TWI652959B - System and method for estimating vehicular traffic by cellular data - Google Patents

Abstract

A system and method for estimating road traffic flow by using action signaling data is to estimate the traffic flow according to the weather data, time data, action signaling data and road segment data on the road segment by estimating the traffic flow model.

Description

System and method for estimating road traffic volume by using action signaling data

The present invention relates to a technique for estimating road traffic flow, and more particularly to a system and method for estimating road traffic using motion signaling data.

At present, if the public sector traffic control center can know the traffic volume in the bottleneck area, it can speed up the vehicle redirection or the vehicle temporarily stops entering the area, so as to avoid too many vehicle cards causing road congestion in a certain area. Passers-by can know which roads have too much traffic and can change channels in advance to reduce travel time. However, the existing way to know the traffic flow is to set up hardware such as vehicle detector VD (vehicle detector) or electronic tag detector (eTag Detector) operation. Traffic flow.

However, the installation of a vehicle detector or an electronic tag detector can be costly in terms of hardware and labor. Therefore, a vehicle detector VD (vehicle detector) or an eTag Detector is not installed on each road segment.

Therefore, how to estimate the traffic volume of the road segment on the road section where the vehicle detector VD (vehicle detector) or the electronic tag detector (eTag Detector) is not installed is the technical problem to be solved by the present invention.

To overcome the deficiencies of the prior art, the present invention provides a system for estimating road traffic using motion signaling data, comprising: a data receiving module for receiving real-time weather data, real-time data, and instant action signaling. The data-corresponding module is used to obtain the real-time road segment data corresponding to the instant-action signaling data; and the road segment traffic flow estimation module is based on the estimated traffic flow model according to the real-time weather data, The real-time data, the instant-action signaling data, and the real-time road segment data estimate the traffic volume of the road segment.

In an embodiment, the data receiving module further receives the instantaneous traffic.

In an embodiment, the system further includes: a database module for storing a plurality of the real-time weather data, the instant time data, the instant action signaling data, the instantaneous traffic flow, and the real-time road segment data, and It is defined as historical weather data, historical time data, historical action signaling data, historical traffic flow and historical road data; and historical data training module, which is the historical weather data in the database module, and the historical time. The data, the historical action signaling data, the historical traffic volume, and the historical road segment data are obtained through the neural network to estimate the traffic flow model.

In one embodiment, the neural network is one of a recursive neural network (RNN) or a deep neural network (DNN) or a combination thereof.

In one embodiment, the traffic flow estimation module of the road section further includes: after continuously receiving the real-time weather data, the real-time data, the real-time action signaling data, and the real-time road segment data reach a period of time, the traffic volume of the road section is pushed The estimated module is based on the estimated traffic flow model for the received time period The real-time weather data, the real-time data, the instant-action signaling data, and the real-time road segment data are estimated to obtain the traffic volume of the road segment.

The invention further provides a method for estimating road traffic volume by using action signaling data, which comprises the following steps: (1) obtaining real-time data of weather data, time data and action signaling data; and (2) obtaining the action signaling data. The real-time data of the road segment data corresponding to the real-time data; and (3) using a estimator traffic flow model to estimate the traffic volume of the road segment based on the weather data, the time data, the action signaling data and the real-time data of the road segment data .

In an embodiment, the real-time data in the step (3) is real-time data in a period of time, and the estimated vehicle traffic model is used to estimate the real-time data received during the period of time to obtain the Road traffic.

In an embodiment, the method further includes transmitting the traffic volume of the road section to the display end to display the traffic volume of the road section.

In an embodiment, the method for estimating the traffic flow model includes the following steps: (3-1) obtaining weather data, time data, action signaling data, and historical data of traffic flow; (3-2) obtaining the action The historical data of the link data corresponding to the historical data of the signaling data; and (3-3) the historical data obtained by the steps (3-1) and (3-2) are obtained through the neural network. Traffic flow model.

In an embodiment, the neural network in the step (3-3) is one of a recursive neural network (RNN) or a deep neural network (DNN) or a combination thereof.

The technical features of the present invention are as follows:

1. The present invention analyzes the amount of action signaling data in historical data, and the type of road segment State (national expressway, urban or scenic section, etc.), number of lanes, weather data, time data (spike or off-peak time) to find out the relationship with traffic flow, so that real-time data can be calculated and estimated flow.

2. The invention can still find similar number of action signaling data, road type (national fast track, urban area or scenic section, etc.) and road section lane on the road section without hardware detection equipment such as VD and eTag. The neural network model trained by the number, weather data, time data (spike or off-peak time) estimates the traffic flow.

3. The invention proposes a road segment traffic calculation module, which is different according to the road segment type (national expressway, urban area or scenic section, etc.), road segment number, weather data, time data (spike or off-peak time). The characteristic sections use different parameter settings to more accurately calculate the traffic flow.

It can be seen from the above that the present invention transmits traffic detected by a vehicle detector VD (vehicle detector) or an eTag Detector on certain sections of the historical data and the action on the road section. Number of signaling data, road type (national fast track, urban or scenic section, etc.), number of lanes, weather data and time data (spike or off-peak time) using a neural algorithm to find the corresponding push Estimating the traffic flow model, in the remaining VD-free or eTag detector sections, the traffic volume of the road section can be estimated through a similar estimated traffic flow model. The estimated traffic volume of the road section can be presented through the APP or the webpage. Save time by either posting traffic measures to avoid road congestion.

10‧‧‧Server side

11‧‧‧ Display

101‧‧‧ data receiving module

102‧‧‧ Signaling data section corresponding module

103‧‧‧Database Module

104‧‧‧Historical Data Training Module

105‧‧‧Street traffic flow estimation module

106‧‧‧First Network Transmission Module

111‧‧‧Second network transmission module

112‧‧‧Car flow display module

201‧‧‧ Input parameters

202‧‧‧ true value

203‧‧‧Recommended traffic flow learning

S401~S404‧‧‧Steps

S501~S503‧‧‧Steps

Figure 1 is an illustration of the use of mobile signaling data to estimate road traffic Schematic diagram of the system; FIG. 2 is a schematic diagram of a method for learning the estimated traffic flow model by using the neural network training of the present invention; FIG. 3 is a schematic diagram showing the traffic volume of the road section according to the present invention; A flow chart of a method for estimating road traffic using the action signaling data of the present invention; and FIG. 5 is a flow chart showing steps of obtaining the estimated vehicle flow model of the present invention.

The other embodiments of the present invention will be readily understood by those skilled in the art from this disclosure.

It is to be understood that the structure, the proportions, the size, and the like of the present invention are intended to be used in conjunction with the disclosure of the specification, and are not intended to limit the invention. The conditions are limited, so it is not technically meaningful. Any modification of the structure, change of the proportional relationship or adjustment of the size should remain in this book without affecting the effects and the objectives that can be achieved by the present invention. The technical content disclosed in the invention can be covered. In the meantime, the terms "instant" and "history" as used in this specification are for convenience of description only, and are not intended to limit the scope of the invention, and the relative relationship is changed or adjusted. Substantially changing the technical content is considered to be within the scope of the invention.

Please refer to FIG. 1 for the use of the mobile signaling information of the present invention. A schematic architecture diagram of a system for estimating road traffic flow includes: a server end 10 for estimating road traffic flow; and a display end 11 for presenting the road traffic flow.

The action signaling data utilized by the present invention is that a mobile device (such as a mobile phone) can be automatically connected to a base station when the mobile device is powered on, and the base station obtains action signaling data by receiving a signaling signal of the mobile device. The signal is the signal that the mobile device is connected to the base station.

The server terminal 10 includes: a data receiving module 101 for receiving real-time weather data, real-time data, instant action signaling data, and real-time traffic, wherein the mobile signaling data is determined by coordinate position and identification. The code component; the signaling data segment corresponding module 102 is configured to receive the real-time weather data, the real-time data, the instant action signaling data, and the real-time traffic flow transmitted by the data receiving module 101, and according to the instant action signaling The coordinate position of the data is obtained by the corresponding real-time road segment data, wherein the road segment data is composed of a road name, a road segment number and a road segment type; the database module 103 is configured to store a plurality of the real-time weather data and an instant time. Data, real-time action signaling data, real-time traffic flow and the real-time road segment data, and respectively define it as historical weather data, historical time data, historical action signaling data, historical traffic flow and historical road segment data; historical data training learning model The group 104 is historical weather data, historical time data, and historical action signaling data in the database module 103. The historical traffic flow and historical road segment data are obtained through a neural network to estimate a traffic flow model; the road traffic flow estimation module 105 is configured to receive the real-time weather data transmitted by the signaling module segment corresponding module 102, The instant time data, the instant line And the real-time weather data, the real-time data, the instant-action signaling data, and the instant information are obtained from the estimated traffic flow model obtained from the historical data training module 104. The data of the road section is similar to the data to estimate the traffic volume of the road section; the first network transmission module 106 is configured to transmit the traffic volume of the road section estimated by the road traffic estimation module 105 to the display end 11.

The display end 11 includes a second network transmission module 111 for receiving the traffic volume of the road section transmitted by the server end 10, and a traffic flow display module 112 for displaying the traffic volume of the road section.

In an embodiment, the server end 10 is a cloud server.

In an embodiment, the data receiving module 101 is a wired or wireless entity receiving interface.

In one embodiment, the database module 103 can be a solid state hard disk or a database (such as SQL or Access Database) stored in a hard disk.

In an embodiment, the first network transmission module 106 is a wired or wireless entity receiving interface.

In an embodiment, the signaling data segment corresponding module 102, the historical data training learning module 104, and the road traffic estimation module 105 can be functions generated by software or instructions running by the processor in the server terminal 10. .

In one embodiment, the display terminal 11 is any one of a mobile phone application, a computer webpage, and a navigator page.

In this embodiment, the coordinate position of the action signaling data represents the location of the mobile device user, and the base station analyzes the location, weather, and time of the mobile device user only on the road segment according to the coordinate position of the mobile signaling data.

In this embodiment, the instantaneous traffic volume received by the data receiving module 101 is obtained by a Vehicle Detector (VD) and an eTag Detector on the road segment.

In this embodiment, the real-time data received by the data receiving module 101 and the historical time data in the database module 103 can be classified into a peak time or an off-peak time.

In an embodiment, the road segment type obtained by the signaling data segment corresponding module 102 is a national expressway, an urban area, or a scenic road section.

In an embodiment, the road segment type obtained by the signaling data segment corresponding module 102 is classified according to the highway classification published by the General Administration of Highways, and can be classified into national roads, provincial roads, city roads, county roads, and district roads according to the administrative system. The township roads can be classified into highways, expressways, major roads, secondary roads and regional roads according to the transportation function. According to the geographical environment, they can be classified into highway system, expressway system, round-the-island road system, cross-road system, and running road. Systems, coastal road systems and liaison roads.

In an embodiment, the information about the instant road segment obtained by the signaling data segment corresponding module 102 is obtained by dividing the whole road network into different grids, and then confirming the coordinates of the instant action signaling data. The grid in which the location is located, and calculates the linear distance between the road segment in the grid and the location of the instant action signaling data. After the distance sorting, the road segment taking the minimum distance is the road segment where the instant action signaling data is located, and simultaneously obtains the same The relevant information of the road section includes the road name, the number of lanes and the type of the road segment.

In an embodiment, the neural network of the historical data training module 104 is performed by a recurrent neural network (RNN) or a deep neural network. One of the roads (DNN) or a combination thereof.

Please refer to FIG. 2 , which is a schematic diagram of a method for obtaining the estimated traffic flow model through the neural network based on the historical data training module 104 in FIG. 1 . The historical data training learning module 104 needs to be configured to the database. Group 103 reads historical weather data, historical time data, historical action signaling data, and historical road data as input parameters 201 of the neural network and reads historical traffic to the database module 103 as a neural network. The true value 202 can be used to estimate the traffic flow 203 of the neural network to obtain the estimated traffic flow model for the road traffic estimation module 105 to receive the real data after the estimation Find similar data in the traffic flow model to estimate the traffic volume of the road segment.

In an embodiment, the data of each estimated traffic flow learning 203 is composed of the input parameter 201 and the true value 202, and the data format of each estimated traffic flow learning 203 is: 0, 09: 50, 3, and the city Road, 150, 40 (weather, time, number of lanes, road type, number of mobile signaling data, traffic flow), the last field is the true value of the traffic volume in the first five fields, where the weather Different weather patterns can be set by different numbers, for example, 0 is sunny, 1 is rainy, 2 is snowing, etc., but is not limited thereto.

In an embodiment, the data format of each estimated traffic flow learning 203 is: 0, 09: 50, 3, city road, citizen road, second section, 150, 40 (weather, time, number of lanes, road type) State, road name, road segment, number of mobile signaling data, traffic flow), the last field is the true value of traffic flow in the first 7 fields, but is not limited to this.

In an embodiment, the historical data training learning module 104 to the capital The library module 103 reads three months of historical weather data, historical time data, historical action signaling data, historical traffic flow, and historical road segment data to perform a neural network learning training to estimate a traffic flow model, wherein The time range in which the historical data training learning module 104 reads the historical data from the database module 103 varies according to the user's needs, and is not limited to three months. In this embodiment, three months is a preferred value.

In the previous embodiment, the historical data training learning module 104 completes reading 3 months of historical weather data, historical time data, historical action signaling data, historical traffic flow, and historical sections to the database module 103. After the data-based neural network learning training is performed to estimate the traffic flow model, the database module 103 accumulates one month of historical weather data, historical time data, historical action signaling data, historical traffic flow, and historical sections for each new storage. The historical data training learning module 104 re-reads the historical weather data, historical time data, historical action signaling data, historical traffic flow, and historical road data of the past three months to the database module 103 to perform a neural network. The road learns training and updates the estimated traffic flow model to ensure that the estimated quality of the traffic flow model is estimated.

In an embodiment, the traffic flow estimation module 105 of the road section continuously pushes the real-time weather data, the real-time data, the real-time action signaling data, and the real-time road segment data for a period of time, so that the traffic volume of the road section is pushed. The estimation module 105 estimates the real-time weather data, the real-time data, the instant-action signaling data, and the real-time road segment data that have been received during the period of time according to the estimated vehicle traffic model to obtain the road traffic volume. In this embodiment, the period of time may be 5 minutes, which is known to those skilled in the art. The length of time can be determined according to its needs.

Please refer to FIG. 3 , which is a schematic diagram of the traffic flow display module 112 in the display end 11 of FIG. 1 showing the traffic volume of the road segment. The traffic flow on each road segment can represent the congestion level of the traffic flow in different colors.

Referring to FIG. 4, a flow chart of a method for estimating road traffic using the action signaling data of the present invention includes: S401: obtaining real-time data of weather data, time data, and action signaling data, wherein the action letter The data is composed of the coordinate position and the identification code; S402: the real-time data of the road segment data corresponding to the coordinate position of the real-time data of the action signaling data, wherein the road section data is the road name, the number of road segments and the road segment type State composition; S403: using an estimated traffic flow model to estimate the traffic volume of the road segment based on the weather data, the time data, the action signaling data, and the real-time data of the road segment data; and S404: transmitting the traffic volume to the road segment to The display shows the traffic volume of the road section.

In the embodiment, the road traffic volume in the step S403 is used to find information similar to the real-time weather data, the real-time data, the real-time action signaling data, and the real-time road segment data from the estimated traffic flow model. To estimate the traffic volume of the road section.

In an embodiment, the steps S401 and S402 obtain the five-minute weather data, the time data, the action signaling data, and the real-time data of the road segment data. The estimated traffic flow model in the step S403 is every five minutes. Correct The five-minute weather data, the time data, the action signaling information and the real-time data of the road section data are used to estimate the traffic volume of the road section. In the present embodiment, five minutes is only an example, and the present invention is not limited thereto.

In an embodiment, the instant data system of the link data corresponding to the coordinate position of the real-time data of the action signaling data is obtained by the following method: first dividing the whole road network into different grids, and secondly confirming the action The grid where the signaling data is located, and calculates the linear distance between the link in the grid and the coordinates of the action signaling data. After the distance is sorted, the link with the smallest distance is the segment where the action signaling data is located, and simultaneously obtained. The relevant information of the road section includes the road name, the number of lanes and the type of the road segment.

In an embodiment, the display terminal 11 can be a computer, a mobile phone, a tablet computer or a navigator, and the traffic display module 112 can be any one of a mobile phone application, a computer webpage, a browser webpage, and a navigator page. By.

Please refer to FIG. 5 , which is a flow chart of the steps of obtaining the estimated traffic flow model in step S403 of FIG. 4 , which includes: S501: obtaining weather data, time data, action signaling data, and historical data of traffic flow; S502 : obtaining historical data of the link data corresponding to the coordinate position of the historical data of the action signaling data; and S503: the weather data obtained by the steps S501 and S502, the time data, the action signaling data, and the traffic flow And the historical data of the road section data is obtained through the neural network to estimate the traffic flow model.

In an embodiment, the neural network in step S503 is one of a recursive neural network (RNN) or a deep neural network (DNN) or a combination thereof.

In this embodiment, the weather data, the time data, the action signaling data, and the historical data of the road segment data in the step S503 are used as input parameters of the neural network, and the historical data system of the traffic flow is used. As the true value of the neural network of this kind, the estimated traffic flow learning 203 of the neural network like FIG. 2 is performed, and the estimated traffic flow model is obtained.

In an embodiment, the data of each estimated traffic flow learning 203 is composed of the input parameter and the true value, and the data format of each estimated traffic flow learning 203 is: 0, 09: 50, 3, the market road , 150, 40 (weather, time, number of lanes, road type, number of mobile signaling data, traffic flow), the last field is the true value of traffic flow in the first five fields, where the weather can Different numbers set different weather patterns, for example, 0 is sunny, 1 is rainy, 2 is snowing, etc., but is not limited thereto.

In an embodiment, the data format of each estimated traffic flow learning 203 is: 0, 09: 50, 3, city road, citizen road, second section, 150, 40 (weather, time, number of lanes, road type) State, road name, road segment, number of mobile signaling data, traffic flow), the last field is the true value of traffic flow in the first 7 fields, but is not limited to this.

In this embodiment, the coordinate position of the action signaling data represents the location of the mobile phone user, and the base station analyzes the location, weather, and time of the mobile phone user only on the road segment according to the coordinate position of the mobile signaling data.

In this embodiment, the traffic flow in the step S501 is obtained by a Vehicle Detector (VD) and an eTag Detector on the road segment.

In an embodiment, the real-time data and historical data of the time data may be Classified as peak time or off-peak time.

In an embodiment, the road section type is a national expressway, an urban area, or a scenic section.

In an embodiment, the road segment type is classified according to the highway classification published by the General Administration of Highways, and can be classified into national highways, provincial highways, city roads, county roads, district roads, and rural roads according to the administrative system, and can be classified into highways according to transportation functions. Expressways, major roads, secondary roads and regional roads can be classified into highway systems, expressway systems, round-the-island road systems, transverse highway systems, long-distance highway systems, coastal road systems and liaison roads depending on the geographical environment.

In an embodiment, if the weather data, the time data, the action signaling data, the traffic volume, and the historical data of the road segment data obtained in the steps S501 and S502 are three months, the step S503 is performed. The weather data, the time data, the motion signaling data, the traffic flow, and the historical data of the road section data are subjected to neural network learning training to obtain the estimated traffic flow model.

In the previous embodiment, the neural network learning training is performed on the weather data, the time data, the motion signaling data, the traffic volume, and the historical data of the road segment data for the three months in the step S503. After obtaining the estimated traffic volume model, the weather data, the time data, the motion signaling data, the traffic volume, and the historical data of the road segment data are newly acquired for each step S501 and S502, and the step S503 is performed. The weather data, the time data, the action signaling data, the traffic flow and the historical data of the road section data of the past three months are subjected to neural network learning training to obtain the estimated traffic flow model.

It can be seen from the above that the present invention transmits traffic detected by a vehicle detector VD (vehicle detector) or an eTag Detector on certain sections of the historical data and the action on the road section. Number of signaling data, road type (national fast track, urban or scenic section, etc.), number of lanes, weather data and time data (spike or off-peak time) using a neural algorithm to find the corresponding push Estimating the traffic flow model, in the remaining VD-free or eTag detector sections, the traffic volume of the road section can be estimated through a similar estimated traffic flow model. The estimated traffic volume of the road section can be presented through the APP or the webpage. Save time by either posting traffic measures to avoid road congestion.

The above embodiments are intended to illustrate the principles of the invention and its effects, and are not intended to limit the invention. Any of the above-described embodiments may be modified by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention should be as set forth in the appended claims.

Claims (10)

A system for estimating road traffic using action signaling data includes: a data receiving module for receiving real-time weather data, real-time data, and real-time action signaling data, wherein the action signaling data is from a base Provided by the station, and the action signaling data is that when the mobile device is automatically connected to the base station in the power-on state, the base station obtains the action signaling data by receiving the signaling signal of the mobile device, and the signaling signal is The mobile device is connected to the base station, and the action signaling data includes a coordinate position; the signaling data segment corresponding module is used to obtain the real-time link data corresponding to the instant action signaling data; and the road segment The traffic flow estimation module estimates the traffic volume of the road segment based on the real-time weather data, the real-time time data, the instantaneous action signaling data and the real-time road segment data. The system of claim 1, wherein the data receiving module further receives the instantaneous traffic flow. The system of claim 2, further comprising: a database module for storing a plurality of the real-time weather data, the instant time data, the instant action signaling data, the instantaneous traffic flow and the instant The road segment data is defined as historical weather data, historical time data, historical action signaling data, historical traffic flow and historical road segment data; and historical data training module, which is in the database module. The historical weather data, the historical time data, the historical action signaling data, the historical traffic volume, and the historical road segment data are obtained by the neural network to estimate the traffic flow model. The system of claim 3, wherein the neural network is one of a recurrent neural network (RNN) or a deep neural network (DNN) or a combination thereof. The system of claim 1, wherein the road traffic estimation module further comprises continuously receiving the real-time weather data, the instant time data, the instant action signaling data, and the real-time road segment data. After the time, the traffic flow estimation module of the road section is based on the estimated traffic flow model for pushing the real-time weather data, the instant time data, the instant action signaling data and the real-time road segment data received during the period of time. Estimate to get the traffic volume of the road section. A method for estimating road traffic using action signaling data includes the following steps: (1) obtaining real-time data of weather data, time data, and action signaling data, wherein the action signaling data is provided by a base station. And the action signaling data is that when the mobile device is automatically connected to the base station in the power-on state, the base station obtains the action signaling data by receiving a signaling signal of the mobile device, and the signaling signal is the mobile device. a signal connected to the base station, and the action signaling data includes a coordinate location; (2) real-time data of the link data corresponding to the real-time data of the action signaling data; and (3) use of a EV The traffic model is based on the weather data and the time The inter-information data, the operational signaling information and the real-time data of the road segment data estimate the traffic volume of the road section. The method of claim 6, wherein the instant data in the step (3) is real-time data for a period of time, and the estimated vehicle flow model is used for the instant received during the period of time. The data is estimated to obtain the traffic volume of the road section. The method of claim 6, further comprising: transmitting the traffic volume of the road section to the display end to display the traffic volume of the road section. The method of claim 6, wherein the method for obtaining the estimated traffic flow model comprises the following steps: (3-1) obtaining historical data of weather data, time data, action signaling data, and traffic flow; 3-2) obtaining historical data of the link data corresponding to the historical data of the action signaling data; and (3-3) passing the historical data obtained by the steps (3-1) and (3-2) through the class The neural network obtains the estimated vehicle traffic model. The method of claim 9, wherein the neural network in the step (3-3) is one of a recurrent neural network (RNN) or a deep neural network (DNN). Or a combination thereof.