A method and system for detecting road and bridge risks using mobile phone signals

Abstract

The present invention provides a method and system for detecting road and bridge risks using mobile phone signals, the method comprising: obtaining mobile phone positioning data and meteorological data of each user on the target road and bridge; determining the movement mode of the vehicle corresponding to each user according to the mobile phone positioning data; determining the traffic condition of the target road and bridge according to the movement mode, and determining the potential risk area of ​​the target road and bridge; inputting the mobile phone positioning data and meteorological data of each user in the potential risk area into a trained road and bridge anomaly detection model to obtain a detection result. The present invention comprehensively utilizes mobile phone signals and other relevant data to effectively and automatically monitor and evaluate the safety status of roads and bridges in real time, timely discover and deal with potential risks, thereby improving public safety and traffic efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of road and bridge monitoring, and in particular to a method and system for detecting road and bridge risks by using mobile phone signals. Background Art

[0002] As a key component of the modern transportation network, roads and bridges undertake important transportation tasks, connect various parts of the city, and promote economic and social development. However, the safety of roads and bridges has always been a matter of great concern in the fields of engineering and transportation. Roads and bridges may have structural safety problems due to a variety of reasons, including but not limited to natural disasters (such as floods and earthquakes), design defects, construction quality problems, material aging, excessive loads, and lack of timely maintenance and inspection. In severe weather conditions, such as ice and snow, heavy rain, etc., roads and bridges may freeze or be damaged, resulting in vehicle loss of control, accidents, and traffic jams.

[0003] Traditional methods of road and bridge inspection mainly rely on regular visual inspections, structural monitoring using various sensors, and regular maintenance. These methods can be divided into two types: direct and indirect. Direct inspection methods are regular inspections conducted by professionals, including direct observation of the surface and structure of the road and bridge. This method is simple and direct, but may miss some internal or imperceptible defects. Indirect inspection methods predict possible structural problems by installing stress, strain and other data sensors at key parts of the road and bridge, monitoring the changes in these indicators, but it requires expensive sensors and complex installations to monitor real-time road and bridge risks, and existing technologies cannot cope with the safety hazards brought about by sudden natural disasters. Summary of the invention

[0004] In order to overcome the deficiencies of the prior art, the object of the present invention is to provide a method and system for detecting road and bridge risks using mobile phone signals.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for detecting road and bridge risks using mobile phone signals, comprising:

[0007] Obtain mobile phone location data and weather data of each user on the target road and bridge;

[0008] Determine the movement mode of the vehicle corresponding to each user according to the mobile phone positioning data;

[0009] Determining the traffic condition of the target road and bridge according to the movement pattern, and determining the potential risk area of ​​the target road and bridge;

[0010] The mobile phone positioning data of each user in the potential risk area and the meteorological data are input into the trained road and bridge anomaly detection model to obtain the detection results.

[0011] Preferably, obtaining the mobile phone positioning data and weather data of each user on the target road or bridge includes:

[0012] Acquiring registration information of each of the user's mobile phones in the mobile network based on the network operator; the registration information includes signal strength and base station information;

[0013] Using the GPS module built into the mobile phone of each user to obtain the location data of users passing through the target road and bridge;

[0014] Correcting the location data according to the registration information to obtain the mobile phone positioning data; the mobile phone positioning data includes latitude and longitude, timestamp and moving speed;

[0015] The real-time meteorological data is obtained from the meteorological department; the meteorological data includes temperature, humidity, wind speed and precipitation.

[0016] Preferably, determining the movement mode of the vehicle corresponding to each user according to the mobile phone positioning data includes:

[0017] Constructing a convolutional neural network; the convolutional neural network comprises an input layer, a hidden layer, a fully connected layer and an output layer connected in sequence; the hidden layer is composed of at least one convolutional layer and at least one pooling layer arranged at intervals;

[0018] The convolutional neural network is trained using training sample data to obtain a trained motion pattern recognition model; the training sample data includes positioning data of each mobile phone and corresponding vehicle motion pattern data;

[0019] The mobile phone positioning data is input into a trained motion pattern recognition model to obtain the motion pattern of the vehicle; the motion pattern includes a regular driving state and an irregular driving state; the irregular driving state includes a slow driving and deceleration state, a stop and wait state, an abnormal behavior state and a lane change state.

[0020] Preferably, determining the traffic condition of the target road bridge according to the movement pattern, and determining the potential risk area of ​​the target road bridge includes:

[0021] The time series of the mobile phone positioning data corresponding to the unconventional driving state is used as a Markov random process;

[0022] Based on the Markov random process, a dangerous event chain model is formed by determining the initial probabilities of different states in the irregular driving state and the distribution of transition probabilities between states in the chain;

[0023] According to the road danger event state and the vehicle driving risk coefficient distribution in the danger event chain model, the range representation hidden state of the road section passed at different times is set respectively, and the range representation hidden state is divided into levels;

[0024] Establishing a prediction model based on a training data set; the training data set includes existing road hazard event status;

[0025] Calculate the number of times the training data set is different from the state at the previous moment to the next moment, and use the formula to calculate the state transfer matrix and the observation variable transfer probability matrix;

[0026] Determine an initial probability vector according to the state transfer matrix, the observed variable transfer probability matrix and the prediction model;

[0027] Determining a predicted value of a hidden state of a range representation according to the prediction model;

[0028] Predicting the traffic condition of a real dangerous event according to the initial probability vector and the hidden state prediction value;

[0029] The minimum bounding rectangle of the data points of each mobile phone positioning data corresponding to the traffic condition is determined, and the minimum bounding rectangle is determined as the potential risk area.

[0030] Preferably, the mobile phone positioning data of each user in the potential risk area and the meteorological data are input into a trained road and bridge anomaly detection model to obtain a detection result, including:

[0031] Obtain the preset mobile phone positioning data training set and meteorological data training set;

[0032] Build a deep learning model for multi-object segmentation;

[0033] The deep learning model for multi-target segmentation is trained according to the mobile phone positioning data training set and the meteorological data training set respectively to obtain a trained first classifier and a second classifier;

[0034] Cascading the trained first classifier and the second classifier to obtain a classification network;

[0035] Connecting the trained LSTM neural network after the classification network to obtain the road and bridge anomaly detection model;

[0036] The mobile phone positioning data and the meteorological data of each user in the potential risk area are detected according to the road and bridge anomaly detection model to obtain the detection result.

[0037] A system for detecting road and bridge risks using mobile phone signals, comprising:

[0038] An acquisition module is used to obtain mobile phone positioning data and meteorological data of each user on the target road and bridge;

[0039] A pattern recognition module, used to determine the movement pattern of the vehicle corresponding to each user according to the mobile phone positioning data;

[0040] An area positioning module, used to determine the traffic condition of the target road and bridge according to the movement pattern, and determine the potential risk area of ​​the target road and bridge;

[0041] The detection module is used to input the mobile phone positioning data of each user in the potential risk area and the meteorological data into the trained road and bridge anomaly detection model to obtain the detection result.

[0042] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0043] The present invention provides a method and system for detecting road and bridge risks using mobile phone signals, the method comprising: obtaining mobile phone positioning data and meteorological data of each user on the target road and bridge; determining the movement mode of the vehicle corresponding to each user according to the mobile phone positioning data; determining the traffic condition of the target road and bridge according to the movement mode, and determining the potential risk area of ​​the target road and bridge; inputting the mobile phone positioning data and meteorological data of each user in the potential risk area into a trained road and bridge anomaly detection model to obtain a detection result. The present invention comprehensively utilizes mobile phone signals and other relevant data to effectively and automatically monitor and evaluate the safety status of roads and bridges in real time, timely discover and deal with potential risks, thereby improving public safety and traffic efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] 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 will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0045] Figure 1 A flow chart of a method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0046] 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.

[0047] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] Figure 1 A flow chart of a method provided by an embodiment of the present invention, such as Figure 1 As shown, the present invention provides a method for detecting road and bridge risks using mobile phone signals, comprising:

[0049] Step 100: Obtaining mobile phone positioning data and weather data of each user on the target road or bridge;

[0050] Step 200: determining the movement mode of the vehicle corresponding to each user according to the mobile phone positioning data;

[0051] Step 300: determining the traffic condition of the target road or bridge according to the movement pattern, and determining a potential risk area of ​​the target road or bridge;

[0052] Step 400: inputting the mobile phone positioning data of each user in the potential risk area and the meteorological data into the trained road and bridge anomaly detection model to obtain the detection result.

[0053] Preferably, obtaining the mobile phone positioning data and weather data of each user on the target road or bridge includes:

[0054] Acquiring registration information of each of the user's mobile phones in the mobile network based on the network operator; the registration information includes signal strength and base station information;

[0055] Using the GPS module built into the mobile phone of each user to obtain the location data of users passing through the target road and bridge;

[0056] Correcting the location data according to the registration information to obtain the mobile phone positioning data; the mobile phone positioning data includes latitude and longitude, timestamp and moving speed;

[0057] The real-time meteorological data is obtained from the meteorological department; the meteorological data includes temperature, humidity, wind speed and precipitation.

[0058] Mobile phone positioning data acquisition:

[0059] Specifically, this embodiment uses the built-in GPS or other positioning systems (such as GLONASS, Beidou, etc.) of the mobile phone to obtain the precise location data of the user passing the target road bridge. These data include latitude and longitude, timestamp and moving speed. By cooperating with the mobile network operator, the registration information of the user's mobile phone in the mobile network can be obtained, including signal strength and base station information. This embodiment uses the registration information to assist positioning when the GPS signal is poor.

[0060] Furthermore, this embodiment obtains real-time meteorological data from the meteorological department or a third-party meteorological service provider, including temperature, humidity, wind speed, precipitation, etc. This embodiment takes into account that these factors may affect the safety status and traffic behavior of roads and bridges.

[0061] Furthermore, this embodiment ensures that all data (mobile phone location data and weather data) have a unified time stamp to facilitate accurate time series analysis. Time synchronization is critical to tracking and analyzing the development of events.

[0062] Preferably, determining the movement mode of the vehicle corresponding to each user according to the mobile phone positioning data includes:

[0063] Constructing a convolutional neural network; the convolutional neural network comprises an input layer, a hidden layer, a fully connected layer and an output layer connected in sequence; the hidden layer is composed of at least one convolutional layer and at least one pooling layer arranged at intervals;

[0064] The convolutional neural network is trained using training sample data to obtain a trained motion pattern recognition model; the training sample data includes positioning data of each mobile phone and corresponding vehicle motion pattern data;

[0065] The mobile phone positioning data is input into a trained motion pattern recognition model to obtain the motion pattern of the vehicle; the motion pattern includes a regular driving state and an irregular driving state; the irregular driving state includes a slow driving and deceleration state, a stop and wait state, an abnormal behavior state and a lane change state.

[0066] Furthermore, the motion modes of this embodiment include:

[0067] Normal driving state, that is, normal vehicle driving: the vehicle drives on the road bridge at a stable speed without frequent acceleration or deceleration. This mode indicates smooth traffic without obvious congestion or abnormal phenomena.

[0068] Slow Driving and Slow Down: Vehicles are moving at slower than normal speeds, possibly due to traffic congestion, road construction or other obstructions. Slow Driving mode indicates heavy traffic or other problems on a bridge.

[0069] Stop and Wait: The vehicle comes to a complete stop or frequently stops and starts. This often occurs during rush hour or when traffic is blocked by lights, accidents or road conditions.

[0070] Lane change status: By analyzing the direction changes in the positioning data, it is possible to identify whether the vehicle is changing lanes. Frequent lane changes indicate complex traffic conditions or problems with the driver trying to avoid certain road sections.

[0071] Abnormal behavior: includes behaviors that do not conform to normal traffic patterns, such as parking in illegal areas, turning around on bridges, or driving on prohibited roads. These behaviors may indicate that the driver is not adapting to the current road conditions or violating traffic regulations. Exemplarily, abnormal behaviors in this embodiment also include situations where multiple vehicles suddenly stop or collapse (this means that the speed of the vehicle drops to zero instantly).

[0072] Specifically, this embodiment can make a more accurate assessment of the traffic conditions on the road and bridge by analyzing these movement patterns, and indirectly monitor the load and potential risks of the road and bridge. This information is very valuable for road and bridge maintenance and safety management, and can help relevant departments to adjust traffic management measures in a timely manner or conduct necessary road and bridge inspections and repairs.

[0073] Preferably, determining the traffic condition of the target road bridge according to the movement pattern, and determining the potential risk area of ​​the target road bridge includes:

[0074] The time series of the mobile phone positioning data corresponding to the unconventional driving state is used as a Markov random process;

[0075] Based on the Markov random process, a dangerous event chain model is formed by determining the initial probabilities of different states in the irregular driving state and the distribution of transition probabilities between states in the chain;

[0076] According to the road danger event state and the vehicle driving risk coefficient distribution in the danger event chain model, the range representation hidden state of the road section passed at different times is set respectively, and the range representation hidden state is divided into levels;

[0077] Establishing a prediction model based on a training data set; the training data set includes existing road hazard event status;

[0078] Calculate the number of times the training data set is different from the state at the previous moment to the next moment, and use the formula to calculate the state transfer matrix and the observation variable transfer probability matrix;

[0079] Determine an initial probability vector according to the state transfer matrix, the observed variable transfer probability matrix and the prediction model;

[0080] Determining a predicted value of a hidden state of a range representation according to the prediction model;

[0081] Predicting the traffic condition of a real dangerous event according to the initial probability vector and the hidden state prediction value;

[0082] The minimum bounding rectangle of the data points of each mobile phone positioning data corresponding to the traffic condition is determined, and the minimum bounding rectangle is determined as the potential risk area.

[0083] Specifically, the Markov random process applies the hidden Markov model (HMM), which is a statistical model used to describe a Markov process with hidden unknown parameters. A standard hidden Markov model has five variables (q, O, X, A, h). Among them, q represents the hidden state, which means the state that cannot be directly observed; O represents the observable state, which means the state that can be directly observed. In the model, these states: have a specific relationship with the hidden state, X is the initial state probability matrix, which represents the state transition probability distribution of the hidden state at the initial moment: A is the hidden state transition probability matrix, which is used to describe the probability of mutual conversion between states in the model; N represents the number of hidden states, then A ij =P(S j |S i ) means that at a certain moment, the state is S i Under the condition, the state at time t+1 is S j The probability of; h is the observed state transition probability matrix, let M represent the number of observable states, then h ij =P(O i |S j ) means that at time t, the implicit state is q j Under the condition, the observation state is O i probability.

[0084] Preferably, the mobile phone positioning data of each user in the potential risk area and the meteorological data are input into a trained road and bridge anomaly detection model to obtain a detection result, including:

[0085] Obtain the preset mobile phone positioning data training set and meteorological data training set;

[0086] Build a deep learning model for multi-object segmentation;

[0087] The deep learning model for multi-target segmentation is trained according to the mobile phone positioning data training set and the meteorological data training set respectively to obtain a trained first classifier and a second classifier;

[0088] Cascading the trained first classifier and the second classifier to obtain a classification network;

[0089] Connecting the trained LSTM neural network after the classification network to obtain the road and bridge anomaly detection model;

[0090] The mobile phone positioning data and the meteorological data of each user in the potential risk area are detected according to the road and bridge anomaly detection model to obtain the detection result.

[0091] Optionally, this embodiment combines the time series analysis capability of the LSTM network, and the detection results may analyze the long-term trends and patterns of road and bridge traffic conditions and safety risks.

[0092] In order to evaluate the prediction performance of the road and bridge anomaly detection model, this embodiment uses the root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) to evaluate the error between the predicted value and the true value. The root mean square error is expressed as:

[0093]

[0094] The mean absolute error is expressed as:

[0095]

[0096] The mean absolute percentage error is expressed as:

[0097]

[0098] Where n is the number of samples; y i is the predicted value; is the true value.

[0099] Furthermore, this embodiment provides early warning information based on the abnormal detection results to alert relevant departments to potential dangers in specific areas or time periods. At the same time, response suggestions may be provided, such as adding traffic signs, adjusting traffic lights, etc., to improve traffic safety and smoothness.

[0100] Furthermore, the present embodiment also provides an open interface for other applications to call early warnings to serve the society, and the present embodiment can collect other official disaster information based on AI algorithms. When users or subscribers go to areas near disasters, the present embodiment can forward the abnormal detection results or disaster information to users or subscribers in the target area for their awareness.

[0101] Corresponding to the above method, this embodiment further provides a system for detecting road and bridge risks using mobile phone signals, including:

[0102] An acquisition module is used to obtain mobile phone positioning data and meteorological data of each user on the target road and bridge;

[0103] A pattern recognition module, used to determine the movement pattern of the vehicle corresponding to each user according to the mobile phone positioning data;

[0104] An area positioning module, used to determine the traffic condition of the target road and bridge according to the movement pattern, and determine the potential risk area of ​​the target road and bridge;

[0105] The detection module is used to input the mobile phone positioning data of each user in the potential risk area and the meteorological data into the trained road and bridge anomaly detection model to obtain the detection result.

[0106] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0107] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for detecting road and bridge risks using mobile phone signals. It is characterized in that include: Obtain mobile phone location data and weather data of each user on the target road and bridge; Determine the movement mode of the vehicle corresponding to each user according to the mobile phone positioning data; Determining the traffic condition of the target road and bridge according to the movement pattern, and determining the potential risk area of ​​the target road and bridge; Inputting the mobile phone positioning data of each user in the potential risk area and the meteorological data into the trained road and bridge anomaly detection model to obtain the detection result; Determining the movement mode of the vehicle corresponding to each user according to the mobile phone positioning data includes: Constructing a convolutional neural network; the convolutional neural network comprises an input layer, a hidden layer, a fully connected layer and an output layer connected in sequence; the hidden layer is composed of at least one convolutional layer and at least one pooling layer arranged at intervals; The convolutional neural network is trained using training sample data to obtain a trained motion pattern recognition model; the training sample data includes positioning data of each mobile phone and corresponding vehicle motion pattern data; The mobile phone positioning data is input into a trained motion pattern recognition model to obtain the motion pattern of the vehicle; the motion pattern includes a regular driving state and an irregular driving state; the irregular driving state includes a slow driving and deceleration state, a stop and wait state, an abnormal behavior state and a lane change state; Determining the traffic condition of the target road and bridge according to the movement pattern, and determining the potential risk area of ​​the target road and bridge, including: The time series of the mobile phone positioning data corresponding to the unconventional driving state is used as a Markov random process; Based on the Markov random process, a dangerous event chain model is formed by determining the initial probabilities of different states in the irregular driving state and the distribution of transition probabilities between states in the chain; According to the road danger event state and the vehicle driving risk coefficient distribution in the danger event chain model, the range representation hidden state of the road section passed at different times is set respectively, and the range representation hidden state is divided into levels; Establishing a prediction model based on a training data set; the training data set includes existing road hazard event status; Calculate the number of times the training data set is different from the state at the previous moment to the next moment, and use the formula to calculate the state transfer matrix and the observation variable transfer probability matrix; Determine an initial probability vector according to the state transfer matrix, the observed variable transfer probability matrix and the prediction model; Determining a predicted value of a hidden state of a range representation according to the prediction model; Predicting the traffic condition of a real dangerous event according to the initial probability vector and the hidden state prediction value; The minimum bounding rectangle of the data points of each mobile phone positioning data corresponding to the traffic condition is determined, and the minimum bounding rectangle is determined as the potential risk area.

2. The method for detecting road and bridge risks using mobile phone signals according to claim 1, It is characterized in that Obtain the mobile phone location data and weather data of each user on the target road and bridge, including: Acquiring registration information of each of the user's mobile phones in the mobile network based on the network operator; the registration information includes signal strength and base station information; Using the GPS module built into the mobile phone of each user to obtain the location data of users passing through the target road and bridge; Correcting the location data according to the registration information to obtain the mobile phone positioning data; the mobile phone positioning data includes latitude and longitude, timestamp and moving speed; The real-time meteorological data is obtained from the meteorological department; the meteorological data includes temperature, humidity, wind speed and precipitation.

3. The method for detecting road and bridge risks using mobile phone signals according to claim 1, It is characterized in that The mobile phone positioning data of each user in the potential risk area and the meteorological data are input into the trained road and bridge anomaly detection model to obtain the detection results, including: Obtain the preset mobile phone positioning data training set and meteorological data training set; Build a deep learning model for multi-object segmentation; The deep learning model for multi-target segmentation is trained according to the mobile phone positioning data training set and the meteorological data training set respectively to obtain a trained first classifier and a second classifier; Cascading the trained first classifier and the second classifier to obtain a classification network; Connecting the trained LSTM neural network after the classification network to obtain the road and bridge anomaly detection model; The mobile phone positioning data and the meteorological data of each user in the potential risk area are detected according to the road and bridge anomaly detection model to obtain the detection result.

4. A system for implementing the method for detecting road and bridge risks using mobile phone signals as described in any one of claims 1 to 3, It is characterized in that include: An acquisition module is used to obtain mobile phone positioning data and meteorological data of each user on the target road and bridge; A pattern recognition module, used to determine the movement pattern of the vehicle corresponding to each user according to the mobile phone positioning data; An area positioning module, used to determine the traffic condition of the target road and bridge according to the movement pattern, and determine the potential risk area of ​​the target road and bridge; The detection module is used to input the mobile phone positioning data of each user in the potential risk area and the meteorological data into the trained road and bridge anomaly detection model to obtain the detection result.

Images