CN101510357B - Method for detecting traffic state based on mobile phone signal data - Google Patents

The invention provides a method for detecting traffic state based on mobile phone signals. A virtual sensor network is set up to acquire the real-time signal data sent from all mobile phones within a time interval from a mobile phone network at a fixed time interval, a virtual sensor section and the traveling speed vi can be obtained according to the real-time signal data sent from an i-th mobilephone; the virtual sensor section is superposed onto a road network to obtain the road network section and the traveling speed Vi of the i-th mobile phone; a real-time traffic flow intensity K and a section traffic flow Q can be obtained by collecting the number n of the mobile phones on each road network section and the traveling speed V of each road network section; and finally, through a traffic parameter prediction model, the predicted traffic flow speed V, the traffic flow intensity K and the section traffic flow Q can be obtained. The method has the advantages that the large range of real-time traffic data acquisition in the city can be completed in a short time by fully relying on the existing mobile communication network resources, and utilizing the information in the existing mobile phone communication network, meanwhile, the initial investment is relatively small, the data coverage is wide, and data accuracy is high.

A kind of method based on mobile phone signal Data Detection traffic behavior

Technical field

The present invention relates to a kind of method, be applicable to Urban Transportation management and traffic-information service industry, belong to the method and technology field of detecting traffic behavior with mobile phone signal based on mobile phone signal Data Detection traffic behavior.

Background technology

Intelligent transportation system (Intelligent Transportation System, be called for short ITS) be the exploit information communication technology, with people, car, road three's close coordination, harmony, on a large scale comprehensive play a role in real time, traffic management system accurately and efficiently.ITS can effectively utilize existing means of transportation and reduce traffic loading and environmental pollution, guarantee traffic safety, improve conevying efficiency, thereby promote socio-economic development, improve people's living standard, and generally be subjected to the attention of countries in the world with its ability that promotes social informatization and form new industry.

How to obtain dynamic real-time traffic information and become a ring important in the intelligent transportation system evolution.From technology trends, tradition fixed point acquisition technique:, can only gather the urban road traffic information in the limited range as inductive coil, radar, infrared and video; And Floating Car (normally taxi, bus or the freight etc.) technology of loading GPS equipment is limited by the vehicle number scale of charging appliance also, and the dynamic information of city subrange can only be provided.How to gather the interior Real-time Traffic Information in city on a large scale and become a technical barrier.The development of wireless communication networks, wireless communication networks signal collecting and the technology such as monitor supervision platform, safety encipher mechanism of the universal and mobile operator of mobile phone terminal and perfect, for utilizing mobile phone terminal as checkout equipment, obtain the wireless signal parameter in the normal use of mobile phone, the technology of gathering the wide area Real-time Traffic Information based on this realizes providing the important techniques guarantee.

In the past, traffic data all is to obtain with traditional fixation of sensor, such as, inductive coil, radar, infrared and video etc.Because high installation and maintenance expense, the traffic data collection technical development is to the Floating Car technology of the positional information of utilizing mobile phone or vehicle GPS.Floating Car based on GPS is meant the mobile device that the GPS unit is housed that operation vehicle on the way carries.The mobile device that the GPS unit is housed comprises GPS mobile phone, personal navigation equipment, the vehicle of GPS equipment etc. is housed.The signal that passes out from the GPS unit of GPS sample can be through handling and with the electronic map data coupling, thereby obtain travel locus, travelling speed and the hourage of GPS sample on different road sections.Floating Car based on mobile phone is meant mobile terminal of mobile telephone, and it may not comprise the GPS unit.Cell phone network generates various physical records constantly, and such as receiving short message, short-message sending, start, shutdown etc., the cellphone subscriber can connect just that to make a phone call and guarantee to converse be continuous through the border of cell phone network base station the time like this.Method and system in the past can not be handled the mobile phone signal data with the rudimentary knowledge of traffic science or traffic engineering, obtains in real time and the traffic behavior of prediction.

Summary of the invention

The purpose of this invention is to provide a kind of rudimentary knowledge and handle the mobile phone signal data, obtain in real time and the method based on mobile phone signal Data Detection traffic behavior of the traffic behavior of prediction with traffic science or traffic engineering.

In order to achieve the above object, technical scheme of the present invention has provided a kind of method based on mobile phone signal Data Detection traffic behavior, and step is:

Step 1, set up the virtual-sensor network;

Step 2, obtain the live signal data that all mobile phones send in this time interval from cell phone network with Fixed Time Interval T;

Base Station Identification in step 3, the live signal data sent according to i portion mobile phone number and Location Area Identification number obtain the virtual-sensor highway section at this mobile phone place and the travelling speed v on this virtual-sensor highway section by the geographical decoding of virtual-sensor network _i

Step 4, from the road network data storehouse, read the electronic map information of road network, the virtual-sensor highway section that step 3 is obtained is superimposed upon on the road network, obtain the road network highway section at i portion mobile phone place, make the travelling speed V of this mobile phone on the road network highway section _i=v _i

Step 6, calculate the traffic flow density K and the road section traffic volume flow Q in road corresponding network highway section in the time interval T according to the mobile phone number n on every road network highway section and travelling speed V.

Step 7, according to the time varying characteristic of different category of roads traffic flow parameters, set up corresponding traffic flow parameter forecast model.

The present invention is by gathering, analyze the communication data in the mobile communications network in real time, the mobile terminal of mobile telephone that domestic consumer is used is as a kind of effective traffic detecting device, the method of utilizing the present invention to propose, analyze movement locus and the movement velocity of calculating each mobile phone, obtain in real time and the road traffic state information of prediction.The present invention need not any specific installation to be installed on mobile phone terminal, to be need not to install any software, the regular handset that each personal user is used is as acquisition terminal, having broken through the conventional traffic acquisition technique needs the initial investment of installation acquisition terminal in advance to build bottleneck, can save a large amount of infrastructure investments.

The present invention can provide effective detection and monitoring means for Urban Transportation management, is applicable to vehicle supervision department of relevant government, for road infrastructure planning and operation maintenance, traffic control and management, traffic organization design provide information for supporting some decision.The present invention simultaneously also can be for the traffic-information service industry provides effective Real-time Traffic Information, for Real-time and Dynamic navigation, the transport information issue of various medium, fleet's management and running, special vehicle management and running provide source of traffic information.

Advantage of the present invention is: leverage fully on the existing mobile communication network resource, utilize the information in the existing mobile communication network, can finish real time traffic data collection on a large scale in the city in the short time, initial investment simultaneously is less relatively, data coverage is big, data precision is high.

Description of drawings

Fig. 1 is an overview flow chart of the present invention;

Fig. 2 is for setting up virtual-sensor flow through a network figure.

Embodiment

Specify the present invention below in conjunction with embodiment.

Embodiment

A kind of method provided by the invention based on mobile phone signal Data Detection traffic behavior, step is:

Step 1, set up the virtual-sensor network;

Step 1.1, obtain the position of virtual-sensor node and the sequence of cell base station by field test, wherein the virtual-sensor node definition in the intersection region of wireless network and road network because the cell phone network signal changes the point that produces the wireless network incident:

Step 1.1.1, implement the radio network route measuring operation by test carriage, collect the particular location of test carriage when on the handover information of specifying the Um Interface on the highway section and location updating message and the highway section these information taking place on the spot, specify the highway section to choose different scopes according to the needs of different Real-time Traffic Informations, generally, so long as the public way of permission vehicle ' can;

The definition in highway section is the general knowledge of traffic aspect, for the surface road that traffic lights control is arranged, the definition in highway section be from the crossing of traffic lights control to the crossing of another adjacent traffic lights control, the length in this type of highway section is usually at 200 meters to 600 meters; For the expressway or through street that do not have traffic lights control, choosing of highway section is to an other particular point from a particular point, this type of particular point is up and down ring road and expressway or through street intersection point, road turn round point, number of track-lines catastrophe point etc., and the length in this type of highway section is usually at 400 meters to 800 meters;

Step 1.1.2, will repeatedly measure the same handover information of gained or the pairing position data of location updating message to the drive test information that obtains by step 1.1.1 along road direction and average and obtain the position of all virtual-sensor nodes;

Step 1.1.3, owing to the virtual-sensor node definition in the intersection region of wireless network and road network because the cell phone network signal changes the point produce the wireless network incident, therefore the virtual-sensor nodal information comprises the cell information that switch the sub-district or the position renewal front and back pass in and out, and the sequence that promptly obtains cell base station is enumerated out in the turnover sub-district, front and back that all virtual-sensor nodes are represented separately;

The position of step 1.2, the virtual-sensor node that obtains according to step 1.1, the sequence of pressing the cell base station that drive test gathers is respectively with starting point and the terminal point of latter two virtual-sensor node of adjacent elder generation as the virtual-sensor highway section, set up the corresponding relation of virtual-sensor highway section and virtual-sensor node, obtain the travel direction in virtual-sensor highway section simultaneously;

Step 1.3, obtain the length in each virtual-sensor highway section and virtual-sensor highway section and road network are mated mutually, determine the corresponding relation of itself and road network according to the space overlap in virtual-sensor highway section and the road network highway section relation and the consistance of travel direction, wherein, the road network highway section carries directional information;

Step 1.4, all virtual-sensor node and virtual-sensor highway sections have constituted the virtual-sensor network;

Step 2, obtain the live signal data that all mobile phones send in this time interval from cell phone network with Fixed Time Interval T;

Common time interval adopts 2 minutes or 5 minutes as the time interval.Suppose that system's zero-time is 8:00AM in the morning, if the time interval is 2 minutes, then, first subsequent computation period for the morning 8:00:00 to 8:01:59, be 8:02:00-8:03:59 then successively ... 9:42:00-9:43:59, If the time interval is 5 minutes, then, first subsequent computation period for the morning 8:00:00 to 8:04:59, be 8:05:00-8:09:59 then successively ... 9:45:00-9:49:59 ...The mobile phone quantity that relies on time range to come constrained sampling;

The geographical decoding of virtual-sensor network that Base Station Identification in the live signal data that step 3, the i portion mobile phone that obtains according to step 2 send number and Location Area Identification number are set up by step 1 obtains the virtual-sensor highway section and the travelling speed v of this mobile phone on the virtual-sensor highway section at this mobile phone place _i:

Two virtual-sensor nodes that Base Station Identification in step 3.1, the live signal data sent according to i portion mobile phone number and Location Area Identification number obtain being complementary and through the timestamp of these two virtual-sensor nodes are t through the timestamp of first virtual-sensor node ₁, be t through the timestamp of second virtual-sensor node ₂

Step 3.2, the corresponding relation by virtual-sensor highway section and virtual-sensor node are matched to each virtual-sensor highway section respectively with the i portion mobile phone that obtains in the step 3.1 successively virtual-sensor node of process, obtain the virtual-sensor highway section at this mobile phone place, and calculate the length d in virtual-sensor highway section by the geometric formula of routine;

Step 3.3, the travelling speed v of i portion mobile phone on the virtual-sensor highway section _i=d/ (t ₂-t ₁);

Step 4, from the road network data storehouse, read the electronic map information of road network, the virtual-sensor highway section that step 3 is obtained is superimposed upon on the road network, obtain the road network highway section at i portion mobile phone place, and the travelling speed V of this mobile phone on the road network highway section _i=v _i

Owing to when setting up the virtual-sensor network, established the corresponding relation in virtual-sensor highway section and road network highway section, therefore can finish stack according to the relation of above-mentioned establishment;

V=e * V (up)+f * V (down), the degree of correlation factor that e, f obtain for the training according to historical data, and e+f=1, V (up) is the travelling speed with this adjacent upstream highway section, highway section, V (down) is the travelling speed with this adjacent downstream highway section, highway section, upstream and downstream is a relative notion, and the upstream and downstream relation is to be decided by the actual location distribution of road and travel direction.Each computation period step 5 is finished is calculating to all road network highway section travelling speeds, so that the speed of upstream and downstream also is is known.

Mobile phone number n on step 6, every the road network highway section obtaining according to step 5 and travelling speed V calculate traffic flow density K and the road section traffic volume flow Q in the time interval T:

(B/N), wherein, K is for needing the road section traffic volume current density of estimation, and N is the mobile phone number n that obtains by step 5, and A and B are the system model parameter by historical data training gained for step 6.1, traffic flow density K=A * B * N * EXP;

Step 6.2, road section traffic volume flow Q=K * V, wherein, the traffic flow density of K for obtaining, the travelling speed of V for obtaining by step 5 by step 6.1;

Step 7, according to the time varying characteristic of different category of roads traffic flow parameters, set up corresponding traffic flow parameter forecast model:

Step 7.1, one day is divided into 24/T time period, sets up highway section travelling speed linear prediction model according to the described time interval T of step 2:

V(k+1)＝a×V(k)+b×V(k-1)+c×V(k-2)+d×V(k-3)，

Wherein, a, b, c, d is the degree of correlation factor, training obtains according to historical data, and a+b+c+d=1, a＞=b＞=c＞=d, k+1 is a time period numbering to be predicted, V (k+1) is the travelling speed of time period k+1 to be predicted, V (k) is the travelling speed of the previous time period of the time period k+1 to be predicted that obtains by step 5, V (k-1) is the travelling speed of the first two time period of the time period k+1 to be predicted that obtains by step 5, V (k-2) is the travelling speed of first three time period of the time period k+1 to be predicted that obtains by step 5, and V (k-3) is the travelling speed of preceding four time periods of the time period k+1 to be predicted that obtains by step 5;

Step 7.2, set up road traffic delay density linear prediction model:

K(k+1)＝a×K(k)+b×K(k-1)+c×K(k-2)+d×K(k-3)，

Wherein, a, b, c, d are the degree of correlation factor, and training obtains according to historical data, and a+b+c+d=1, and a＞=b＞=c＞=d, k+1 are time period numbering to be predicted; K (k+1) is the traffic flow density of time period k+1 to be predicted; K (k) is the traffic flow density of the previous time period of the time period k+1 to be predicted that obtains by step 6;

K (k-1) is the traffic flow density of the first two time period of the time period k+1 to be predicted that obtains by step 6;

K (k-2) is the traffic flow density of first three time period of the time period k+1 to be predicted that obtains by step 6;

K (k-3) is the traffic flow density of preceding four time periods of the time period k+1 to be predicted that obtains by step 6.

Step 7.3, set up road traffic delay volume forecasting model:

Q(k+1)＝V(k+1)×K(k+1)

Wherein, Q (k+1) is the traffic flow flow of time period k+1 to be predicted, and V (k+1) is the traffic speed of the time period k+1 to be predicted of step 7.1 gained, and V (k+1) is the traffic speed of the time period k+1 to be predicted of step 7.2 gained.

For example at a certain highway section RL1 (length of RL1 is 800m) corresponding three virtual-sensor highway section VL1, VL2 and VL3, its corresponding relation is as shown in the table:

Being located at has four mobile phones on this highway section, then mobile phone signal is matched and obtain following table behind the virtual-sensor node:

According to the corresponding relation in the table 1, can calculate the hourage and the travelling speed of the mobile phone sample of each virtual-sensor highway section correspondence, as shown in table 2;

Table 1: virtual-sensor highway section and virtual-sensor node corresponding tables

Table 2: each mobile phone sample hourage and travelling speed result of calculation, and match each virtual-sensor highway section

As seen from the above table, these four mobile phone samples are the sample of highway section RL1, then:

The estimation travelling speed of highway section RL1 is: (36+34.29+32.14+34.11)/4=34.13 (kilometer/hour);

The estimation of highway section RL1 is hourage: 3.6 * 800/34.13=84.4 (second);

Calculate after highway section travelling speed and the effective mobile phone sample size,, can further obtain other traffic flow parameter according to traffic flow model.

As follows as the funtcional relationship that can set up road section traffic volume current density and effective sample quantity: K=A * B * N * EXP (B/N), wherein, K---needs the road section traffic volume current density of estimation, N---and refers to the mobile phone sample size, A, B---system model parameter is by historical data training gained, A=0.2 in this example, B=5, N=4, then, the highway section estimates that traffic flow density is: K=0.2 * 5 * 4 * EXP (1/4)=12.5 (car/every kilometer per car road).

Again according to the relation between traffic flow three parameter flows, density and the speed: Q=K * V, then, the highway section estimates that the magnitude of traffic flow is: Q=12.5 * 34.20=426 (car/per hour per car road), predict that the travelling speed of next time period is: 0.4 * 34.2+0.3 * 36.4+0.2 * 37.1+0.1 * 39.9=36.01 (kilometer/hour).