WO2018064931A1 - 考虑出租车运营状态下的城市道路行程时间分布估算方法 - Google Patents

考虑出租车运营状态下的城市道路行程时间分布估算方法 Download PDF

Info

Publication number
WO2018064931A1
WO2018064931A1 PCT/CN2017/102693 CN2017102693W WO2018064931A1 WO 2018064931 A1 WO2018064931 A1 WO 2018064931A1 CN 2017102693 W CN2017102693 W CN 2017102693W WO 2018064931 A1 WO2018064931 A1 WO 2018064931A1
Authority
WO
WIPO (PCT)
Prior art keywords
link
travel time
taxi
state
path
Prior art date
Application number
PCT/CN2017/102693
Other languages
English (en)
French (fr)
Inventor
钟绍鹏
隽海民
刘锴
王全志
邹延权
王坤
张路
唐天力
Original Assignee
大连理工大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 大连理工大学 filed Critical 大连理工大学
Priority to US15/778,064 priority Critical patent/US10748421B2/en
Publication of WO2018064931A1 publication Critical patent/WO2018064931A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/123Traffic control systems for road vehicles indicating the position of vehicles, e.g. scheduled vehicles; Managing passenger vehicles circulating according to a fixed timetable, e.g. buses, trains, trams
    • G08G1/127Traffic control systems for road vehicles indicating the position of vehicles, e.g. scheduled vehicles; Managing passenger vehicles circulating according to a fixed timetable, e.g. buses, trains, trams to a central station ; Indicators in a central station
    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/01Detecting movement of traffic to be counted or controlled
    • G08G1/0104Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0108Measuring and analyzing of parameters relative to traffic conditions based on the source of data
    • G08G1/0112Measuring and analyzing of parameters relative to traffic conditions based on the source of data from the vehicle, e.g. floating car data [FCD]
    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/20Monitoring the location of vehicles belonging to a group, e.g. fleet of vehicles, countable or determined number of vehicles
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/04Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047Optimisation of routes or paths, e.g. travelling salesman problem
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063Operations research, analysis or management
    • G06Q10/0631Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40Business processes related to the transportation industry
    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/01Detecting movement of traffic to be counted or controlled
    • G08G1/0104Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125Traffic data processing
    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/01Detecting movement of traffic to be counted or controlled
    • G08G1/0104Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125Traffic data processing
    • G08G1/0129Traffic data processing for creating historical data or processing based on historical data

Definitions

  • the invention belongs to the technical field of urban traffic planning and management, and relates to the estimation of urban road travel time and the field of ITS intelligent transportation system, and is particularly suitable for estimating the travel time distribution of urban routes based on taxi data.
  • Chen used the floating car data in "Dynamic Freeway Travel Time Prediction Using ProbeVehicle Data: Link-based vs. Path-based" to compare the travel time estimation methods based on road segments and paths, and then discussed the proportion of floating cars. Estimating the influence of accuracy, the Kalman filter is used to estimate the travel time of the road segment based on the floating car data, and the simulation is verified.
  • the present invention proposes a more accurate path travel time estimation method based on taxi data, and fully considers the influence of taxi operation status on path travel time estimation, and proposes a more accurate improvement method.
  • the technical problem to be solved by the invention is to first obtain the travel time distribution of each road section by using the taxi GPS data of each road section, and then construct a model to estimate the route travel time distribution of the taxi in a certain operating state, and finally operate according to two different operations.
  • the number of vehicles is proportional, and the weight is set to obtain the final path travel time distribution.
  • a method for estimating the travel time distribution of urban roads under the condition of taxi operation the steps are as follows:
  • the collected taxi GPS data is filtered, corrected and matched, and the taxi GPS data of each road section containing the license plate number, longitude, latitude, speed and passenger status field is obtained, which is recorded as table a. ;
  • i road unit distance travel time also known as travel time rate, unit s / m;
  • v i the average speed of a vehicle in the i section
  • the new taxi data table containing the license plate number, longitude, latitude, travel time rate and passenger status field is recorded as table b.
  • the table is classified according to the empty car and the passengers (0 and 1), and the table c with the passenger state of 0 and the table d with the passenger state 1 are obtained.
  • the travel time rate data in table c and table d are clustered to obtain multiple driving state data in two operating states.
  • the taxi travel time rate data in the same driving state is fitted, and the probability density distribution function of the travel time rate of each link in each running state is obtained.
  • the travel time rate of the vehicle on the current road segment is determined only by the upstream road section, and is independent of the travel time rate of the previous road section. Therefore, the traffic space change on the current road section is similar to the structure of the Markov chain.
  • the time rate has a typical Markov property, and the Markov chain can be used to model the correlation between the travel time rates of the segments.
  • the central distribution range of each taxi travel time rate defining the current road segment is the state of the Markov chain. It is the boundary value set of link l, and m l is the number of link l states. For the boundary value set of link l+1, n l is the number of link l+1 states; therefore, the first state of link l is represented as The last state is represented as Where ⁇ l represents the taxi travel time rate on the link l;
  • the probability distribution of each driving state defining the link 1 of the road segment is the initial state probability distribution of the Markov chain:
  • N(i) represents the state i of the link 1 of the link, for example, the travel time rate of the link 1 is at The number of data points within.
  • the general state probability transfer matrix P is expressed as:
  • S l represents the state of the link l of the link
  • N(i, j) represents the number of data points that are in the i state on the link l and in the j state on the link l+1.
  • the intermediate link segment l and its downstream link segment l-1 can match the vehicle and the intermediate link between the link l and the downstream link l l+1.
  • the vehicle may be different for the intermediate segment.
  • the driving state classification is completely the same.
  • the vehicle passes the path Any of a variety of states, each of which is referred to as a Markov path.
  • the product of the transition probabilities between states of all segments is the probability of occurrence of the Markov path:
  • the convolution operation can be used to obtain the path travel time rate distribution of a new Markov path:
  • the path travel time rate distribution in the two operating states is calculated by the method in the step (2), and then the weights of the respective distribution functions are set according to the data amounts in the two operating states.
  • the formula for calculating the total path travel time rate distribution is as follows:
  • the ratio of vehicles to the total number of taxis is the ratio of empty to car and passenger to passenger.
  • the invention proposes to calculate the final path travel time distribution according to the proportion of the number of vehicles in two different operating states, and calculate the proposed model by using the data of Shenzhen City, and finds the calculated travel after considering the operating state.
  • the time distribution function is more precise.
  • FIG. 1 is a method for superimposing the travel time rate of each group of vehicles.
  • Figure 2 is a schematic diagram of the study area.
  • Fig. 3(a) is a travel time rate probability density function image of link1 in the "fast” state in the empty state.
  • Fig. 3(b) is a travel time rate probability density function image of link3 in the "slow” state in the empty vehicle state.
  • Figure 4 is a probability density distribution curve of the path travel time rate in the empty car and passenger state.
  • Fig. 5 is a cumulative probability distribution curve of the path travel time rate in the empty vehicle and passenger state.
  • Fig. 6 is a schematic diagram (partial) of comparing the actual value of the travel time rate distribution of the road section with the simulation value.
  • FIG. 7 is a comparison diagram of whether the operational state estimated travel time rate probability density distribution and the real travel time rate probability density distribution are distinguished.
  • FIG. 8 is a comparison diagram of whether or not the operational state estimation travel time rate cumulative probability distribution and the real travel time rate cumulative probability distribution are distinguished.
  • FIG. 9 is a flow chart of a method for estimating a city road travel time distribution in consideration of a taxi operating state.
  • Binhe Avenue is one of the three main roads in Shenzhen. It is close to Shenzhen Convention and Exhibition Center, shopping park and other commercial centers and Futian Port. Use the actual taxi data on June 3, 2014, on the three sections of the intersection of Caitian Road and Mintian Road, east of the Binhe Avenue, Futian District, Shenzhen.
  • the taxi data from the east to the west during this time period is selected.
  • the driving state of the vehicle will be divided into two driving state categories of "fast” and "slow". Clusters are respectively performed for each road segment in each operation state.
  • the probability density function of each link state and the travel time rate of each link state and the probability of occurrence of each Markov path have been obtained, and then the total path travel time in each operation state can be calculated according to the already constructed model, such as Figure 4 and Figure 5. From the image, it can be found that the peak occurrence position of the probability density function distribution curve in the empty vehicle state differs greatly from the position where the peak appears in the passenger state, that is to say: the estimated path travel time rate under the empty vehicle state is significantly lower than the indistinguishable The path travel time rate under the operating state; the estimated path travel time rate under the passenger state is significantly higher than the path travel time rate without distinguishing the operating state. After calculation, the operation time of the route is 85.
  • the time limit is 1.7259, that is, the path travel time is 275.92 seconds, and the 85th minute of the empty vehicle state is 1.1362, that is, the path travel time is 497.61 seconds, which is more than the indistinguishable operation state. 80.35%, the 85-digit number in the passenger state is 1.9047, that is, the path travel time is 230.74 seconds, which is 16.37% lower than the undifferentiated operating state, and 53.63% lower than the empty vehicle state.
  • the VISSIM software abstracts the road network to simulate the empty car and passenger conditions.
  • the simulation model is built according to the real road network in the study area.
  • the simulated road network includes one-way three 3-car road sections and three three-lane ramps.
  • the total simulation time is 5000 seconds, and the simulation is performed 10 times with different random seeds, using data from 1000 seconds to 4600 seconds with a total duration of 1 hour.
  • the main road import flow is 10080pcu/h
  • the branch import flow is 800pcu/h
  • the taxi proportion is 0.244
  • the passenger-to-custom ratio is 0.736.
  • the speed distribution in the free-riding and passenger-carrying conditions takes the speed distribution from 15 o'clock to 16 o'clock in one hour.
  • the other social vehicles' speed distribution takes the average value of 53.1 km/h, obeying the double lognormal distribution.
  • the simulation results of the travel time rate distribution of each section of the vehicle under the two operating states of passenger and empty vehicles are corrected.
  • the corrected model simulation results are basically consistent with the actual values, indicating that the corrected model can be used to simulate the actual road traffic flow.
  • Fig. 7 is a comparison diagram of the probability density distribution of the estimated travel time rate and the probability density distribution of the real travel time rate by each method
  • Fig. 8 is a cumulative probability distribution diagram thereof.
  • the result of distinguishing the operational status is the result estimated by equation (13).
  • TTRD estimated and TTRD real respectively represent the probability density distribution function of the estimated travel time rate distribution function and the actual travel time rate
  • RANGE represents the distribution function effective interval length.
  • the calculation results of the MAE and PE max for distinguishing the vehicle operating state (improved method) and the vehicle operating state (original method) are as shown in Table 11. It can be known that the error is significantly reduced after distinguishing the vehicle operating state, the average absolute error is reduced by 51.44%, and the maximum percentage error is reduced by 46.83%. Therefore, after considering the operating state of the taxi, the new route travel time estimation method can greatly improve the accuracy of the estimation.
  • the model in equation (13) is used to estimate the total path travel time rate distribution, compared to the path travel time rate and the true value of the existing method (ie, without distinguishing the operational state). It can be clearly seen that there is a significant error in the cumulative distribution of the total path travel time rate obtained by each estimation method, and the quantiles of the three are compared, as shown in Table 12. The error of the estimation method under the operational state is much smaller than the error when the operational state is not distinguished. After the algorithm improvement, the absolute value of the 15-digit error decreased by 70.73%, the absolute value of the median error decreased by 33.90%, and the absolute value of the 85-digit error decreased by 70.94%. Therefore, this again proves that the method proposed by the present invention can greatly improve the accuracy of estimating the travel time of the route using the taxi data.
  • Table 12 Comparison of cumulative probability distribution quantile obtained by two estimation methods: operating state and non-distinguishing operating state

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Analytical Chemistry (AREA)
  • Chemical & Material Sciences (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Operations Research (AREA)
  • Development Economics (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Traffic Control Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种考虑出租车运营状态下的城市道路行程时间分布估算方法,属于城市交通规划及管理的技术领域。根据出租车不同运营状态分别估算路径行程时间分布。在估算路径行程时间分布时,路网中相邻路段并不是独立的,通过加入马尔科夫模型来描述各相邻路段行程时间分布之间的相关性,提高了估算结果的科学性和准确性。根据两种不同运营状态下车辆数比例大小设定权重,得到最终路径行程时间分布。出租车驾驶员在空车和载客两种运营状态下的驾驶行为会有所差异,直接使用出租车数据估算的行程时间与真实值之间必然存在差异。通过采用深圳市的数据对所提出模型进行计算,发现考虑了运营状态之后,计算得到的行程时间分布函数更加精确。

Description

考虑出租车运营状态下的城市道路行程时间分布估算方法 技术领域
本发明属于城市交通规划及管理的技术领域,涉及城市道路行程时间估计和ITS智能交通系统领域,特别适用于基于出租车数据对城市路径行程时间分布的估算。
背景技术
近年来,随着交通拥堵的日益严重和人们对时间价值的重视,路径行程时间分布成为出行者最关心的问题之一。目前,最广泛使用的是基于出租车数据的行程时间估算,因为相比其他数据来源,出租车具有覆盖范围广,实时性高,收集数据成本低等优点。Jenelius E在《Travel time estimation for urban road networks using low frequency probe vehicle data》中提出了一种用低频GPS浮动车观测的车辆运行轨迹估计城市路网出行时间的统计学方法,把路段转向特征和出行条件等作为解释变量,刻画了速度变异性的时空分布背后的影响因素,这对于交通预测十分实用。Chen在《Dynamic Freeway Travel Time Prediction Using ProbeVehicle Data:Link-based vs.Path-based》中利用浮动车数据,分别对基于路段和路径的行程时间估计方法进行比较,然后探讨了浮动车所占比例对于估计精度影响,提出通过卡尔曼滤波进行基于浮动车数据的路段行程时间估计,并进行仿真验证。
关于这类研究方法,目前存在着两个问题:一是将各路段行程时间分布简单地叠加作为路径行程时间分布,会增加路径行程时间估算的误差。二是,出租车驾驶员在空车和载客两种运营状态下的驾驶行为会有差异,所以这种驾驶行为的差异必然导致直接使用出租车数据估算行程时间会与真实值之间存在差异。本发明在此基础上,提出一种基于出租车数据更精确的路径行程时间估算方法,并充分考虑出租车运营状态对路径行程时间估算的影响,从而提出更精确的改进方法。
发明内容
本发明要解决的技术问题是先利用各路段出租车GPS数据得到各路段行程时间分布,然后在此基础上构建模型估算出租车在某一运营状态下路径行程时间分布,最后根据两种不同运营状态下车辆数比例大小,设定权重得到最终路径行程时间分布。
本发明的技术方案:
一种考虑出租车运营状态下的城市道路行程时间分布估算方法,步骤如下:
(1)路段行程时间分布
根据所要研究的路段和时段,对收集到的出租车GPS数据进行筛选、校正和匹配,得到各个路段含有车牌号、经度、纬度、速度和载客状态字段的出租车GPS数据,记为表a;
根据表a中出租车GPS数据,计算同一车牌号的出租车平均速度,再根据各出租车平均速度,通过公式(1)计算同一车牌号出租车的行程时间率:
Figure PCTCN2017102693-appb-000001
Figure PCTCN2017102693-appb-000002
i路段单位距离行程时间,也可以称为行程时间率,单位s/m;
vi:i路段某一车辆平均速度;
新建各个路段含有车牌号、经度、纬度、行程时间率和载客状态字段的出租车数据表,记为表b。按照空车和载客(0和1)对表进行分类,得到载客状态为0的表c和载客状态1的表d。然后用聚类算法,对表c和表d中的行程时间率数据进行聚类,得到两种运营状态下的多种行驶状态数据。最后,分别对同一行驶状态下的出租车行程时间率数据进行拟合,得到各运营状态下、各行驶状态下各路段行程时间率概率密度分布函数。
(2)各运营状态下路径行程时间分布模型
(2.1)马尔科夫链构建
车辆在当前路段上的行程时间率只决定于其上游路段,而与之前路段的行程时间率无关,因此当前路段上的交通空间变化情况与马尔科夫链的结构相类似,路径上各路段行程时间率具有典型的马尔可夫性质,可以使用马尔科夫链对各路段行程时间率之间的相关性进行建模。
定义当前路段的各出租车行程时间率集中分布区间为马尔科夫链的状态,
Figure PCTCN2017102693-appb-000003
为路段link l的边界值集合,ml为link l状态个数,
Figure PCTCN2017102693-appb-000004
为路段link l+1的边界值集合,nl为link l+1状态个数;因此,link l的第一个状态表示为
Figure PCTCN2017102693-appb-000005
最后一个状态表示为
Figure PCTCN2017102693-appb-000006
其中,τl表示link l上各出租车行程时间率;
定义路段link 1的各行驶状态概率分布为马尔科夫链的初始状态概率分布:
Figure PCTCN2017102693-appb-000007
其中,N(i)代表路段link 1的状态i下,例如link 1的行程时间率处于
Figure PCTCN2017102693-appb-000008
内的数据点的数量。
定义两个连续路段的行程时间率分布为马尔科夫链的状态转移概率,则一般状态概率转移矩阵P表示为:
Figure PCTCN2017102693-appb-000009
pi,j=Pr(Sl+1=j|Sl=i)    (4)
其中,Sl表示路段link l的状态;N(i,j)表示在路段link l上处于i状态且在路段link l+1上处于j状态的数据点的数量。
中间路段link l与其下游路段link l-1之间所能匹配上的车辆和中间路段link l与其下游路段link l+1之间所能匹配上的车辆可能不同,对于中间路段而言,当其作为上游路段或下游路段时,分类情况会存在差异;根据这种差异分为两种情形:
(2.2)情形1
任意中间路段link l在其作为上游路段或下游路段时,行驶状态分类完全一致。车辆通过路径时处于
Figure PCTCN2017102693-appb-000010
种不同状态之中的任意一种状态,每一种状态都称为马尔科夫路径。对于给定的马尔科夫路径,所有路段各状态之间的转移概率之积为马尔科夫路径的发生概率:
Figure PCTCN2017102693-appb-000011
假设同一路段上各状态之间的行程时间率分布是条件独立的,某一马尔科夫路径的路径行程时间率分布直接用卷积运算得到:
Figure PCTCN2017102693-appb-000012
式中,运算符(*)表示卷积运算,具体的运算规则表示为:
Figure PCTCN2017102693-appb-000013
(2.3)情形2
任意中间路段link l在其作为上游路段和作为下游路段时,各状态分类不一致的情况。因此,需要对马尔科夫链状态进行一定的修订。模型中引入过渡路段link’l和link”l(l=2,3,…,k-1);其中,link’l的所有状态均与路段link l-1和路段link l中作为下游路段的link l相等,link”l的所有状态均与路段link l和路段link l+1中作为上游路段的link l相等;车辆在路径上的运行过程的新路段序列表示为link 1、…、link l-1、link’l、link”l、link l+1、…、link k;令
Figure PCTCN2017102693-appb-000014
为过渡路段link’l的边界值集合,ml为link’l状态个数;
Figure PCTCN2017102693-appb-000015
为过渡路段link”l的边界值集合,nl为link”l状态个数;τl表示link l上出租车行程时间率;那么,过渡路段link’l与过渡路段link”l之间的状态转移概率矩阵为
Figure PCTCN2017102693-appb-000016
Figure PCTCN2017102693-appb-000017
Figure PCTCN2017102693-appb-000018
新构建的马尔科夫链形成
Figure PCTCN2017102693-appb-000019
条马尔科夫路径。对于给定的新马尔科夫路径,包含过渡路段link’l和link”l的所有路段各状态之间的转移概率之积为马尔科夫路径的发生概率:
Figure PCTCN2017102693-appb-000020
同理,考虑过路路段link’l和link”l之后,可以利用卷积运算得到某一新马尔科夫路径的路径行程时间率分布:
Figure PCTCN2017102693-appb-000021
(2.4)路径行程时间分布叠加
已经知道各马尔科夫路径行程时间率分布以及其发生概率根据马尔科夫链进行叠加便可以得到通过某一路径行程时间率分布(叠加方法如图1所示):
Figure PCTCN2017102693-appb-000022
(3)总路径行程时间估算模型
用步骤(2)中方法分别计算两种运营状态下路径行程时间率分布,然后根据两种运营状态下的数据量对各自分布函数设定权重。总路径行程时间率分布计算公式如下:
TTRD(x)=α0·ttrd0(x)+α1·ttrd1(x)    (13)
其中,ttrd0(x)和ttrd1(x)分别表示在空车状态下和在载客状态下总路径行程时间率的概率密度函数;α0、α1为比例参数,表示各运营状态的车辆占出租车总量的比例即空车比和载客比。
本发明的有益效果:
目前关于道路行程时间的研究多是以路段为对象,来探讨路段行程时间分布,而基于路径的行程时间估算多是将各路段行程时间分布简单的叠加。事实证明,路网中各路段的行程时间分布并不是相互独立的,相邻两条路段之间的车辆运行状态有较强的相关性,因此传统的估算方法忽略了路段之间的时空相关性,这将会产生较大的估算误差。本发明加入马尔科夫模型来描述各相邻路段行程时间分布之间的相关性,将很大程度提高结果的科学性和准确性。
虽然出租车数据可以很好反应和模拟路网中车辆或交通流运行的情况,大部分出租车驾驶员在行驶过程中驾驶行为和决策与一般出行者是不太相同的。出租车驾驶员在空车和载客两种运营状态下的驾驶行为会有所差异。因此直接使用出租车数据估算的行程时间与真实值之间必然存在差异。本发明提出根据两种不同运营状态下车辆数比例大小,设定权重计算最终路径行程时间分布,并通过采用深圳市的数据对所提出模型进行计算,发现考虑了运营状态之后,计算得到的行程时间分布函数更加精确。
附图说明
图1是各组车辆行程时间率叠加方法。
图2是研究区域示意图。
图3(a)是空车状态下link1在“fast”状态下行程时间率概率密度函数图像。
图3(b)是空车状态下link3在“slow”状态下行程时间率概率密度函数图像。
图4是空车和载客状态下路径行程时间率的概率密度分布曲线。
图5是空车和载客状态下路径行程时间率的累积概率分布曲线。
图6是路段行程时间率分布真实值与仿真值比较示意图(部分)。
图7为是否区分运营状态估算行程时间率概率密度分布和真实行程时间率概率密度分布对比图。
图8为是否区分运营状态估算行程时间率累积概率分布和真实行程时间率累积概率分布对比图。
图9为考虑出租车运营状态下的城市道路行程时间分布估算方法的流程图。
具体实施方式
以下结合实例详细叙述本发明的具体实施方式,并模拟发明的实施效果。
1研究对象
选取深圳市福田区滨河大道东往西方向彩田路路口至民田路路口作为案例研究对象,研究区域示意图见附图2。滨河大道是深圳市内的三条主干道之一,邻近深圳会展中心、购物公园等商业中心和福田口岸,交通流量大。使用2014年6月10日深圳市福田区滨河大道东往西方向彩田路路口至民田路路口三个路段上所有的出租车实际数据。
2路段行程时间分布
由于14点至17点期间出租车流量分布以及载客比分布比较一致,所以选取此时段内自东向西方向运行的出租车数据。通过K均值聚类方法,将车辆的行驶状态将分成“fast”和“slow”两种行驶状态类别。对各运营状态下各路段分别进行聚类。
统计各运营状态下各路段的行程时间率,然后对不同行驶状态下的行程时间率分布分别进行拟合。由于当速度越大时,各单位距离行程时间之间差值会变得越小,造成拟合的困难和误差。因此,在计算时取
Figure PCTCN2017102693-appb-000023
以减少误差。其拟合的结果如表1、表2和表3所示。部分拟合结果的函数图像见图3(a)和图3(b)。
表1空车状态下各路段各行驶状态行程时间率的概率密度函数拟合类型和参数估算值
Figure PCTCN2017102693-appb-000024
Figure PCTCN2017102693-appb-000025
表2载客状态下各路段各行驶状态行程时间率的概率密度函数拟合类型和参数估算值
Figure PCTCN2017102693-appb-000026
表3全状态下各路段各行驶状态行程时间率的概率密度函数拟合类型和参数估算值
Figure PCTCN2017102693-appb-000027
3路径行程时间分布
在三种不同的运营状态下,中间路段Link 2在其作为下游路段和上游路段时,它的状态分类是一致的,因此可以直接构成8条马尔科夫路径,满足情形1的使用条件。分别对空车状态和载客状态的数据构建马尔科夫链模型,各运营状态下的初始概率分布如表4所示,概率转移矩阵结果如表5、表6和表7所示,各马尔科夫路径发生概率如表8、表9和表10所示。
表4各运营状态下的初始概率分布表
Figure PCTCN2017102693-appb-000028
Figure PCTCN2017102693-appb-000029
表5空车状态下马尔科夫链模型概率转移矩阵
Figure PCTCN2017102693-appb-000030
表6载客状态下马尔科夫链模型概率转移矩阵
Figure PCTCN2017102693-appb-000031
表7全状态下马尔科夫链模型概率转移矩阵
Figure PCTCN2017102693-appb-000032
表8空车状态下各马尔科夫路径发生概率
Figure PCTCN2017102693-appb-000033
表9载客状态下各马尔科夫路径发生概率
Figure PCTCN2017102693-appb-000034
表10全状态下各马尔科夫路径发生概率
Figure PCTCN2017102693-appb-000035
已经求得各运营状态和各行驶状态下的路段行程时间率的概率密度函数和各马尔科夫路径发生概率,那么可以按照已经构建好的模型计算在各运营状态下的总路径行程时间,如图4和图5所示。从图像可以发现:空车状态时概率密度函数分布曲线的高峰出现位置与载客状态下高峰出现的位置相差较大,也就是说:空车状态下估算的路径行程时间率明显低于不区分运营状态下路径行程时间率;载客状态下估算的路径行程时间率明显高于不区分运营状态下路径行程时间率。经计算,不区分运营状态路径行程时间率85分位数为1.7259即路径行程时间为275.92秒,空车状态下85分位数为1.1362即路径行程时间为497.61秒,比不区分运营状态下增加80.35%,载客状态下85分位数为1.9047即路径行程时间为230.74秒,比不区分运营状态下减少16.37%,比空车状态下减少53.63%。这说明不同运营状态下出租车驾驶员的驾驶行为与决策并不相同,导致其各自的行程时间率的分布情况有明显差异。所以,出租车的运营状态对路径行程时间的估算有较大的影响。并且,空车和载客两种运营状态下行程时间率的分布情况也会根据载客比的不同,对最终估算值产生较大影响。
3.仿真结果对比
为了对比区分运营状态和不区分运营状态下路径行程时间估算方法的准确度和分析运营状态对路径行程时间估算的影响的显著性,通过VISSIM软件抽象研究路网,模拟空车和载客状态下出租车的运行状态以及包括出租车、社会车辆、公交车等所有车辆的运行状态,对比各方法估算路径行程时间分布情况。
根据研究区域真实路网建立仿真模型,仿真路网包括单向3条3车道路段和3条3车道匝道。仿真总时长5000秒,以不同的随机种子仿真10次,采用从第1000秒到第4600秒,总时长为1个小时的数据进行的分析。其中,主干道进口流量为10080pcu/h,支路进口流量为800pcu/h,出租车比重为0.244,载客比为0.736。出租车空车和载客状态下的速度分布取15点至16点一个小时时间段内的速度分布,其他社会车辆速度分布情况取均值53.1km/h,服从双对数正态分布。并通过对仿真参数的调整,校正载客和空车两种运营状态下车辆的各路段行程时间率分布的仿真结果。如附图6所示,校正后的模型模拟得到的结果与实际值基本一致,说明校正后的模型可以用来模拟实际的道路交通流。
图7为各方法估算行程时间率概率密度分布和真实行程时间率概率密度分布对比图,图8为其累计概率分布图。其中区分运营状态结果为式(13)估算的结果。
使用平均绝对误差(Mean Absolute Error)和最大百分误差(Maximum Percent Error)对两种评价指标所估算结果进行精度分析。
Figure PCTCN2017102693-appb-000036
Figure PCTCN2017102693-appb-000037
其中TTRDestimated和TTRDreal分别代表估算行程时间率分布函数和实际行程时间率的概率密度分布函数,RANGE表示分布函数有效区间长度。区分车辆运 营状态(改进方法)和不区分车辆运营状态(原始方法)各自的MAE和PEmax的计算结果如表11所示。可以得知,在区分车辆运营状态之后误差明显减少,平均绝对误差降低了51.44%,最大百分误差降低46.83%。因此,考虑出租车的运营状态之后,新的路径行程时间估算方法可以极大程度地提高估算的准确性。
表11区分运营状态和不区分运营状态两种估算方法所得概率密度分布差值对比表
Figure PCTCN2017102693-appb-000038
使用式(13)中的模型估算总路径行程时间率分布,与现有方法下(即不区分运营状态)的路径行程时间率和真实值对比。可以很明显的看出,各估算方法所得总路径行程时间率的累积分布存在明显的误差,比较三者的分位数,如表12所示。区分运营状态下的估算方法的误差远小于不区分运营状态时的误差。经过算法改进后,15分位数的误差绝对值下降了70.73%,中位数的误差绝对值下降了33.90%,85分位数的误差绝对值下降了70.94%。因此,这再一次证明了本发明所提出的方法能大幅度提高使用出租车数据估算路径行程时间的准确性。
表12区分运营状态和不区分运营状态两种估算方法所得累计概率分布分位数比较表
Figure PCTCN2017102693-appb-000039
1)“绝对误差”与“相对误差”中,正数表示估算值大于实际值,负数表示估算值小于实际值。
2)“数值”为行程时间率的实际值或估算值。

Claims (1)

  1. 一种考虑出租车运营状态下的城市道路行程时间分布估算方法,其特征在于,步骤如下:
    (1)路段行程时间分布
    根据所要研究的路段和时段,对收集到的出租车GPS数据进行筛选、校正和匹配,得到各个路段含有车牌号、精度、纬度、速度和载客状态字段的出租车GPS数据;
    根据出租车GPS数据,计算同一车牌号的出租车平均速度,再根据各出租车平均速度,通过公式(1)计算同一车牌号出租车的行程时间率:
    Figure PCTCN2017102693-appb-100001
    Figure PCTCN2017102693-appb-100002
    i路段单位距离行程时间,称为行程时间率,单位s/m;
    vi:i路段某一车辆平均速度;
    建立各个路段含有车牌号、精度、纬度、行程时间率和载客状态字段的出租车数据表;按照空车0和载客1对出租车数据表进行分类,得到载客状态为0的出租车数据表和载客状态为1的出租车数据表;然后用聚类算法,对载客状态为0的出租车数据表和载客状态为1的出租车数据表中的行程时间率进行聚类,得到两种运营状态下的多种行驶状态数据;最后,分别对同一行驶状态下的出租车行程时间率进行拟合,得到各运营状态下、各行驶状态下各路段行程时间率概率密度分布函数;
    (2)各运营状态下路径行程时间分布模型
    (2.1)马尔科夫链构建
    定义当前路段的各出租车行程时间率集中分布区间为马尔科夫链的状态,
    Figure PCTCN2017102693-appb-100003
    为路段link l的边界值集合,ml为link l状态个数,
    Figure PCTCN2017102693-appb-100004
    为路段link l+1的边界值集合,nl为link l+1状态个数;因此,link l的第一个状态表示为
    Figure PCTCN2017102693-appb-100005
    最后一个状态表示为
    Figure PCTCN2017102693-appb-100006
    其中,τl表示link l上各出租车行程时间率;
    定义路段link 1的各行驶状态概率分布为马尔科夫链的初始状态概率分布:
    Figure PCTCN2017102693-appb-100007
    其中,N(i)代表路段link 1的状态i下,link 1的行程时间率处于
    Figure PCTCN2017102693-appb-100008
    内的数据点的数量;
    定义两个连续路段的行程时间率分布为马尔科夫链的状态转移概率,则一般状态概率转移矩阵P表示为:
    Figure PCTCN2017102693-appb-100009
    pi,j=Pr(Sl+1=j|Sl=i)  (4)
    其中,Sl表示路段link l的状态;N(i,j)表示在路段link l上处于i状态且在路段link l+1上处于j状态的数据点的数量;
    中间路段link l与其下游路段link l-1之间所能匹配上的车辆和中间路段link l与其下游路段link l+1之间所能匹配上的车辆可能不同,对于中间路段,当其作为上游路段或下游路段时,分类情况会存在差异;根据差异分为两种情形:
    (2.2)情形1
    任意中间路段link l,在其作为上游路段或下游路段时,行驶状态分类完全一致;车辆通过路径时处于
    Figure PCTCN2017102693-appb-100010
    l=1,2,…,k种不同状态之中的任意一种状态,每一种状态都称为马尔科夫路径;对于给定的马尔科夫路径,所有路段各状态之间的转移概率之积为马尔科夫路径的发生概率:
    Figure PCTCN2017102693-appb-100011
    假设同一路段上各状态之间的行程时间率分布是条件独立的,某一马尔科夫路径的路径行程时间率分布直接用卷积运算得到:
    Figure PCTCN2017102693-appb-100012
    式中,运算符(*)表示卷积运算,具体的运算规则表示为:
    Figure PCTCN2017102693-appb-100013
    (2.3)情形2
    任意中间路段link l,在其作为上游路段和作为下游路段时,各状态分类不一致的情况;对马尔科夫链状态进行修订,引入过渡路段link’l和link”l,l=2,3,…,k-1;其中,link’l的所有状态均与路段link l-1和路段link l中作为下游路段的link l相等,link”l的所有状态均与路段link l和路段link l+1中作为上游路段的link l相等;车辆在路径上的运行过程的新路段序列表示为link 1、…、link l-1、link’l、link”l、link l+1、…、link k;令
    Figure PCTCN2017102693-appb-100014
    为过渡路段link’l的边界值集合,ml为link’l状态个数;
    Figure PCTCN2017102693-appb-100015
    为过渡路段link”l的边界值集合,nl为link”l状态个数;τl表示link l上出租车行程时间率;那么,过渡路段link’l与过渡路段link”l之间的状态转移概率矩阵为
    Figure PCTCN2017102693-appb-100016
    Figure PCTCN2017102693-appb-100017
    新构建的马尔科夫链形成
    Figure PCTCN2017102693-appb-100018
    条马尔科夫路径;对于给定的新马尔科夫路径,包含过渡路段link’l和link”l的所有路段各状态之间的转移概率之积为马尔科夫路径的发生概率:
    Figure PCTCN2017102693-appb-100019
    同理,引入过路路段link’l和link”l,利用卷积运算得到某一新马尔科夫路径的路径行程时间率分布:
    Figure PCTCN2017102693-appb-100020
    Figure PCTCN2017102693-appb-100021
    (2.4)路径行程时间分布叠加
    将各马尔科夫路径行程时间率分布以及其发生概率根据马尔科夫链进行叠加,得到通过某一路径行程时间率分布:
    Figure PCTCN2017102693-appb-100022
    (3)总路径行程时间估算模型
    用步骤(2)中方法分别计算两种运营状态下路径行程时间率分布,然后根据两种运营状态下的数据量对各自分布函数设定权重,总路径行程时间率分布计算公式如下:
    TTRD(x)=α0·ttrd0(x)+α1·ttrd1(x)  (13)
    其中,ttrd0(x)和ttrd1(x)分别表示在空车状态下和在载客状态下总路径行程时间率的概率密度函数;α0、α1为比例参数,表示各运营状态的车辆占出租车总量的比例即空车比和载客比。
PCT/CN2017/102693 2016-10-08 2017-09-21 考虑出租车运营状态下的城市道路行程时间分布估算方法 WO2018064931A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/778,064 US10748421B2 (en) 2016-10-08 2017-09-21 Method for estimating distribution of urban road travel time in considering operation state of taxi

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610873710.4A CN106384509A (zh) 2016-10-08 2016-10-08 考虑出租车运营状态下的城市道路行程时间分布估算方法
CN201610873710.4 2016-10-08

Publications (1)

Publication Number Publication Date
WO2018064931A1 true WO2018064931A1 (zh) 2018-04-12

Family

ID=57937240

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/102693 WO2018064931A1 (zh) 2016-10-08 2017-09-21 考虑出租车运营状态下的城市道路行程时间分布估算方法

Country Status (3)

Country Link
US (1) US10748421B2 (zh)
CN (1) CN106384509A (zh)
WO (1) WO2018064931A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109379762A (zh) * 2018-09-10 2019-02-22 中国联合网络通信集团有限公司 一种随机信号数据统计方法和系统
CN109885639A (zh) * 2019-03-21 2019-06-14 江苏智通交通科技有限公司 可视化的出租车上下车特征分析方法
CN110992233A (zh) * 2019-12-13 2020-04-10 中国科学院深圳先进技术研究院 一种针对城市聚集事件的应急疏散方法与系统
CN111126819A (zh) * 2019-12-18 2020-05-08 南京理工大学 一种城市行驶工况智能分析方法
CN111223293A (zh) * 2018-11-27 2020-06-02 北京嘀嘀无限科技发展有限公司 分析交通拥堵的系统和方法

Families Citing this family (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106384509A (zh) * 2016-10-08 2017-02-08 大连理工大学 考虑出租车运营状态下的城市道路行程时间分布估算方法
CN107016851B (zh) * 2017-05-24 2019-06-28 大连理工大学 一种量化分析城市建成环境对道路行程时间影响的方法
US10769558B2 (en) * 2018-07-03 2020-09-08 Lyft, Inc. Systems and methods for managing dynamic transportation networks using simulated future scenarios
CN109308803A (zh) * 2018-07-31 2019-02-05 北京航空航天大学 基于随机波动模型的路径行程时间可靠性分析
CN109712389B (zh) * 2018-12-14 2020-06-09 北京航空航天大学 一种基于Copula和Monte-Carlo模拟的路径行程时间估计方法
CN109670703B (zh) * 2018-12-17 2021-10-26 湖南大学 一种基于马尔科夫决策过程的乘客等出租车优化方法
CN111582527A (zh) * 2019-02-15 2020-08-25 拉扎斯网络科技(上海)有限公司 行程时间的预估方法、装置、电子设备和存储介质
CN109978267B (zh) * 2019-03-28 2023-01-03 东南大学 基于城市轨道交通数据的城市微循环公交线路规划方法
CN110264706A (zh) * 2019-04-07 2019-09-20 武汉理工大学 一种基于大数据挖掘的空载出租车辅助系统
US11687840B2 (en) * 2019-05-20 2023-06-27 Honeywell International Inc. Forecasting with state transitions and confidence factors
CN110580404B (zh) * 2019-07-26 2023-05-12 东南大学 一种基于城市多模式交通网络的网络运能确定方法
CN110599769B (zh) * 2019-09-10 2022-08-09 南京城建隧桥经营管理有限责任公司 一种城市路网中分时段道路重要性的分级排序方法
CN111209966B (zh) * 2020-01-07 2023-09-26 中南大学 一种基于马尔科夫链的路径行程时间确定方法及系统
CN111311910B (zh) * 2020-02-19 2020-12-22 中南大学 多层次道路级浮动车异常轨迹探测方法
CN111553539A (zh) * 2020-05-09 2020-08-18 上海大学 一种基于概率模型检验的行驶路径规划方法
CN111983480B (zh) * 2020-08-19 2023-02-28 华晟(青岛)智能装备科技有限公司 一种基于马氏过程的agv电量状态预测方法及系统
WO2022126152A1 (en) * 2020-12-15 2022-06-23 Robert Koelbl Methods for estimating traffic loads and optimising road networks
CN113192342B (zh) * 2021-04-27 2022-03-22 中山大学 基于浮动车数据的城市道路自由流状态百分位车速确定方法
CN114530038B (zh) * 2022-01-11 2023-05-09 江苏大学 一种基于时空数据聚类的出行兴趣区域提取方法及系统
US20230222901A1 (en) * 2022-01-13 2023-07-13 GM Global Technology Operations LLC System and process for mitigating road network congestion
CN115019507B (zh) * 2022-06-06 2023-12-01 上海旷途科技有限公司 城市路网行程时间可靠性实时估计方法
CN115455681B (zh) * 2022-09-02 2023-08-25 江苏城乡建设职业学院 一种面向多载运工具的通勤交通碳排放空间分布估算方法
CN116739322B (zh) * 2023-08-16 2023-12-22 天津市政工程设计研究总院有限公司 基于服务获取指数的出租车供需评估方法、设备及介质
CN117273197B (zh) * 2023-08-28 2024-05-28 长江水上交通监测与应急处置中心 基于航迹及生产信息融合的船舶运营状态预测方法及系统
CN117669998A (zh) * 2024-02-01 2024-03-08 聊城大学 一种考虑乘客载荷变化的公交工况构建方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101727746A (zh) * 2009-09-18 2010-06-09 苏州大学 信号灯控制的城市道路机动车动态行程时间估计方法
CN101739823A (zh) * 2009-12-21 2010-06-16 浙江工业大学 一种适用低频采样的路段平均行程时间测量方法
JP2011075345A (ja) * 2009-09-30 2011-04-14 Nomura Research Institute Ltd 走行リンク確定装置及びリンク旅行時間算出装置
CN104537836A (zh) * 2014-12-30 2015-04-22 北京通博科技有限公司 路段行驶时间分布预测方法
CN104778274A (zh) * 2015-04-23 2015-07-15 山东大学 基于稀疏出租车gps数据的大范围城市路网旅行时间估计方法
CN106384509A (zh) * 2016-10-08 2017-02-08 大连理工大学 考虑出租车运营状态下的城市道路行程时间分布估算方法

Family Cites Families (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4255007B2 (ja) * 2003-04-11 2009-04-15 株式会社ザナヴィ・インフォマティクス ナビゲーション装置、およびその旅行時間算出方法
JP4899756B2 (ja) * 2006-09-29 2012-03-21 アイシン・エィ・ダブリュ株式会社 交通情報作成装置及び交通情報作成方法
CN100492434C (zh) * 2006-11-30 2009-05-27 上海交通大学 交通流状态分析所需的探测车采样量的获取方法
JP5311793B2 (ja) * 2007-10-16 2013-10-09 トヨタ自動車株式会社 脇見状態判定装置
JP5504135B2 (ja) * 2010-11-02 2014-05-28 日本電信電話株式会社 機器状態推定装置
US20130166350A1 (en) * 2011-06-28 2013-06-27 Smart Software, Inc. Cluster based processing for forecasting intermittent demand
DE102012204306A1 (de) * 2012-03-19 2013-09-19 Bayerische Motoren Werke Aktiengesellschaft Verfahren zur Steuerung eines Bereitstellens von Verkehrsinformationsdaten zur Aktualisierung einer Verkehrsinformation
DE102013207895A1 (de) * 2013-04-30 2014-10-30 Kuka Laboratories Gmbh Fahrerloses Transportfahrzeug, System mit einem Rechner und einem fahrerlosen Transportfahrzeug, Verfahren zum Betreiben eines fahrerlosen Transportfahrzeugs
SG11201700895YA (en) * 2014-08-04 2017-03-30 Beijing Didi Infinity Technology & Dev Co Ltd Methods and systems for distributing orders
US10175054B2 (en) * 2015-01-11 2019-01-08 Microsoft Technology Licensing, Llc Predicting and utilizing variability of travel times in mapping services
CN104615897B (zh) * 2015-02-13 2017-07-11 北方工业大学 基于低频gps数据的路段行程时间估算方法
CN104851287B (zh) * 2015-04-15 2017-02-22 浙江大学 基于视频检测器的城市道路路段行程时间检测方法
CN105006147B (zh) * 2015-06-19 2017-03-15 武汉大学 一种基于道路时空关联关系的路段行程时间推断方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101727746A (zh) * 2009-09-18 2010-06-09 苏州大学 信号灯控制的城市道路机动车动态行程时间估计方法
JP2011075345A (ja) * 2009-09-30 2011-04-14 Nomura Research Institute Ltd 走行リンク確定装置及びリンク旅行時間算出装置
CN101739823A (zh) * 2009-12-21 2010-06-16 浙江工业大学 一种适用低频采样的路段平均行程时间测量方法
CN104537836A (zh) * 2014-12-30 2015-04-22 北京通博科技有限公司 路段行驶时间分布预测方法
CN104778274A (zh) * 2015-04-23 2015-07-15 山东大学 基于稀疏出租车gps数据的大范围城市路网旅行时间估计方法
CN106384509A (zh) * 2016-10-08 2017-02-08 大连理工大学 考虑出租车运营状态下的城市道路行程时间分布估算方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CHEN, JIN: "Study on Evaluation Methodology of Path Travel Time Reliability", SCIENCE -ENGINEERING (B), CHINA MASTER'S THESES FULL-TEXT DATABASE, vol. II, no. 3, 15 March 2016 (2016-03-15) *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109379762A (zh) * 2018-09-10 2019-02-22 中国联合网络通信集团有限公司 一种随机信号数据统计方法和系统
CN109379762B (zh) * 2018-09-10 2021-09-14 中国联合网络通信集团有限公司 一种随机信号数据统计方法和系统
CN111223293A (zh) * 2018-11-27 2020-06-02 北京嘀嘀无限科技发展有限公司 分析交通拥堵的系统和方法
CN111223293B (zh) * 2018-11-27 2021-03-02 滴滴智慧交通科技有限公司 分析交通拥堵的系统和方法
CN109885639A (zh) * 2019-03-21 2019-06-14 江苏智通交通科技有限公司 可视化的出租车上下车特征分析方法
CN110992233A (zh) * 2019-12-13 2020-04-10 中国科学院深圳先进技术研究院 一种针对城市聚集事件的应急疏散方法与系统
CN110992233B (zh) * 2019-12-13 2024-04-23 中国科学院深圳先进技术研究院 一种针对城市聚集事件的应急疏散方法与系统
CN111126819A (zh) * 2019-12-18 2020-05-08 南京理工大学 一种城市行驶工况智能分析方法
CN111126819B (zh) * 2019-12-18 2022-08-16 南京理工大学 一种城市行驶工况智能分析方法

Also Published As

Publication number Publication date
US20180350237A1 (en) 2018-12-06
CN106384509A (zh) 2017-02-08
US10748421B2 (en) 2020-08-18

Similar Documents

Publication Publication Date Title
WO2018064931A1 (zh) 考虑出租车运营状态下的城市道路行程时间分布估算方法
CN106781468B (zh) 基于建成环境和低频浮动车数据的路段行程时间估计方法
CN104778834A (zh) 一种基于车辆gps数据的城市道路交通拥堵判别方法
CN108335483B (zh) 交通拥堵扩散路径的推断方法及其系统
Lu et al. Using taxi GPS data for macroscopic traffic monitoring in large scale urban networks: calibration and MFD derivation
WO2022166239A1 (zh) 一种车辆行驶方案的规划方法、装置和存储介质
CN106251642A (zh) 一种基于实时公交车gps数据的公交路链速度计算方法
CN105869402B (zh) 基于多类型浮动车数据的高速公路路段速度修正方法
CN106971546B (zh) 基于公交车gps数据的路段公交车渗透率估计方法
CN110796876A (zh) 基于卡尔曼滤波的路段车辆总数估计方法
Oh et al. Real-time inductive-signature-based level of service for signalized intersections
Zhang et al. Assessing the impacts of signal coordination on the crash risks of various driving cohorts
Sobreira et al. Disaggregated traffic conditions and road crashes in urban signalized intersections
Wei et al. Turn-level network traffic bottleneck identification using vehicle trajectory data
Srikanth et al. A modified approach for estimation of Passenger Car Units on intercity divided multilane highways
Bakibillah et al. Sustainable eco-driving strategy at signalized intersections from driving data
Qin et al. Spatiotemporal K-Nearest Neighbors Algorithm and Bayesian Approach for Estimating Urban Link Travel Time Distribution From Sparse GPS Trajectories
Liang et al. Traffic impedance estimation driven by trajectories for urban roads
Kim Simultaneous calibration of a microscopic traffic simulation model and OD matrix
Behrisch et al. Comparing performance and quality of traffic assignment techniques for microscopic road traffic simulations
CN107590998B (zh) 一种基于浮动车数据的道路节点运行状态识别系统
Iscan Evaluation the level of service at a roundabout: A case study on Al-Ibrahimeya roundabout in Alexandria, Egypt
Crowther A Comparison of CORSIM and INTEGRATION for the Modeling of Stationary Bottlenecks
Talvitie et al. A parametric access network model
Han et al. A study on Autonomous Vehicles’ Driving Behavior Based on Open Road Test Data

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17857812

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17857812

Country of ref document: EP

Kind code of ref document: A1