A method and system for identifying means of transport based on signaling data

Abstract

The present invention discloses a method for identifying vehicles based on signaling data, including: S1 obtaining terminal signaling data, and performing preprocessing to obtain user trajectory data; S2 identifying and obtaining the corresponding vehicle type based on the user trajectory data; S3 extracting conventional feature values ​​based on the user trajectory data and the corresponding vehicle type, and obtaining the distribution of each feature value; S4 establishing conventional feature rule strategies based on conventional feature values ​​and distribution; S5 establishing advanced feature rule strategies based on user trajectory data and distribution; S6 establishing expert system rules based on conventional feature rule strategies and advanced feature rule strategies; S7 verifying and adjusting the expert system rules to obtain the optimal expert system rules; S8 identifying the terminal signaling data to be tested based on the optimal expert system rules to obtain the vehicle type. The vehicle type to which the signaling data belongs is efficiently and accurately identified through the optimal expert system rules.

Description

Technical Field

[0001] The present invention relates to the technical field of communication data mining, and more particularly to a method and system for identifying vehicles based on signaling data. Background Art

[0002] At present, in the fields of technical reconnaissance, security control, etc., there is a growing demand for accurate identification of target behaviors and activities. Traditional monitoring and reconnaissance methods usually rely on fixed-position cameras, sensor networks or other special equipment. The limitation of traditional monitoring methods is that they cannot provide enough information to accurately identify the behavior or activities of the target, especially in complex urban environments; secondly, traditional monitoring methods require a lot of manpower and time for data analysis and event identification, which limits the rapid response to real-time activities; in addition, some public transportation, such as subways and airplanes, have large passenger flow and difficult installation and maintenance of monitoring equipment or sensor networks, which makes it difficult to obtain user data and thus cannot identify the mode of transportation they take.

[0003] In the field of technical reconnaissance and security control, communication big data has become a potential solution. When users' mobile communication devices move, they interact with nearby communication base stations, generating a large amount of signaling trajectory data. These data contain the user's location information and communication history with the base station, providing important information for identifying target behaviors and activities. However, accurately identifying the target behavior and the type of transportation they take from these massive communication data is still a complex and challenging task. Existing methods cannot quickly and accurately identify the type of transportation the target takes.

[0004] Therefore, how to efficiently and accurately identify the type of transportation a user takes is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0005] In view of this, the present invention provides a method and system for identifying vehicles based on signaling data, which analyzes the signaling trajectory data of the user's mobile phone, constructs conventional features and advanced features, explores based on the features, and constructs expert system rules. The expert system rules are used to efficiently and accurately identify the type of vehicle to which the signaling trajectory data belongs.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A method for identifying a vehicle based on signaling data, comprising:

[0008] S1 obtains terminal signaling data and performs preprocessing to obtain user trajectory data;

[0009] S2 identifies the corresponding transportation type based on the user trajectory data;

[0010] S3 extracts conventional feature values ​​based on the user trajectory data and the corresponding transportation tool type, and obtains the distribution of each feature value;

[0011] S4 establishes a regular feature rule strategy based on the regular feature value and the distribution;

[0012] S5 establishes an advanced feature rule strategy based on the user trajectory data and the distribution situation;

[0013] S6 establishes expert system rules based on the conventional feature rule strategy and the advanced feature rule strategy;

[0014] S7 verifies and adjusts the expert system rules to obtain optimal expert system rules;

[0015] S8 identifies the signaling data of the terminal to be tested based on the optimal expert system rule to obtain the type of transportation tool.

[0016] Preferably, the user trajectory data is obtained, and the specific process is as follows:

[0017] Obtaining terminal signaling trajectory data based on the terminal signaling data, and performing data cleaning to filter out data with abnormal trajectory status;

[0018] The terminal signaling trajectory data after data cleaning is sorted in chronological order to obtain initial trajectory data;

[0019] The initial trajectory data is subjected to outlier noise removal, ping-pong data aggregation, and trajectory completion to obtain user trajectory data.

[0020] Preferably, the conventional characteristic values ​​at least include:

[0021] The moving distance between adjacent trajectory points, the time difference between adjacent trajectory points, the maximum instantaneous speed between adjacent trajectory points, the straight-line distance from the first trajectory point, the cumulative sum of the distances up to the current trajectory point, the sum of the moving distances of the signaling in trajectory point windows of different sizes, the average moving distances of the signaling in trajectory point windows of different sizes, the average speed of the signaling in trajectory point windows of different sizes, and the number of base stations passed per unit time.

[0022] Preferably, a conventional feature rule strategy is established, and the specific process is as follows:

[0023] Summarizing thresholds for different types of vehicles based on the distribution;

[0024] A regular feature rule policy is established based on the threshold and the regular feature value.

[0025] Preferably, the advanced feature rule strategy is specifically:

[0026] According to different modes of transportation, there are airplane identification rules, train or high-speed rail identification rules, car identification rules and subway identification rules;

[0027] The aircraft identification rule is as follows: select user trajectory data whose distance from any airport is less than a preset value A from the user trajectory data; extract and calculate the time difference between trajectory points near the airport, and remove the user trajectory data whose time difference is greater than a preset value B; select user trajectory data whose stay time at the airport is greater than a preset value C; select user trajectory data whose distance from 24:00 on the same day to the disappearance of the user trajectory data is greater than a preset value D hours;

[0028] The train or high-speed rail identification rule is: select user trajectory data in which the base station location data in the user trajectory data continuously appears the string "train" or "high-speed rail" for more than a preset value E times; select user trajectory data of a single sport in which the user trajectory data continuously appears for more than a preset value F times; select user trajectory data with a maximum movement distance greater than a preset value G;

[0029] The car identification rule is: based on filtering the user trajectory data of trains or high-speed railways, select the user trajectory data with the character string "highway" or "highway" or "toll station" appearing in the base station location data; select the user trajectory data with a daily moving distance greater than a preset value H;

[0030] The subway identification rules are as follows: select user trajectory data in which the character string "subway" or "line" or "track" appears for more than a preset value K times in a row in the base station location data; select numbers in which the different values ​​of "subway", "line" or "track" appearing in the base station location data are greater than or equal to a preset value L.

[0031] Preferably, if the conditions in the airplane identification rule, the train or high-speed rail identification rule, the car identification rule and the subway identification rule are all met, the vehicle is identified as the corresponding type of transportation.

[0032] Preferably, the expert system rule verification process is specifically as follows:

[0033] Extract new terminal signaling data as test data;

[0034] Identify the test data based on the expert system rules to obtain rule determination results;

[0035] Process the test data based on S1-S2 to determine the type of vehicle as a test data label;

[0036] The rule determination result is compared with the test data label, and the recognition accuracy of the expert system rule is counted.

[0037] Preferably, the optimal expert system rules are obtained, specifically including:

[0038] Determining whether the recognition accuracy meets the requirement;

[0039] If the requirements are met, the expert system rules are used as the optimal expert system rules;

[0040] If the requirements are not met, more terminal signaling data is obtained, and S1 to S7 are repeated. The rule threshold is adjusted according to the feature distribution, and the expert system rules are reorganized and verified until the recognition accuracy meets the requirements. The expert system rules at this time are output as the optimal expert system rules.

[0041] A vehicle identification system based on signaling data, comprising: a data acquisition processing module, a first identification module, a conventional feature extraction module, a conventional strategy establishment module, an advanced strategy establishment module, an expert rule establishment module, a verification adjustment module and a second identification module;

[0042] The data acquisition processing module is used to acquire terminal signaling data and perform preprocessing to obtain user trajectory data;

[0043] The first identification module is used to identify the corresponding transportation type based on the user trajectory data;

[0044] The conventional feature extraction module is used to extract conventional feature values ​​based on the user trajectory data and the corresponding transportation tool type, and obtain the distribution of each feature value;

[0045] The conventional strategy establishment module is used to establish a conventional feature rule strategy based on the conventional feature value and the distribution situation;

[0046] The advanced strategy establishment module is used to establish an advanced feature rule strategy based on the user trajectory data and the distribution situation;

[0047] The expert rule establishment module is used to establish expert system rules based on the conventional feature rule strategy and the advanced feature rule strategy;

[0048] The verification and adjustment module is used to verify and adjust the expert system rules to obtain the optimal expert system rules;

[0049] The second identification module is used to identify the signaling data of the terminal to be tested based on the optimal expert system rule to obtain the type of transportation tool.

[0050] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method and system for identifying vehicles based on signaling data, which combines relevant technologies in the fields of data analysis, feature engineering, and expert system rules. By analyzing the signaling trajectory data of the user's mobile phone, conventional features and advanced features are constructed, and expert system rules are constructed based on the features for exploration. The types of vehicles to which the signaling trajectory data belongs are efficiently and accurately identified through expert system rules. It has the following beneficial effects:

[0051] 1. The present invention realizes accurate identification of the user's mode of transportation based on various features of the user's mobile phone signaling data in the communication data and expert system rules. It solves the problems that the current traditional monitoring method requires a large number of professional and technical personnel to perform data analysis and event identification, cannot quickly obtain identification results, and has difficulty in obtaining user data.

[0052] 2. The present invention is based on communication data, feature engineering and expert system rules. It does not require expensive hardware materials and has timely response. It can not only be used for traffic management and monitoring, but also can be widely used in intelligent transportation systems, smart city management, technical reconnaissance and security control and other businesses.

[0053] 3. The present invention has the advantages of high accuracy, real-time and wide applicability, can effectively improve the efficiency and accuracy of vehicle identification, and contribute to the development and progress of intelligent traffic management, technical reconnaissance and security control.

[0054] 4. The present invention combines communication big data with rule strategies and uses advanced data analysis technology and algorithms to achieve the identification and tracking of transportation vehicles. According to communication-related data, rule strategies are designed, combined with characteristic parameters, to identify the type of transportation vehicle that the user is riding. These rules can be formulated based on machine learning algorithms, pattern matching and other technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0056] Figure 1 A flow chart of a vehicle identification method based on signaling data provided by the present invention.

[0057] Figure 2 This is a flow chart of the expert system rule verification method provided by the present invention.

[0058] Figure 3A schematic diagram of density estimation of the mean speeds of different types of vehicles in different trajectory windows provided by the present invention.

[0059] Figure 4 A schematic diagram of the confusion matrix for predicting various types of transportation using the optimal expert system rules provided by the present invention.

[0060] Figure 5 A schematic diagram of the structure of a vehicle identification system based on signaling data provided by the present invention. DETAILED DESCRIPTION

[0061] 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.

[0062] Example 1

[0063] like Figure 1 As shown, an embodiment of the present invention discloses a method for identifying a vehicle based on signaling data, comprising:

[0064] S1 obtains terminal signaling data and performs preprocessing to obtain user trajectory data;

[0065] S2 identifies the corresponding transportation type based on the user trajectory data;

[0066] S3 extracts conventional feature values ​​based on user trajectory data and corresponding transportation tool types, and obtains the distribution of each feature value;

[0067] S4 establishes a regular feature rule strategy based on regular feature values ​​and distribution;

[0068] S5 establishes advanced feature rule strategies based on user trajectory data and distribution;

[0069] S6 establishes expert system rules based on conventional feature rule strategies and advanced feature rule strategies;

[0070] S7 verifies and adjusts the expert system rules to obtain the optimal expert system rules;

[0071] S8 identifies the signaling data of the terminal to be tested based on the optimal expert system rules to obtain the type of transportation tool.

[0072] Example 2

[0073] The embodiment of the present invention discloses a method for identifying a vehicle based on signaling data, comprising:

[0074] S1 obtains terminal signaling data and performs preprocessing to obtain user trajectory data.

[0075] Preferably, the user trajectory data is obtained, and the specific process is as follows:

[0076] Based on the terminal signaling data, the terminal signaling trajectory data is obtained, and the data is cleaned to delete the data with empty or zero key fields and abnormal trajectory status due to base station data parsing errors or terminal device signal abnormalities, so as to remove data noise;

[0077] The terminal signaling trajectory data after data cleaning is sorted in chronological order to obtain initial trajectory data;

[0078] And use the folium tool in python to visualize the initial trajectory data on the map;

[0079] The visualized initial trajectory data is subjected to outlier noise removal, ping-pong data aggregation, and trajectory completion to obtain user trajectory data.

[0080] Preferably, trajectory point aggregation is performed on the ping-pong switching data in the initial trajectory data, and trajectory completion is performed on the data with few trajectory points or missing data fields in the initial trajectory data.

[0081] Preferably, the terminal signaling data mainly includes trajectory location data, and the trajectory location data of each terminal signaling data includes different trajectory points. A trajectory point is a piece of sample data, including information such as mobile phone number, time, longitude and latitude.

[0082] S2 identifies the corresponding transportation type based on the user trajectory data.

[0083] Preferably, the user trajectory data is visualized on a map by using the folium tool in Python, and the type of transportation the user takes is determined and confirmed by manually observing and analyzing the presentation image of the user trajectory data on the map. For example, if the user trajectory is along a highway or an urban road, the transportation is considered to be a car; if the user trajectory is along a railway, the transportation is considered to be a train or a high-speed railway; if the trajectory is along a subway line, the transportation is considered to be a subway.

[0084] S3 extracts conventional feature values ​​based on the user trajectory data and the corresponding transportation tool type, and obtains the distribution of each feature value.

[0085] Preferably, the conventional characteristic values ​​include at least:

[0086] The moving distance between adjacent trajectory points, the time difference between adjacent trajectory points, the maximum instantaneous speed between adjacent trajectory points, the straight-line distance from the first trajectory point, the cumulative sum of the distances up to the current trajectory point, the sum of the moving distances of the signaling in trajectory point windows of different sizes, the average moving distances of the signaling in trajectory point windows of different sizes, the average speed of the signaling in trajectory point windows of different sizes and the number of base stations passed per unit time, the maximum distance, the minimum distance, the average distance, the maximum driving speed, the minimum driving speed and the average driving speed.

[0087] Preferably, in order to obtain effective information about the time before and after each trajectory point and capture potential features at different time scales, multi-trajectory point fusion features acquired by sliding windows on the structured data with different window sizes are added. For example, in terms of time, distance and speed, feature fusion is performed with window sizes of 3 and 5, respectively, to derive new features.

[0088] Preferably, the trajectory location data of each terminal signaling data includes different trajectory points, and one trajectory point is a piece of sample data, including information such as mobile phone number, time, longitude and latitude. For example, user A has 400 trajectory points on the day, and user B has 600 trajectory points on the day. The window size refers to the calculation of distance, speed and other features for every 3 or 5 adjacent trajectory points after sorting them in chronological order for the 400 trajectory points of user A. Then 3 and 5 are the window sizes.

[0089] Preferably, the data features within the window are calculated on the trajectory points in a sliding window manner, which can accommodate more time series information and also increase fault tolerance. For example, the distance between two adjacent trajectory points may be quite different, so the calculated speed is relatively large. If the average speed within the window is used for calculation, the average speed value within this window can be calculated, which can better reflect the true effectiveness of the features.

[0090] Preferably, the data distribution of each feature is checked in order to find a suitable threshold as a rule to filter out different types of transportation.

[0091] S4 establishes a regular feature rule strategy based on regular feature values ​​and distribution.

[0092] Preferably, a conventional feature rule strategy is established, and the specific process is: summarizing the thresholds of different types of transportation vehicles based on the distribution; and establishing a conventional feature rule strategy based on the thresholds and conventional feature values.

[0093] S5 establishes advanced feature rule strategies based on user trajectory data and distribution.

[0094] Preferably, the advanced feature rule strategy is as follows:

[0095] According to different modes of transportation, there are airplane identification rules, train or high-speed rail identification rules, car identification rules and subway identification rules;

[0096] The rules for aircraft identification are as follows: select user trajectory data that is less than 800 meters away from any airport; extract and calculate the time difference between trajectory points near the airport, and remove user trajectory data with a time difference greater than 1 hour, so as to filter out user trajectory data that may drive back and forth through the airport; select user trajectory data that stays at the airport for more than 1 hour, so as to filter out users waiting at the airport; select user trajectory data that disappears more than 5 hours from 24:00 on the same day, so as to ensure that the data is from users flying from the province to other provinces;

[0097] The train or high-speed rail identification rules are as follows: select user trajectory data in which the base station location data contains the string "train" or "high-speed rail" more than 4 times in a row; select user trajectory data in which a single sport occurs more than 10 times in a row; select user trajectory data in which the maximum movement distance is greater than 100,000 meters;

[0098] The rules for identifying cars are as follows: based on filtering the user trajectory data of trains or high-speed railways, select the user trajectory data with the string "highway" or "highway" or "toll station" appearing in the base station location data; select the user trajectory data with a daily moving distance greater than 70,000 meters;

[0099] The subway identification rules are as follows: select user trajectory data in which the base station location data contains the string "subway", "line" or "track" for more than three consecutive times; select numbers in which the base station location data contains "subway", "line" or "track" with different values ​​greater than or equal to 3, which can record the approximate location of the current base station.

[0100] Preferably, if the conditions in the airplane identification rule, the train or high-speed rail identification rule, the car identification rule and the subway identification rule are all met, then the vehicle is identified as the corresponding type of transportation.

[0101] S6 establishes expert system rules based on conventional feature rule strategies and advanced feature rule strategies.

[0102] S7 verifies and adjusts the expert system rules to obtain the optimal expert system rules.

[0103] Preferably, Figure 2 As shown in Figure 1, the expert system rule verification process is as follows:

[0104] Extract new terminal signaling data as test data;

[0105] Identify the test data based on the expert system rules and obtain the rule judgment results;

[0106] Process the test data based on S1-S2 and determine the type of transportation as the test data label;

[0107] Compare the rule judgment results with the test data labels and count the recognition accuracy of the expert system rules.

[0108] Preferably, the optimal expert system rules are obtained, specifically including:

[0109] Determine whether the recognition accuracy meets the requirements;

[0110] If the requirements are met, the expert system rules are taken as the optimal expert system rules;

[0111] If the requirements are not met, more terminal signaling data is obtained, and S1 to S7 are repeated. The rule threshold is adjusted according to the feature distribution, and the expert system rules are reorganized and verified until the recognition accuracy meets the requirements. The expert system rules at this time are output as the optimal expert system rules.

[0112] S8 identifies the signaling data of the terminal to be tested based on the optimal expert system rules to obtain the type of transportation tool.

[0113] Preferably, when identifying the signaling data of the terminal to be tested based on the optimal expert system rules, a preliminary judgment and screening is first performed through a conventional feature rule strategy, and then a further accurate identification is performed through an advanced feature rule strategy, and finally the type of vehicle is obtained.

[0114] Example 3

[0115] According to the distribution of each feature value obtained in S3, the density distribution statistics of the conventional features are performed to obtain the rule threshold that can better distinguish the types of transportation tools. Figure 3 As shown, the kernel density estimation of speed features of different vehicles (except airplanes) is shown, and the speed mean distribution diagram of different vehicles in different window sizes is shown. In order to make the most possible and relatively accurate use of speed features to effectively distinguish vehicles, the area to the left of the density estimation peak of each feature is used as the interval for adjusting the rule threshold, and the final threshold is determined in combination with the quantile statistics of the feature.

[0116] like Figure 3 In the figure, the peak value of the density estimation graph corresponding to "Train_Window 5_Speed ​​Mean" corresponds to the horizontal coordinate of 30. In order to filter out most samples with this feature, the rule threshold corresponding to this feature is set to "greater than 20". Here we only list a method for filtering a feature and its threshold. When judging a type of transportation, we need to combine multiple features for judgment.

[0117] Combining quantile statistics with features means that for each feature in the conventional features, after determining the approximate threshold range according to the kernel density estimation, the threshold correction is also performed with reference to the quantile of the feature. For example, the 0.3, 0.4, and 0.5 quantiles of the feature are checked to determine the distribution of the number of people corresponding to different quantile values ​​of the feature.

[0118] Example 4

[0119] 100 new samples are extracted to perform prediction statistics on the optimal expert system rules obtained by the present invention. The classification of the category of the vehicle belonging to the new samples by the optimal expert system rules is as follows: Figure 4 As shown in the figure, we can see that there are 15 airplane samples, all of which are predicted correctly, with a recognition accuracy of 100%; there are 21 high-speed rail or train samples, 19 of which are predicted correctly, with a recognition accuracy of 90.5%; there are 22 car samples, 18 of which are predicted correctly, with a recognition accuracy of 81.8%; there are 42 subway samples, 31 of which are predicted correctly, with a recognition accuracy of 73.8%.

[0120] In summary, the precise identification technology of transportation vehicles provided by the embodiment of this patent realizes a relatively accurate identification of the user's transportation method, and this technology can be applied to the fields of smart city management, technical reconnaissance and security control.

[0121] Example 5

[0122] like Figure 5 As shown, a vehicle identification system based on signaling data includes: a data acquisition processing module, a first identification module, a conventional feature extraction module, a conventional strategy establishment module, an advanced strategy establishment module, an expert rule establishment module, a verification adjustment module and a second identification module;

[0123] The data acquisition and processing module is used to acquire terminal signaling data and perform preprocessing to obtain user trajectory data;

[0124] A first identification module, used to identify the corresponding transportation type based on the user trajectory data;

[0125] A conventional feature extraction module is used to extract conventional feature values ​​based on user trajectory data and corresponding transportation tool types, and obtain the distribution of each feature value;

[0126] A conventional strategy establishment module is used to establish conventional feature rule strategies based on conventional feature values ​​and distribution conditions;

[0127] Advanced strategy building module, used to build advanced feature rule strategies based on user trajectory data and distribution;

[0128] An expert rule building module, used for building expert system rules based on conventional feature rule strategies and advanced feature rule strategies;

[0129] Verification and adjustment module, used to verify and adjust the expert system rules to obtain the optimal expert system rules;

[0130] The second identification module is used to identify the signaling data of the terminal to be tested based on the optimal expert system rules to obtain the type of transportation tool.

[0131] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method and system for identifying vehicles based on signaling data, which combines relevant technologies in the fields of data analysis, feature engineering, and expert system rules. By analyzing the signaling trajectory data of the user's mobile phone, conventional features and advanced features are constructed, and expert system rules are constructed based on the features for exploration. The types of vehicles to which the signaling trajectory data belongs are efficiently and accurately identified through expert system rules. It has the following beneficial effects:

[0132] 1. The present invention realizes accurate identification of the user's mode of transportation based on various features of the user's mobile phone signaling data in the communication data and expert system rules. It solves the problems that the current traditional monitoring method requires a large number of professional and technical personnel to perform data analysis and event identification, cannot quickly obtain identification results, and has difficulty in obtaining user data.

[0133] 2. The present invention is based on communication data, feature engineering and expert system rules. It does not require expensive hardware materials and has timely response. It can not only be used for traffic management and monitoring, but also can be widely used in intelligent transportation systems, smart city management, technical reconnaissance and security control and other businesses.

[0134] 3. The present invention has the advantages of high accuracy, real-time and wide applicability, can effectively improve the efficiency and accuracy of vehicle identification, and contribute to the development and progress of intelligent traffic management, technical reconnaissance and security control.

[0135] 4. The present invention combines communication big data with rule strategies and uses advanced data analysis technology and algorithms to achieve the identification and tracking of transportation vehicles. According to communication-related data, rule strategies are designed, combined with characteristic parameters, to identify the type of transportation vehicle that the user is riding. These rules can be formulated based on machine learning algorithms, pattern matching and other technologies.

[0136] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0137] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying a vehicle based on signaling data, characterized in that: include: S1 obtains terminal signaling data and performs preprocessing to obtain user trajectory data; S2 identifies the corresponding transportation type based on the user trajectory data; S3 extracts conventional feature values ​​based on the user trajectory data and the corresponding transportation tool type, and obtains the distribution of each feature value; The conventional characteristic values ​​include at least: The moving distance between adjacent trajectory points, the time difference between adjacent trajectory points, the maximum instantaneous speed between adjacent trajectory points, the straight-line distance from the first trajectory point, the cumulative sum of the distances up to the current trajectory point, the sum of the moving distances of the signaling in trajectory point windows of different sizes, the average moving distances of the signaling in trajectory point windows of different sizes, the average speed of the signaling in trajectory point windows of different sizes, and the number of base stations passed per unit time; S4 establishes a regular feature rule strategy based on the regular feature value and the distribution; Establish a general feature rule strategy. The specific process is as follows: Summarizing thresholds for different types of vehicles based on the distribution; Establishing a regular feature rule strategy based on the threshold and the regular feature value; S5 establishes an advanced feature rule strategy based on the user trajectory data and the distribution situation; The advanced feature rule strategy is specifically: According to different modes of transportation, there are airplane identification rules, train or high-speed rail identification rules, car identification rules and subway identification rules; The aircraft identification rule is as follows: select user trajectory data whose distance from any airport is less than a preset value A from the user trajectory data; extract and calculate the time difference between trajectory points near the airport, and remove the user trajectory data whose time difference is greater than a preset value B; select user trajectory data whose stay time at the airport is greater than a preset value C; select user trajectory data whose distance from 24:00 on the same day to the disappearance of the user trajectory data is greater than a preset value D hours; The train or high-speed rail identification rule is: select user trajectory data in which the base station location data in the user trajectory data continuously appears the string "train" or "high-speed rail" for more than a preset value E times; select user trajectory data of a single sport in which the user trajectory data continuously appears for more than a preset value F times; select user trajectory data with a maximum movement distance greater than a preset value G; The car identification rule is: based on filtering the user trajectory data of trains or high-speed railways, select the user trajectory data with the character string "highway" or "highway" or "toll station" appearing in the base station location data; select the user trajectory data with a daily moving distance greater than a preset value H; The subway identification rule is: select user trajectory data in which the character string "subway" or "line" or "track" appears in the base station location data for more than a preset value K times in a row; select numbers in which the different values ​​of "subway", "line" or "track" appear in the base station location data are greater than or equal to a preset value L; S6 establishes expert system rules based on the conventional feature rule strategy and the advanced feature rule strategy; S7 verifies and adjusts the expert system rules to obtain optimal expert system rules; S8 identifies the signaling data of the terminal to be tested based on the optimal expert system rule to obtain the type of transportation tool.

2. A method for identifying a vehicle based on signaling data according to claim 1, characterized in that: Get user trajectory data. The specific process is as follows: Obtaining terminal signaling trajectory data based on the terminal signaling data, and performing data cleaning to filter out data with abnormal trajectory status; The terminal signaling trajectory data after data cleaning is sorted in chronological order to obtain initial trajectory data; The initial trajectory data is subjected to outlier noise removal, ping-pong data aggregation, and trajectory completion to obtain user trajectory data.

3. A method for identifying a vehicle based on signaling data according to claim 1, characterized in that: If the conditions in the airplane identification rule, the train or high-speed rail identification rule, the car identification rule and the subway identification rule are all met, the corresponding transportation type is identified.

4. The method for identifying a vehicle based on signaling data according to claim 2, characterized in that: The expert system rule verification process is as follows: Extract new terminal signaling data as test data; Identify the test data based on the expert system rules to obtain rule determination results; Process the test data based on S1-S2 to determine the type of vehicle as a test data label; The rule determination result is compared with the test data label, and the recognition accuracy of the expert system rule is counted.

5. A method for identifying a vehicle based on signaling data according to claim 4, characterized in that: The optimal expert system rules are obtained, including: Determining whether the recognition accuracy meets the requirement; If the requirements are met, the expert system rules are used as the optimal expert system rules; If the requirements are not met, more terminal signaling data is obtained, and S1 to S7 are repeated. The rule threshold is adjusted according to the feature distribution, and the expert system rules are reorganized and verified until the recognition accuracy meets the requirements. The expert system rules at this time are output as the optimal expert system rules.

6. A vehicle identification system based on signaling data, applied to a vehicle identification method based on signaling data as claimed in any one of claims 1 to 5, characterized in that: include: A data acquisition processing module, a first identification module, a conventional feature extraction module, a conventional strategy establishment module, an advanced strategy establishment module, an expert rule establishment module, a verification adjustment module, and a second identification module; The data acquisition processing module is used to acquire terminal signaling data and perform preprocessing to obtain user trajectory data; The first identification module is used to identify the corresponding transportation type based on the user trajectory data; The conventional feature extraction module is used to extract conventional feature values ​​based on the user trajectory data and the corresponding transportation tool type, and obtain the distribution of each feature value; The conventional strategy establishment module is used to establish a conventional feature rule strategy based on the conventional feature value and the distribution situation; The advanced strategy establishment module is used to establish an advanced feature rule strategy based on the user trajectory data and the distribution situation; The expert rule establishment module is used to establish expert system rules based on the conventional feature rule strategy and the advanced feature rule strategy; The verification and adjustment module is used to verify and adjust the expert system rules to obtain the optimal expert system rules; The second identification module is used to identify the signaling data of the terminal to be tested based on the optimal expert system rule to obtain the type of transportation tool.

Images