CN100492434C - A method for obtaining the sampling amount of the probe car required for traffic flow state analysis - Google Patents

Abstract

This invention relates to an obtaining method for sampling volume of detection cars needed by analyzing traffic flow state including: 1, pre-processing data of a GPS detection car, 2, matching the car with a GIS map, 3, obtaining mean velocity of highway sections, 4, obtaining sampling volume of a road net detection car, which traces vehicles to eliminated interference data, applies a nearest method to match maps then uses a space time curvature fitting model to get a mean velocity of directional sections, calculates historical data of the GPS detection car, its mean number of weight and empty in each time section network of a day, applies a sampling formula to compute the sample frequency of each time section and get the sample volume of the detection car of the road network finally.

Description

A method for obtaining the sampling amount of the probe car required for traffic flow state analysis

technical field

The invention relates to an information processing method in the technical field of transportation, in particular to a method for acquiring the sampling amount of a probe car required for traffic flow state analysis.

Background technique

With the development of intelligent transportation technology, the dynamic estimation of urban traffic flow has received more and more attention. The traffic state estimation is inseparable from accurate and reliable traffic information, so the accuracy of the detection information method determines the accuracy of the traffic state estimation. There are many ways to detect traffic flow information in urban road network, mainly including magnetic frequency induction coil detection method, video detection method and so on. These studies have achieved success in some aspects and have practical value, but there are limitations in the calculation of urban road network traffic flow velocity. The induction coil monitor can obtain a variety of traffic flow parameters, but it is difficult to lay the induction coil, the cost of adding equipment is high, and the service life is greatly affected by human-destructive factors. Video detection has high requirements on hardware equipment and is greatly affected by weather. Applying the information detection method of the Global Satellite Positioning System (GPS (Global Satellite Positioning System)), through real-time monitoring of moving vehicles, dynamic acquisition of time, position, speed and other vehicle positioning information, with high precision, large amount of data, within the city limits Wide distribution, less affected by weather and other advantages, is currently the most effective way to obtain real-time traffic information. Therefore, more and more scholars devote themselves to the research of traffic flow state analysis using GPS (Global Satellite Positioning System) probe vehicle data, and have obtained rapid development. However, how many samples of probe cars are needed to accurately estimate the traffic state is the primary problem to be solved at present.

In recent years, many scholars in the world have tried to use GPS (Global Satellite Positioning System) probe car information to calculate the average speed and travel time of road sections. A large number of studies have been carried out on the problem of how many samples of probe cars (that is, the sampling volume of probe cars) are needed. Research. Among them, Quiroga and Bullock published the article "Determination of sample sizes for travel time studies" in the "Institute of Transportation Engineers" journal, Volume 68, Issue 8, Page 92-98 in 1998, and proposed to obtain the sampling amount of the probe car. standard variance model. The model obtains the minimum number of probe vehicles by introducing a t-distribution with a confidence interval of 1-α and a degree of freedom of n-1, the standard deviation s of the sample velocity and a custom velocity error ε _α . Applying this model can effectively obtain the sampling amount of the probe car. In 2000, Chen and Chien published "Determining the number of probe vehicles for freeway travel time estimation using microscopic simulation" at the 79th Annual Meeting of the "Transportation Research Board". The standard variance model was improved and the relative speed error was introduced. ε _γ , and the average speed x of the road section calculated by n probe cars to obtain more accurate probe car sampling. These two models have been widely used in sampling volume estimation in recent years because they can effectively obtain the sampling volume of the probe vehicle. However, both models assume that the vehicles in the road section obey the normal distribution, which is impossible in actual road traffic; in addition, because the standard deviation s of the sample speed needs to be calculated by n, the model equation has no closed solution, and must Iterative solution, which also increases the complexity of the problem in practical applications.

After searching the prior art documents, it was found that Singaporean scholars Cheu and Lee published "Vehicle Population and Sample Size for Arterial Speed Estimation" ( In this paper, a model for obtaining the sampling volume of probe vehicles is proposed. This model not only retains the advantages of the standard variance model, but the most important thing is to overcome the shortcomings of the standard variance model that need to be solved iteratively. At the same time, Cheu proposed to ensure that the absolute error of the calculated average speed of the road network is lower than 95% (or less than 5.0km/h), and the number of GPS (Global Satellite Positioning System) detection vehicles on each road section must be greater than 10. , which greatly facilitates the actual road traffic management. However, Cheu did not take into account the difference in the sending cycle of GPS (Global Satellite Positioning System) empty-loaded vehicle information, which seriously affects the accuracy of the calculation results. In addition, because no detailed analytical formula was proposed, the sampling volume of the probe vehicle could not be obtained quickly and accurately, which greatly restricted the promotion of the method.

Contents of the invention

The object of the present invention is to propose a method for obtaining the sampling volume of probe cars required for traffic flow state analysis in view of the above-mentioned deficiencies and actual needs. The present invention effectively overcomes the problems of the prior art that require a large amount of prior and complex information, has the advantages of simple operation, high reliability, and strong practicability, and provides information for the control of the entire urban traffic.

The present invention is achieved through the following technical solutions, and the present invention comprises the following steps:

① Preprocessing the GPS (Global Satellite Positioning System) rover data

Currently available GPS (Global Satellite Positioning System) probe vehicle data mainly includes vehicle location, time, status, vehicle speed, driving direction and other attributes. The preprocessing is mainly aimed at the GPS (Global Satellite Positioning System) detection vehicle whose speed is 0 on the road section. These probe cars include two parts: the probe car whose speed is 0 due to severe congestion, and the probe car whose speed is 0 because the vehicle has been parked somewhere for a long time without driving.

The former is the key part of traffic flow analysis on road sections, and it is correct data; the latter is interference data, so it is necessary to judge and process this part of data accordingly. The discrimination method adopted by the present invention is to track the vehicle. If it is found that the speed is always 0 for a long period of time exceeding one signal cycle, it indicates that the data is interference data, and the data is eliminated during preprocessing.

②Match the GPS (Global Satellite Positioning System) rover with the GIS (Geographic Information System) map

Obtain the road network GIS (Geographic Information System) information provided by the geographic information system and the preprocessed GPS (Global Satellite Positioning System) probe vehicle information, and treat the vehicle GPS (Global Satellite Positioning System) position data as scattered data to the surrounding The road is vertically projected and the projection distance is calculated. If the shortest projection distance of a scatter point data is greater than the preset threshold, it is considered as a wrong matching point and filtered out. Otherwise, the road corresponding to the shortest projection distance is taken as the vehicle. The corresponding projection point is the position of the vehicle after matching, and the preliminary result is obtained, and the map matching from point to line is completed.

Third, obtain the average speed of the road section

Intersections in urban road networks isolate roads into two directional sections, uplink and downlink. Taking the GPS data sampling points on the directional road section of the unit in the time period T as the object, the distance, time, and speed three-dimensional surface fitting modeling are carried out on it, and the spatial-temporal distribution of the directional road section of the unit in the time period T is obtained. Velocity distribution surface. The specific method is to determine the order of the surface fitting model according to the number of effective data points, and then determine the minimum value M of GPS data points required by the surface fitting model. Then the speed distribution surface is integrated in the direction of the road to obtain the average speed v _i (t) of the unit directed road section in this period.

Fourth, the acquisition of the sampling volume of the road network detection vehicle

The sampling amount n of the probe car is the number of GPS probe cars needed to accurately calculate the average speed of each road section in the road network. It depends on factors such as vehicle speed, road grade, road section length, and probe vehicle sampling frequency f(t). Sampling frequency f(t) refers to the number of GPS data provided by the probe car driving through the road section within a sampling period, which depends on the average speed of the probe car on the road section.

Since GP detection vehicles mainly include empty vehicles that do not carry passengers and heavy vehicles that carry passengers, and there are great differences in the cycle of sending information between empty and heavy vehicles. Let L _i be the length of road segment i (i=1, 2....N), T _L and T _H are the time for empty vehicle and heavy vehicle to send data respectively, n _L (t) and n _H (t) are The number of empty vehicles and heavy vehicles in the sampling period (need to be obtained through statistical historical GPS (Global Satellite Positioning System) data), v _i (t) is the average speed of the road section in the sampling period. The formula for the sampling frequency of the probe car on road section i is:

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; L</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; *</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; L</span>}}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; *</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}}}{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; L</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}$

The required probe vehicle sampling amount model of the road network proposed by the present invention is:

$\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; \&Greater\; Equal;</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}_{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; )</span>}$

e is the positioning error of the GPS (Global Satellite Positioning System) probe vehicle, ε _γ is the confidence degree of the average speed of the road section, and _Mi is the minimum value of the GPS (Global Satellite Positioning System) data points required for the average speed of the road section to fit the model.

The remarkable effect of the present invention compared with the prior art is that the accuracy of the positioning data of the mobile vehicle (probing vehicle) acquired by the present invention by using the GPS (Global Positioning System) information detection method is as high as 95%, and the data volume is large, which can cover roads. All vehicles in the network. The space-time curved surface fitting model applied in the present invention has fast operation speed (the operation time is 15-25 seconds), and can realize real-time synchronous acquisition of the average speed of road sections. Therefore, the present invention can quickly and accurately acquire the sampling amount of the probe car required for traffic flow state analysis, and provide real-time and accurate information for the traffic control of the whole city. The accuracy rate can reach about 90% by applying the probe car sample acquired by the invention to analyze the urban traffic state.

Description of drawings

Fig. 1 is a block flow diagram of the method for calculating the sampling amount of the probe car required by the urban road network proposed by the present invention.

Figure 2 shows the 24-hour average speed change of a road section.

Figure 3 shows the minimum sampling volume of probe vehicles required for traffic flow state analysis of the inner ring road network in Shanghai. 3(a) is the minimum amount of probe vehicle sampling required for the main road network, 3(b) is the minimum probe vehicle sampling amount required for the secondary road network, and 3(c) is the minimum probe vehicle sampling required for the branch road network 3(d) is the minimum amount of probe car sampling required for the entire inner loop network.

Detailed ways

The embodiments of the present invention are described in detail below in conjunction with the accompanying drawings: the present embodiment is implemented on the premise of the technical solution of the present invention, and detailed implementation methods and processes are provided, but the protection scope of the present invention is not limited to the following implementations example.

In order to better understand the method proposed in this embodiment, an example of traffic flow state analysis is selected from the inner ring road network in Shanghai, which can be applied to different road networks in different cities. The present embodiment requires providing GPS (Global Satellite Positioning System) system vehicle satellite positioning data within one month in the inner circle of Shanghai, including sampling vehicle label, time, position, speed, running direction, vehicle status and other information. There is also GIS (Geographic Information System) information of the inner ring road network in Shanghai.

As shown in Figure 1, the specific implementation steps of this embodiment are as follows:

① Preprocessing the GPS (Global Satellite Positioning System) probe vehicle data:

The acquired GPS (Global Satellite Positioning System) data is stored with the ID number of the vehicle. Firstly, the vehicle is tracked according to the ID number of the detection vehicle. If all the speed values of the detection vehicle are 0, it is concluded that the vehicle is an abnormal vehicle, and the file is deleted; If it is always 0, it is determined that the vehicle provides invalid data during this time period, and all information of the vehicle within this time period is deleted. After abnormal data processing, the GPS (Global Satellite Positioning System) data is stored in chronological order, and the GPS (Global Satellite Positioning System) data of all vehicles within 8 minutes are stored in a file. Facilitate dynamic traffic flow analysis.

②Match the GPS (Global Satellite Positioning System) rover with the GIS (Geographic Information System) map:

Since the position of the probe vehicle in the GPS (Global Satellite Positioning System) data is represented by latitude and longitude coordinates, coordinate transformation should be performed before map matching, and transformed into urban construction coordinates in GIS (Geographic Information System). Then the 8-minute taxi GPS (Global Positioning System) data is read in at the same time, and the position data is regarded as scattered data, and the nearest neighbor method is used for unified map matching, and the GPS (Global Satellite Positioning System) positioning data of the vehicle is Corrected to the road the vehicle is traveling on. Finally, the vehicle trajectory is tracked according to the vehicle's running direction, and then the uncertain positioning data is discriminated according to the vehicle data's front-to-back relationship and road conditions, and the corrected vehicle trajectory is processed to obtain the vehicle's trajectory, and finally the map matching of the vehicle to the road is completed.

③ Obtain the average speed of the road section:

The GPS (Global Satellite Positioning System) data of each probe vehicle is regarded as the traffic sampling points (l _i , t _i , v _i ) on the road section. Taking 8 minutes as a time period (the whole day is divided into 180 time periods), the space-time surface fitting is performed on all GPS (Global Satellite Positioning System) sampling point data on the directional road section of the unit to obtain the surface fitting and the speed distribution curve . Then it is integrated in the time and space directions respectively to obtain the average speed v of road section k within 8 minutes (480 seconds).

${\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; \%</span>}<span\; class="notranslate">\; =</span>\frac{{<span\; class="notranslate">\; \&Integral;</span>}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; \&beta;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}{\mathrm{<span\; class="notranslate">\; l</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}^{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; dl</span>}}{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}}$

$\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 480</span>}}{<span\; class="notranslate">\; \&Integral;</span>}}{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; \%</span>}\mathrm{<span\; class="notranslate">\; lt</span>}}{\mathrm{<span\; class="notranslate">\; 480</span>}}$

Among them, l is the distance from the probe car to the starting point of the road section, and v is the instantaneous average speed of the road section at time t ₀ . a _ij is the coefficient of the fitting polynomial, which is determined by the GPS (Global Positioning System) data information in the sampling period. α, β are the highest times of l and t during fitting, which are determined by the number of valid GPS (Global Positioning System) data points on the fitting road section. The specific values are shown in Table 1:

GPS (Global Positioning System) data information [9, 12] [6, 8] [3, 5] (α,β) (2,3) (2, 2) (1,1) m 9 6 3

Table 1

GPS (Global Satellite Positioning System) data is the number of effective GPS (Global Satellite Positioning System) data points on the road section, which determines the times of α and β in the fitting model. M is the minimum number of GPS (Global Positioning System) data points required to fit the model. Considering the limitation of data and actual road conditions, MAX(α)=2 and MAX(β)=3 are selected in this example. Figure 2 shows the 24-hour average speed of a road section calculated by dynamically selecting the fitting model according to the method in Table 1.

④ Acquisition of the sampling volume of the road network detection vehicle:

Statistics the historical data of GPS (Global Satellite Positioning System) detection vehicles, and calculate the average number of heavy vehicles and empty vehicles in the road network at various times of the day. The period of the feedback information of the heavy vehicle is selected as 2 minutes, and the period of the feedback information of the empty vehicle is 20 seconds. The sampling frequency f(t) of each period is calculated according to the sampling frequency formula. Since the amount of probed vehicle data in the road network varies greatly with time, in this example, the fitting model is adaptively selected according to the actual number of probed vehicles in the road network to calculate the sampling amount. Table 2 lists the method of adaptively selecting the order of the model and the minimum sampling amount of GPS (Global Satellite Positioning System) probe vehicles required for the analysis of the traffic flow state of the inner ring of Shanghai in different periods. See Table 2 for the method and calculation results of adaptively selecting the traffic flow analysis model:

period of time 0:00—5:00 5:00—8:00 8:00—20:00 20:00—24:00 Fitting method one surface fitting Quadratic Surface Fitting Cubic Surface Fitting Quadratic Surface Fitting (α,β) (1,1) (2,3) (3,3) (1,1) Minimum probe car sampling volume (vehicle) 1500 2500 3000 2000

Table 2

Figure 3 shows the minimum sampling amount of probe vehicles required for the analysis of the traffic flow state of the inner ring road network in Shanghai calculated by using the adaptive surface fitting model and the cubic surface fitting model in Table 2 respectively. where a is the minimum amount of probe vehicle sampling required for the main road network, b is the minimum probe vehicle sampling amount required for the secondary arterial network, c is the minimum probe vehicle sampling amount required for the branch road network, and d is the entire inner loop The minimum amount of rover samples required by the network. Among them, there are about 3,000 vehicles during the day, which is in good agreement with the real road conditions of Shanghai's inner ring.

It can be seen from this embodiment that the GPS (Global Positioning System) probe vehicle sample obtained according to this embodiment can be effectively used for traffic flow state analysis of the urban road network.

Claims (3)

1, a kind of acquisition methods of traffic flow state analysis required detection vehicle sampling quantity is characterized in that, may further comprise the steps:

1. GPS probe vehicles data are carried out pre-service, vehicle is followed the tracks of,, conclude that then this car is unusual vehicle, deletion this document if all velocity amplitudes of GPS probe vehicles all are 0; If the speed of finding the GPS probe vehicles is always at 0 o'clock in surpassing long-time section an of signal period, show that these data are interfering datas, when carrying out pre-service, these data are rejected;

2. GPS probe vehicles and GIS map are mated;

3. obtain road-section average speed;

4. the road network detection vehicle sampling quantity obtains;

Described detection vehicle sampling quantity, its value n, be accurately to calculate the needed GPS probe vehicles of each road-section average speed quantity in the road network, it depends on car speed, category of roads, road section length and detection vehicle sampling frequency f (t), sample frequency f (t) was meant in the sampling period, probe vehicles crosses the number of the gps data that this highway section provides, and depends on the average velocity of probe vehicles in the highway section;

Described GPS probe vehicles comprises the empty wagons of not carrying and the loaded vehicle of carrying, makes L _iBe the length of highway section i, i=1,2....N, T _LAnd T _HBe respectively the time of empty wagons and loaded vehicle transmission data, n _L(t) and n _H(t) be the quantity of interior empty wagons of sampling period and loaded vehicle, need obtain v by the statistical history gps data _i(t) be the road-section average speed in sampling period, the detection vehicle sampling frequency formula of highway section i is:

$f_i\left(t\right)=\frac{n_L\left(t\right)*\frac{L_i}{v_i\left(t\right)*T_L}+n_H\left(t\right)*\frac{L_i}{v_i\left(t\right)*T_H}}{n_L\left(t\right)+n_H\left(t\right)};$

The required detection vehicle sampling quantity model of road network is:

$n\&GreaterEqual;\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}\frac{M_i}{f_i\left(t\right){\mathrm{\&epsiv;}}_{\mathrm{\&gamma;}}(1-e)}$

E is a GPS probe vehicles positioning error, ε _γBe the degree of confidence of road-section average speed, M _iMinimum value for the needed gps data point of road-section average speed model of fit.

2, the acquisition methods of traffic flow state analysis required detection vehicle sampling quantity according to claim 1, it is characterized in that, described GPS probe vehicles and GIS map are mated, be meant: obtain road network GIS information and pretreated GPS probe vehicles information that Geographic Information System provides, the position data of vehicle GPS is considered as diffusing point data road vertical projection towards periphery, and calculating projector distance, if wherein the shortest projector distance of certain diffusing point data is greater than the threshold value that sets in advance, then think error matching points, it is filtered out, otherwise getting the pairing road of its shortest projector distance is the travel at vehicle place, corresponding subpoint is the position after the vehicle coupling, obtain PRELIMINARY RESULTS, finish from putting the map match of line.

3, the acquisition methods of traffic flow state analysis required detection vehicle sampling quantity according to claim 1, it is characterized in that, the described road-section average speed of obtaining, be meant: the crossing is isolated into two oriented highway sections of uplink and downlink with road in the city road network, with the gps data sampled point that is in the time period T on the oriented highway section of unit is object, it is carried out distance, time, the three-dimensional surface fitting modeling of speed, obtain this unit velocity distribution curved surface of oriented highway section on space-time in the T time period; Determine the exponent number of surface fitting model then according to the number of active data point, and then the minimum M of the needed gps data point of definite surface fitting model, then to the velocity distribution curved surface in the road direction upper integral, obtain the average velocity v in the oriented highway section of this period unit _i(t).

Images