selected time period T, and g represents the total number of false events of the key node in the selected time period T; S302. Perform normalization on the historical data of each element in the false event historical data set L, and visualize the normalized historical data of false events through visualization tools, thereby obtaining g historical data trend curves, and each historical data trend curve corresponds one-to-one to an element in the corresponding false event historical data set L; for each historical data trend curve, extract the trend features of the corresponding historical data, and convert the extracted trend features into a trend feature vector V, and V = [v1, v2, ..., vh], where v1 represents the first trend feature value of the historical data trend curve, and v2 represents the historical data trend curve The second trend characteristic value of the historical data trend curve, and so on. vh represents the hth trend characteristic value of the historical data trend curve, h represents the number of the trend characteristic value of the historical data trend curve, and the maximum value of the number of trend characteristics in the g historical data trend curves is taken; and the sequential numbering of the trend characteristic values ​​of the trend characteristic vector V corresponding to each historical data trend curve is consistent. When the number of trend characteristics of a historical data trend curve is less than h, the trend characteristics of this historical data trend curve are intersected with the trend characteristics of the historical data trend curve with h trend characteristics, and the trend characteristic values ​​that do not exist in the historical data trend curve with less than h trend characteristics are assigned 0; S303. According to the historical data trend curve and the corresponding historical data trend feature vector V, a mapping relationship between the two is constructed, and the corresponding mapping relationship expression is: Y = f(V) + φ, where Y represents the historical data trend curve, f(V) represents the functional relationship between the trend feature vector V and Y, and φ represents the random error term, which reflects the part that the model fails to capture or the noise in the data.

3. A rail traffic accident monitoring method based on wireless communication according to claim 2, characterized in that: The step S400 includes: S401. Obtain accident notifications and corresponding real-time data of key nodes in the monitoring area in real time, visualize the real-time data, and obtain a real-time data trend curve, and extract corresponding real-time data trend features according to the real-time data trend curve, thereby obtaining a real-time trend feature vector V'; based on the mapping relationship between the trend curve of historical data and the corresponding trend features of historical data in S303, obtain the corresponding prediction trend curve according to the real-time trend feature vector V'; S402. Calculate the deviation index R between the prediction trend curve and the real-time data trend curve. The specific calculation formula is: Among them, u represents the number of data points, which is a positive integer from 1 to U; Y_u represents the value of the predicted trend curve at the u-th data point; P_u represents the value of the real-time data trend curve at the u-th data point; the calculated deviation degree index R is compared with the preset deviation degree threshold R0. If the deviation degree index R is greater than or equal to the preset deviation degree threshold R0, it means that the current accident notification is true, and the accident notification is output to the relevant personnel for processing; if the deviation degree index R is less than the preset deviation degree threshold R0, it means that the current accident notification is false, the accident notification is intercepted and the location data of the current key node is output to the relevant personnel, and the relevant personnel conduct corresponding investigation.

4. A rail transit accident monitoring system based on wireless communication, applied to a rail transit accident monitoring method based on wireless communication according to any one of claims 1 to 3, characterized in that: The system includes: a data acquisition and processing module, a status assessment module, a data analysis module, a real-time monitoring module, and an accident judgment and response module; The data acquisition and processing module collects historical data of wireless sensor nodes in the monitoring area, including track status and environmental data; obtains log data of rail traffic accidents from the database, and matches and marks historical data according to real events and false events recorded in the log data; The state evaluation module obtains the location data of the wireless sensor nodes, and evaluates the operation status of each wireless sensor node in combination with the historical data of the nodes and the log data corresponding to the rail transit accident; and identifies the key nodes in the monitoring area based on the evaluation results; The data analysis module analyzes the trend curve of the historical data of the key nodes and extracts the trend features, thereby establishing a mapping relationship between the trend features of the historical data and the trend curve; The real-time monitoring module obtains accident notifications and related real-time data of key nodes in real time, generates real-time data trend curves and extracts real-time trend features; The accident judgment and response module obtains the predicted trend curve corresponding to the real-time data based on the mapping relationship between the trend characteristics of the historical data and the trend curve, calculates the deviation degree index between the predicted trend curve and the real-time data trend curve, judges the accuracy of the accident notification based on the deviation degree index, and outputs relevant notification information.

5. A rail traffic accident monitoring system based on wireless communication according to claim 4, characterized in that: The data acquisition and processing module includes a data acquisition unit and a data marking unit; The data acquisition unit acquires historical data of the wireless sensor nodes and log data corresponding to the rail transit accident, including the time period of the event, the event type and the related sensor node number; The data marking unit marks the historical data according to the log data, and divides the historical data corresponding to real events and false events.

6. A rail traffic accident monitoring system based on wireless communication according to claim 4, characterized in that: The state assessment module includes a position data acquisition unit, a state assessment index calculation unit and a key node identification unit; The location data acquisition unit acquires the location data of each wireless sensor node from the database; the state evaluation index calculation unit calculates the state evaluation index of the wireless sensor node in combination with the historical data of the wireless sensor node and the log data corresponding to the rail traffic accident; The key node identification unit identifies the key nodes according to the state evaluation index of the wireless sensor nodes.

7. A rail traffic accident monitoring system based on wireless communication according to claim 4, characterized in that: The data analysis module includes a data trend analysis unit and a mapping relationship establishment unit; The data trend analysis unit analyzes the trend curve of the historical data of the identified key node, extracts the trend features, and forms a trend feature vector; The mapping relationship establishing unit constructs a mapping relationship between the historical data trend curve and its trend feature vector.

8. A rail traffic accident monitoring system based on wireless communication according to claim 4, characterized in that: The real-time monitoring module includes an accident notification acquisition unit, a real-time data acquisition unit and a real-time trend feature extraction unit; The accident notification acquisition unit is used to receive accident notifications of key nodes in the monitoring area in real time; The real-time data acquisition unit is used to acquire corresponding real-time data and display it visually, thereby obtaining a real-time data trend curve; The real-time trend feature extraction unit extracts the real-time trend feature according to the real-time data trend curve, thereby obtaining a real-time trend feature vector; The accident judgment and response module includes a prediction trend acquisition unit, a deviation degree calculation unit and a prompt information output unit; The prediction trend acquisition unit obtains the corresponding prediction trend curve according to the real-time trend feature vector based on the mapping relationship between the trend curve of the historical data and the trend features of the corresponding historical data; The deviation degree calculation unit is used to calculate the deviation degree index between the predicted trend curve and the real-time data trend curve; the prompt information output unit judges the authenticity of the accident notification according to the calculation result of the deviation degree calculation unit, and outputs corresponding prompt information according to the judgment result.

Images