WO2023273292A1 - 基于多源数据融合的居民出行链生成方法及共乘查询方法 - Google Patents

基于多源数据融合的居民出行链生成方法及共乘查询方法 Download PDF

Info

Publication number
WO2023273292A1
WO2023273292A1 PCT/CN2022/070988 CN2022070988W WO2023273292A1 WO 2023273292 A1 WO2023273292 A1 WO 2023273292A1 CN 2022070988 W CN2022070988 W CN 2022070988W WO 2023273292 A1 WO2023273292 A1 WO 2023273292A1
Authority
WO
WIPO (PCT)
Prior art keywords
travel
itinerary
bus
user
matching
Prior art date
Application number
PCT/CN2022/070988
Other languages
English (en)
French (fr)
Inventor
张晓春
林涛
陈振武
梁晨
张稷
周勇
高彦
刘星
Original Assignee
深圳市城市交通规划设计研究中心股份有限公司
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 深圳市城市交通规划设计研究中心股份有限公司 filed Critical 深圳市城市交通规划设计研究中心股份有限公司
Publication of WO2023273292A1 publication Critical patent/WO2023273292A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/123—Traffic 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

Definitions

  • the invention relates to the technical field of traffic data processing, in particular to a method for generating travel chains of residents based on multi-source data fusion and a method for carpooling query.
  • Mobile phone signaling data can be continuously supplied in large quantities for a long time, but the mobile phone signaling data only has trajectories, and the analysis of travel modes, vehicles, and specific departure and destinations is insufficient. If only mobile phone signaling data is used to establish a travel chain model, it will be because The type of information is too single, and the accuracy of the model is insufficient.
  • the problem solved by the invention is that the existing travel chain model is established only by using the signaling data of the mobile phone, the information type is single, and the accuracy of the model is insufficient.
  • the present invention proposes a resident travel chain generation method based on multi-source data fusion, including:
  • the mobile phone signaling data perform jump data cleaning processing, drift position confirmation processing, different operators and user identification processing and dwell time analysis processing on the mobile phone signaling data, and obtain the user's travel information, wherein the travel information consists dwell points, dwell time and travel trajectory;
  • the itinerary trajectory of the bus travel itinerary is spatially matched with the bus line trajectory, and then based on the start and end points of the bus travel itinerary and the corresponding times of the start and end points of the itinerary, the bus Time-space matching of travel itinerary and bus number information to generate the first matching result;
  • the user's itinerary includes a track traffic travel itinerary
  • spatially match the itinerary track of the track traffic travel itinerary with the track of the rail transit line, and compare the start and end points of the track traffic travel itinerary, the corresponding time of the start and end points of the track traffic with the track Carry out spatio-temporal matching on the traffic number information to generate the second matching result;
  • the user's itinerary includes a car travel itinerary
  • the start and end points of the itinerary and the corresponding time of the start and end points of the car travel itinerary
  • the time and space of the car travel itinerary and the car in the preset database are compared. matching, generating a third matching result
  • the number of trains taken by the user for each itinerary is obtained, and combined with the start and end points of each itinerary of the user, the user's travel chain is generated.
  • said spatially matching the itinerary trajectory of the bus travel itinerary with the bus line trajectory includes:
  • the first path set into at least two path subsets, wherein, when the path subset contains two or more paths, the two or more paths All are paths that are adjacent in path sequence; at least two of the path subsets are respectively matched with the second path set according to the path order; when each of the path subsets is matched to at least one bus line trajectory
  • the itinerary of the bus travel itinerary is divided according to each of the path subsets to obtain the corresponding itinerary of each of the path subsets, and the bus routes matched by each of the path subsets
  • the trajectories are respectively used as the candidate bus line trajectories corresponding to the corresponding itinerary.
  • performing time-space matching on the bus travel itinerary and bus number information based on the start-destination point of the itinerary of the bus travel itinerary and the time corresponding to the start-destination point of the itinerary, and generating the first matching result includes:
  • the arrival time of all bus times arriving at the starting/destination bus station is obtained, and based on the time corresponding to the starting and ending points of the itinerary and the arrival time, the itinerary and each of the bus times are calculated respectively. Time matching of trains;
  • the bus number with the highest matching degree with the travel time is taken as the bus number matching with the user.
  • the bus number with the maximum matching degree of travel time corresponding to the candidate bus line trajectory is used as the bus number matched with the user, it also includes:
  • the bus card swiping data of the target trip and count the first number of boarders or the first number of people getting off the bus corresponding to the stay time of the target trip at the start/end bus station according to the bus swiping data, wherein the target trip is the bus number with the maximum matching degree of travel time corresponding to the candidate bus route trajectory;
  • the matching result corresponding to the target train is updated based on the difference, and the train matched by the user is updated according to the updated matching result corresponding to the target train.
  • P refers to the probability of carpooling
  • x is the number of vacant seats in the train
  • M 0 , ⁇ + , ⁇ - are the preset values of positive real numbers
  • P( ⁇ ) is the travel characteristic parameter
  • f(x) is the above Car probability
  • f(x)/ ⁇ f(x) is the normalization process for distribution probability.
  • the time and space matching of the car travel itinerary and the cars in the preset database is performed to generate a third matching Results include:
  • the number of trains taken by the user for each itinerary is obtained, and combined with the start and end points of the user's itinerary for each itinerary, the user's travel number is generated.
  • the travel chain includes:
  • the user uses a private car or corresponding public transportation to travel, wherein the public transportation travel includes bus travel, track traffic travel and car travel;
  • the user When the user has more than one public transportation line successfully matched, the user is randomly assigned to one of the trains, and through iterative calculation, the expanded bus card data, rail transit card data and/or car payment data are used as the target data , adjust the number of people matched to each travel mode and each vehicle. After meeting the target data, a certain proportion of users who cannot be matched will be selected as passengers entering the gate at the side gate of the rail transit ground station or coin-operated passengers on the bus, and the rest Users are classified as private car travel;
  • a discrete choice model is established for the individual user, and the departure and destination are selected through the discrete choice model to generate a resident travel chain.
  • the matching of the travel mode for each trip of the user to obtain the travel mode identification of each trip includes:
  • the user's travel mode is rail transit travel
  • P public transport
  • driving refers to the probability that the itinerary is a car travel itinerary
  • T bus refers to said bus time.
  • the present invention also proposes a shared ride query method based on the knowledge graph and the resident travel chain, including:
  • the knowledge graph includes user entities, stay location entities, non-traffic facility entities, train number entities, line entities, vehicle entities, and transportation infrastructure entities.
  • the user entities and the resident The individual users of the travel chain correspond, the stay location entity and the non-traffic facility entity correspond to the activities of the resident travel chain, the train number entity, the line entity, the vehicle entity, and the transportation infrastructure entity Corresponding to the travel chain of the resident travel chain;
  • the present invention also proposes a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is read and run by a processor, the aforementioned multi-source data fusion-based resident travel is realized A chain generation method, or a carpooling query method based on knowledge graphs and resident travel chains as described above.
  • public transport vehicles have bus card data, rail transit AFC data and other actual data to check the matching results, which can ensure a certain matching accuracy;
  • each trip of the user is matched to a specific vehicle , can expand the application scenarios and scope of application of the travel chain, such as obtaining carpooling conditions based on the travel chain, providing a basis for epidemic prevention and control, case detection, and travel feature analysis, which is conducive to the popularization and application of the present invention.
  • Fig. 1 is a schematic diagram of an embodiment of a method for generating a resident travel chain based on multi-source data fusion in the present invention
  • Fig. 2 is a schematic diagram of another embodiment of the method for generating a resident travel chain based on multi-source data fusion in the present invention
  • Fig. 3 is a schematic diagram of an embodiment of determining a user's travel mode in a resident travel chain generation method based on multi-source data fusion in the present invention
  • Fig. 4 is a schematic diagram of an embodiment of user travel trajectory and bus data matching in the resident travel chain generation method based on multi-source data fusion in the present invention
  • Fig. 5 is a schematic diagram of an embodiment of the overall structure of the knowledge map database entity and relationship in the carpooling query method based on the knowledge map and the resident travel chain of the present invention
  • Fig. 6 is a schematic diagram of an embodiment of the carpooling query method based on the knowledge graph and the resident travel chain of the present invention.
  • the mobile phone signaling data based on the mobile phone signaling data, and only consider the individual location and travel trajectory. Considering travel modes, vehicles and other traffic information, as far as traffic analysis is concerned, the parts related to traffic vehicles are not modeled, and the travel chain model based on this is not accurate enough. At the same time, only mobile phone signaling data is used, and the type of data used is single, and the accuracy deviation caused by mobile phone signaling data is difficult to make up for. In addition, the signal point of the mobile phone is the GPS data of the base station, and there are many types of travel places under the coverage of a base station. Without other data support, specific analysis cannot be realized.
  • public transportation below includes public transportation, rail transit, taxis and online car-hailing (in this paper, car travel includes taxis and online car-hailing).
  • Fig. 1 and Fig. 2 are schematic diagrams of an embodiment of a method for generating a resident travel chain based on multi-source data fusion in the present invention.
  • described resident travel chain generation method based on multi-source data fusion comprises:
  • Step S100 obtain the signaling data of the mobile phone, perform jump data cleaning processing, drift position confirmation processing, different operators and user identification processing and residence time analysis processing on the mobile phone signaling data, and obtain the travel information of the user, wherein the The described line information includes dwell points, dwell time and travel trajectory.
  • Mobile phone signaling data is a set of point arrays composed of coordinates of mobile phone signaling base stations.
  • Mobile phone signaling data can be continuously supplied in large quantities for a long time.
  • Through the mobile phone signaling data analyze the dwelling point, dwelling time and travel trajectory of the corresponding individual. After analysis, the dwelling point, dwelling time and travel trajectory of the corresponding individual can be obtained.
  • jumping data cleaning/drifting position confirmation due to the instability of the base station used by the mobile phone, two situations may occur: the first is that the user does not actually travel, but because it is between multiple base stations, its location will be in the Jumping back and forth between several base stations; the second is that there is an unreasonable jump point between a group of continuous travel points of the user. Based on this, it is necessary to clean up the jump data and confirm the drift position for these two types of point data.
  • the user's daily location points Based on the mobile phone signaling data, obtain all the user's daily location points, and judge the dynamic and static status of the points through the change relationship of the points, such as dwell time, interval distance and other information. For a point without motion, it is considered that the user is staying at the point, that is, the user's residence point.
  • the point records between two dwell points are the user's travel trajectory information.
  • the user's day's point is divided into a series of travel activities, and the travel characteristics such as the departure point and the arrival point of each travel activity are obtained through the point position.
  • the travel information of the user including the dwelling point, dwelling time, itinerary track, departure place, arrival place, etc.
  • the statistical expansion of mobile phone signaling data can be carried out, and the travel information of users can be generated based on the expanded mobile phone signaling data .
  • Step S200 segmenting the user's itinerary based on the travel information, and obtaining the starting and ending points of all the user's itineraries, the corresponding time of the starting and ending points of the itinerary, and the itinerary trajectory.
  • the user's itinerary can be cut by methods such as mobile phone signaling point clustering.
  • the specific cutting method is an existing technology and will not be described here.
  • the start and end points of the itinerary can be obtained, for example, the residence time longer than the preset duration is taken as the start and end point of the itinerary.
  • the corresponding time of the starting and ending points of the itinerary can be directly read, that is, the corresponding time of the starting point of the itinerary and the corresponding time of the end of the itinerary.
  • Step S300 traverse each itinerary of the user, perform travel mode matching on each itinerary of the user, and obtain the travel mode identifier of each itinerary.
  • the travel modes include bus travel, rail transit travel, and car travel.
  • car travel includes taxi travel, online car-hailing travel, etc.
  • the user's travel information may contain multiple itineraries, and these itineraries may all use the same travel mode, such as public transportation, or different travel modes, for example, itinerary AB uses public transportation, and itinerary BC uses rail transit. Match the user's travel mode and subsequent matching with public transportation lines and vehicles in units of a single trip.
  • step S200 includes:
  • T bus between the origin and destination points is greater than the driving time T driving , it is a long-distance travel: if T mobile phone /T driving is greater than 1 and T mobile phone /T bus is close to 1, it is a high probability event that the user travels by bus.
  • T bus between the origin and destination points is greater than the driving time T driving , it is a long-distance travel at this time: T mobile phone /T driving is close to 1, while T mobile phone /T bus is less than 1. It is determined that car travel is a high probability event.
  • bus travel When the bus time between the origin and destination is approximately equal to the driving time, it is a short-distance travel: it is difficult to distinguish which mode of transportation, and the confidence level is close to 0.5. Both bus travel and car travel can be used as the user’s travel mode. Traffic vehicle operation data and GPS data are accurately matched to determine the user's final travel mode.
  • the travel mode and possible route of the user's itinerary are preliminarily determined based on the trajectory of the underground base station, for example,
  • the user's travel mode/transportation mode is subway travel, and the possible route is Line 1; if the user's itinerary is not underground rail transit, use the route planning API to capture the bus time and driving time of the start and end points of the user's itinerary, and obtain the start and end points of the user's itinerary
  • the actual travel between the actual travel, the bus travel probability P (public transport) and the car travel probability P (driving) are calculated by the second preset formula above, and it is judged whether P (public transport)/P (driving) is greater than the preset value N, if If it is less than the preset value N, it may be traveling by car.
  • the travel trajectory and the car travel trajectory are fitted at the road level to determine the travel mode and possible route of the user's trip; if P( Bus)/P (driving) is greater than the preset value N, then it may be a bus trip, and the track point data of the bus line is compared with the point array of the mobile phone signaling on the vector (you can use the Euclidean of two vectors distance to represent the similarity), to screen out possible route records, that is, to carry out initial screening of travel routes, and to carry out subsequent matching based on public transport vehicle operation data and GPS data, etc., on the basis of the initially screened travel routes. Precise screening.
  • the route planning API determines whether there is a transfer, and if so, split the bus line trajectory , which identifies all possible routes the user can travel.
  • the user's itinerary analyzed by mobile phone signaling is matched with the public transportation operation data, and the public transportation method and trip number used by each user's itinerary are analyzed through the matching of transportation mode and time and space.
  • Step S400 when the user's itinerary includes a bus travel itinerary, spatially match the itinerary trajectory of the bus travel itinerary with the bus line trajectory, and then based on the start and end points of the bus travel itinerary and the corresponding times of the start and end points of the itinerary, the The bus travel itinerary and the bus number information are matched spatiotemporally to generate a first matching result.
  • the trajectory of the bus line can be obtained by obtaining the GPS data of the bus and performing map matching, including information such as the location of the station and the travel route.
  • the bus number information is obtained by obtaining bus operation data.
  • the bus line operation data includes information such as stops along the line, arrival time, license plate, and bus number.
  • the bus number information includes information such as stops along the line and arrival time.
  • the first matching result that is: whether there is a bus line and bus number matching the user, and if so, the bus line and bus number matching the user.
  • Step S500 when the user's itinerary includes a rail transit travel itinerary, spatially match the travel trajectory of the rail traffic travel itinerary with the rail transit line trajectory, and match the start and end points of the travel itinerary and the start and end points of the travel itinerary Time and space-time matching is performed with rail transit train number information to generate a second matching result.
  • the passenger data of the rail transit underground station is extracted and matched with the rail transit AFC data; based on the mobile phone signaling data of the underground station that has experienced more than two passengers, the time when each train arrives at each station is estimated, so that Obtain the train number; match the user of the mobile phone signaling data to the rail transit line and train number taken by the individual in the underground station.
  • rail transit AFC card swiping data and user travel information obtained based on mobile phone signaling data users are assigned to each train and correspond to the bus card. Through long-term analysis, the corresponding relationship between the bus card used by the user and the mobile phone signaling user ID will be further clarified.
  • the same user's rail-bus data Integrate to obtain a more accurate travel chain for users.
  • the second matching result that is: whether there are rail transit lines and train numbers that match the user, and if so, the rail transit lines and train numbers that match the user.
  • Step S600 when the user's itinerary includes a car travel itinerary, based on the travel trajectory of the car travel itinerary, the start and end point of the itinerary, and the time corresponding to the start and end point of the itinerary, compare the car travel itinerary with the car travel itinerary in the preset database The car performs space-time matching to generate a third matching result.
  • the preset database stores information such as the start and end locations, time, travel route, and passenger status of car travel.
  • Cars include taxis and online car-hailing. Based on the GPS track data of taxis/network-hailing cars, map matching is performed and the complete travel itinerary of taxis/network-hailing cars is obtained throughout the day.
  • Step S700 according to the first matching result, the second matching result, and the third matching result, obtain the number of trains taken by the user for each itinerary, and combine the starting and ending points of each itinerary of the user to generate the user's travel chain.
  • the second matching result and the third matching result After obtaining the first matching result, the second matching result and the third matching result, it is possible to determine whether each trip of the user is successfully matched, and if it is successful, the matched line and train number information, combined with the start and end points of each trip , to generate the travel chain of the user.
  • the number of rides in step S700 when the vehicle is a bus or a rail transit vehicle, it has a relatively fixed operating frequency, and the number of rides includes the vehicle identification (such as the license plate number or the number of the car) and the number of vehicles.
  • the number of rides includes the vehicle identification (such as the license plate number). In one embodiment, it is the number of each journey of the car.
  • the number of rides also includes the number of the journey in addition to the vehicle identification.
  • the travel mode of each trip of the user is roughly matched, and then multi-source data such as mobile phone signaling data, public transportation operation data, and GPS data are fused to perform public transportation lines and public transportation vehicles for each trip of the user.
  • multi-source data such as mobile phone signaling data, public transportation operation data, and GPS data are fused to perform public transportation lines and public transportation vehicles for each trip of the user.
  • the fine matching makes up for the defect of large error in mobile phone signaling data.
  • it analyzes the travel mode of individual users and the specific vehicles they take, and considers people and public transportation vehicles comprehensively, realizing the joint modeling of people and public transportation vehicles.
  • spatially matching the itinerary trajectory of the bus travel itinerary with the bus line trajectory described in step S400 includes:
  • the itinerary trajectory of the bus travel itinerary is fitted to the road, and the first path set and the path sequence corresponding to the itinerary trajectory are obtained.
  • a bus line trajectory corresponds to a second path set, and the second path set and the path sequence corresponding to the bus line track to be matched with the bus travel itinerary are obtained.
  • the first path set has 6 paths such as A, B, C, D, E, and F, and the order of the paths is A->B->C->D->E->F, and the first path set and The second path set performs full path matching, that is, six paths such as A, B, C, D, E, and F are matched with the paths in the second path set.
  • the second path set includes all the paths in the first path set, and the path order of all the paths in the second path set is the same as the path order in the first path set, the second path is determined
  • the set matches all paths of the first path set, that is, the bus line trajectory corresponding to the second path set matches all paths of the first path set.
  • the bus line trajectory matching the full path of the first path set is used as the candidate bus line trajectory corresponding to the bus travel itinerary.
  • the first path set is split into at least two path subsets, wherein, when the path subset contains two or more paths, the two or more paths are sequentially adjacent to each other at least two path subsets are matched with the second path set according to the path order; when each path subset matches at least one bus line trajectory, it is determined that a transfer occurs, and the bus travel itinerary
  • the itinerary is segmented according to each route subset, and the itinerary corresponding to each route subset is obtained, and the bus line trajectories matched by each route subset are respectively used as candidate bus line trajectories corresponding to the corresponding itinerary.
  • the first path set has 6 paths such as A, B, C, D, E, and F, and the order of the paths is A->B->C->D->E->F.
  • the first path set is split into two path subsets "A->B->C” and "D->E->F", and "A->B ->C", "D->E->F" are respectively matched with the path set corresponding to the bus line track, if "A->B->C” matches the bus line track 1, "D->E-> F" is matched to bus line track 2, then it is determined that a transfer occurs, that is, a transfer from bus line track 1 to bus line track 2 occurs.
  • the itinerary of the bus travel itinerary is divided according to each path subset, and two sections of the itinerary "A->B->C" and "D->E->F" are obtained, and the bus line trajectory 1 is used as "A-> For the candidate bus route trajectory corresponding to B->C", the bus route trajectory 2 is used as the candidate bus route trajectory corresponding to "D->E->F".
  • the first path set is split into two path subsets, and the two path subsets are paths that are sequentially adjacent to each other.
  • the concept of split points is introduced. For example, A, The six paths of B, C, D, E, and F are split into two path subsets "A->B->C” and "D->E->F", and then D is the split point. Because splitting the first path set into two path subsets has many different split results, different split points can get different split results, multiple split points can be set in advance, first, according to the first split Split the first path set by points to obtain two path subsets, and perform full path matching on the two path subsets with the second path set according to the path order.
  • each path subset matches at least one bus line trajectory , then the bus line trajectories matched by each path subset are used as the candidate bus line trajectories corresponding to the corresponding itinerary, and the matching of the first path set is completed. points until each path subset matches at least one bus line trajectory or all split points are matched. When all the split points are matched and there are still path subsets that do not match the bus line trajectory, the decision The bus line matching fails, and the bus travel itinerary is classified as a car trip or a private car trip.
  • step S400 based on the start and end points of the bus travel itinerary and the time corresponding to the start and end points of the itinerary, the bus travel itinerary and the bus number information are spatiotemporally matched, and the first matching result is generated including :
  • Step S401 obtaining bus number information corresponding to the trajectory of the candidate bus line.
  • the characteristic of public transportation is that the trajectory of the vehicle is determined, but the repetition degree of time and space is high, and the frequency of departure is high. Therefore, it is necessary to search for all possible lines and trains in time and space, and finally find a matching scheme that meets the threshold requirements.
  • step S401 to step S405 are executed for each candidate bus route track.
  • Step S402 matching the starting and ending points of the itinerary corresponding to the candidate bus route trajectory to the candidate bus route trajectory to obtain the starting/destination bus station corresponding to the starting and ending points of the itinerary.
  • the starting and ending points of the itinerary are the starting and ending locations of the individual travel. Match the start point and end point of the itinerary to the trajectory of the candidate bus line to obtain the start/destination bus stops corresponding to the start and end points of the itinerary, that is, the start bus stop and the end bus stop.
  • Step S403 from the bus number information, obtain the arrival time of all bus times at the start/destination bus station, and calculate the itinerary and each The time matching degree of the bus times.
  • Time determines the time matching degree, wherein, the shorter the waiting time, the greater the time matching degree, the higher the matching degree between the corresponding train number and the user, the greater the waiting time, the smaller the time matching degree, and the lower the matching degree between the corresponding train number and the user lower.
  • the difference between the time corresponding to the starting point of the user's trip and the arrival time of each bus at the bus station can be used as the waiting time.
  • Step S404 when the time matching degrees between the travel trajectory and all bus times are less than the second preset threshold, it is determined that the matching fails, and it is determined that the user travels in a private car.
  • time matching degree it is possible to eliminate the bus lines whose trajectories match the points but have no trains at the corresponding time. If the time matching degrees of the travel trajectory and all the trips of the candidate bus route trajectory are less than the second preset threshold, it means that the matching degree of all the trips of the candidate bus route trajectory and the user is too low, and it is determined that the matching fails, and it is determined that the user uses a private car to travel.
  • Step S405 when the time matching degree between the travel trajectory and at least one bus number is greater than or equal to the second preset threshold, the bus number with the highest matching degree of travel time corresponding to the candidate bus line trajectory is taken as the matching with the user the number of trips.
  • the target train the train with the greatest time matching degree with the travel trajectory.
  • step S405 it also includes:
  • Step S406 obtain the bus card swiping data of the target train, and count the first number of people on board or the first number of people getting off the bus corresponding to the stay time of the target bus at the departure/destination bus station according to the bus card swiping data, wherein, the The target bus number is the bus number with the greatest matching degree with the travel time corresponding to the trajectory of the candidate bus line.
  • step S405 After step S405 is executed, a preliminary matching result is obtained, and in order to ensure that the matching result conforms to reality, the matching result is verified through the bus swiping card data.
  • the stay time of the target train number at the origin/end bus station refers to the time from the arrival of the target train number at the origin/end bus station to the time when it leaves the origin/end bus station, which is relatively close to the corresponding time of the user’s itinerary.
  • the time is to board the target train number and get off the target train number at the starting/destination bus station.
  • the first boarding number corresponding to the stay time of the target train at the departure/destination bus station refers to the first boarding number of the target train number during the stay at the departure bus station
  • the first number of people getting off refers to the number of people on the target bus at the destination bus. The number of people who get off the bus first during the sojourn time of the station.
  • Step S407 counting the second number of boarders or the second number of people who get off at the starting/destination bus station and the stay time matched to the target train number.
  • Step S408 calculating the difference between the first number of people getting on and the second number of people getting on, or calculating the difference between the first number of people getting off and the second number of people getting off.
  • Step S409 when the difference is smaller than the preset difference, keep the matching result corresponding to the target train number.
  • the difference is less than the preset difference, it means that the preliminary matching result is in line with the actual situation, so the matching result corresponding to the target train number can be retained, that is, the matching relationship between the target train number and the user is retained, and the target train number and its line are used as the matching relationship with the user. Matching train numbers and bus lines.
  • Step S410 when the difference is greater than or equal to the preset difference, update the matching result corresponding to the target train based on the difference, and update the user matching result according to the updated matching result corresponding to the target train the number of trains.
  • the difference is greater than or equal to the preset difference, it means that the preliminary matching result does not conform to the actual situation, so it is necessary to adjust the matching result, and update the matching result corresponding to the target train based on the difference.
  • the travel mode of an equal number of users is determined to be traveling by private car, and the matching results of the remaining users are retained. Among them, the users who are determined to travel by private car are excluded from the users to be analyzed, and no further analysis is performed. Checking the matching results through the bus swiping card data can ensure that the matching results conform to the actual situation, thereby ensuring the accuracy of the established resident travel chain.
  • update the possible line set that is: take the bus line track whose coincidence degree is greater than the threshold as the line that the user may take; if not, read another bus line track, return to execute the calculation of the user travel chain track and bus The steps of coincidence degree of line track;
  • the waiting time is less than the threshold, look for the bus number with the least waiting time, and count the number of people getting on/off at the departure/destination bus station of the bus number at the corresponding time according to the bus card data, and count the matching to each station within the same time period Whether the number of people on the bus is consistent with the number of people on the bus based on the bus swiping card data. If not, the matching fails. If so, the bus number and its route with the least waiting time will be used as the bus number and bus line that match the user.
  • P refers to the probability of carpooling
  • x is the number of vacant seats in the train
  • M 0 , ⁇ + , ⁇ - are the preset values of positive real numbers
  • P( ⁇ ) is the travel characteristic parameter
  • f(x) is the above Car probability
  • f(x)/ ⁇ f(x) is the normalization process for distribution probability.
  • the number of trains taken by the user obtained through the above steps is the public transport line/vehicle that the user is most likely to take, and there is a certain probability whether the user actually takes the matching public transport vehicle to the destination.
  • the actual number of trains taken by the first user is known, the travel chain determined through the above steps is obtained, and the second user who takes the same number of trains as the first user is determined, and the ride-sharing probability between the first user and the second user is calculated.
  • the number of vacant seats shared by the first user and the second user is obtained from the number of people getting on and off the bus. Order or OD of large sample statistics is obtained.
  • the probability of carpooling can be calculated through the above-mentioned first preset formula.
  • the time and space matching of the car travel itinerary and the cars in the preset database is performed to generate a third matching Results include:
  • Step S420 traversing the travel route information of each car in the preset database, obtaining the starting and ending points of the car travel, combining the starting and ending points of the trip, judging the travel itinerary of the car and the travel route of the car Whether the start and end positions of the information meet the preset spatial error.
  • Step S421 if yes, that is, the start and end positions of the car travel itinerary and the car travel route information meet the preset spatial error, then judge the car travel itinerary and the car travel time according to the time corresponding to the start and end points of the itinerary Whether the start and end times of the route information meet the preset time error.
  • the car travel route information includes the start and end locations and time.
  • the start and end locations and time can be directly obtained from the car travel route information to combine the user's start and end points and corresponding times of the travel start and end points respectively. , to determine whether the starting and ending points all meet the preset spatial error, and whether the starting and ending time meet the preset time error.
  • step S421 and subsequent steps are not executed.
  • Step S422 if yes, that is, the start and end times of the car travel itinerary and the car travel route information both meet the preset time error, then calculate the trajectory coincidence degree of the car travel itinerary and the car travel route information, and determine the Whether the track coincidence degree is greater than a third preset threshold.
  • trajectory coincidence degree is less than or equal to the third preset threshold, it is determined that the matching fails, and it is determined that the user is traveling by private car.
  • Step S423 if yes, that is, if the trajectory coincidence degree is greater than the third preset threshold, then determine the vehicle travel route information with the highest trajectory coincidence degree, and use the vehicle corresponding to the vehicle travel route information with the highest trajectory coincidence degree as the The train number matched by the user.
  • the car travel path information with the highest trajectory coincidence degree is selected, and the vehicle travel path information with the highest trajectory coincidence degree is selected.
  • the car trips in the car travel route information are used as the trips matched by the user.
  • traverse each piece of car travel path information and store the car travel path information whose trajectory coincidence with the user's travel trajectory is greater than the third preset threshold, and after calculating all the car travel path information and travel trajectory After the trajectory coincidence degree, find the car travel route information with the highest trajectory coincidence degree.
  • the user When the user has more than one public transportation line successfully matched, the user is randomly assigned to one of the public transportation lines, and through iterative calculation, the expanded bus card data, rail transit card data and/or car payment data are used as Target data, adjust the number of people matched to each travel mode and each vehicle. After meeting the target data, select a certain proportion of users who cannot be matched as rail transit ground station side gates or coin-operated passengers on buses , and the rest of the users are classified as private car travel. For example: there are 100 individuals after the sample expansion, 20 of them are matched to public transportation, and the remaining 80 are allocated to various travel modes according to the preset ratio, for example, 10 people take the bus, 5 people take the subway, The remaining 65 people think that they travel by private car.
  • a discrete choice model is established for individual users, and the departure and destination facilities in line with the actual situation are selected through the discrete choice model to generate a resident travel chain.
  • the departure and destination facilities include residences, companies, etc.
  • Obtain the matching results based on the user's time, the location attributes of the departure and arrival places, use long-term data to analyze the user's travel behavior, and obtain travel characteristics, such as long-term stay at home (unemployed/home work, etc.), daily regular round trips (ordinary office workers) ), regular daily round-trips with multiple trips (business, public relations people) in between, as well as the actual travel mode, travel time period, etc., and then perform cluster analysis on different types of users based on the travel characteristics to obtain their travel purpose.
  • the feature function is used to infer the user's travel purpose, because the specific content related to the user's travel purpose obtained through cluster analysis is prior art, and will not be described here.
  • the users who are not successfully matched to one public transportation line are excluded, and are not used as the analysis object.
  • Users who are successfully matched to multiple public transportation lines, through the expanded bus card data and rail transit card data Adjust the matching again to finally obtain a more accurate matching result, based on which a more accurate travel chain is generated.
  • the carpool query method based on the knowledge graph and the resident travel chain of the present invention includes:
  • the resident travel chain is based on the above-mentioned multiple-based It is constructed by a resident travel chain generation method fused with source data; the travel position is read from the travel chain; when the travel position is public transport, the vehicle number of the travel position and the query user ID are obtained to take the The start and end points of the train number, based on the train number and the start and end points, obtain the shared passengers of the query user ID from the knowledge graph.
  • querying the user ID refers to a known query basis, through the above steps, the person sharing the ride with the user ID is found out.
  • the query time range refers to the people who ride with the query user ID within the query time range.
  • Obtain the travel chain within the query time range of the query user ID that is, obtain the activity chain of the user ID within a certain time range, including the travel time and travel location.
  • the travel location can be divided into public transportation and urban facilities.
  • the travel location is a city facility
  • the stay time of the query user ID in the city facility is obtained, and then the personnel in the city facility within the stay time are screened out, that is, the same as the query User IDs are people who may be in the same space at the same time.
  • the knowledge map can be constructed based on the resident travel chain.
  • the entities in the knowledge graph database correspond to individuals, vehicles, and urban facilities in the travel chain model; the relationships in the knowledge graph database correspond to the connections between each link in the individual travel chain model in the travel chain model.
  • Figure 5 is a schematic diagram of the overall structure of graph database entities and relationships. The corresponding relationship between the content and requirements of each main module of the travel chain model (ABM) and the knowledge map (KG) is shown in the following table:
  • the knowledge graph includes user entities, stay location entities, non-traffic facility entities, train number entities, line entities, vehicle entities, and traffic infrastructure entities.
  • the user entities correspond to individual users of the resident travel chain, and the stay location
  • the entity and the non-transport facility entity correspond to the activities of the resident travel chain
  • the train number entity, the line entity, the vehicle entity, and the transportation infrastructure entity correspond to the travel chain of the resident travel chain.
  • Figure 6 shows an epidemic prevention analysis system based on the knowledge map and the sharing query method of the resident travel chain , including a knowledge map building module, a retrieval requirement input module, a data import module, an activity chain acquisition module, and an epidemic prevention analysis module.
  • the knowledge map building module is used to construct the knowledge map through basic data and user travel chain information structure.
  • the basic data includes facility data, public transport train lines, and public transport vehicle information. The structure of transportation facilities.
  • the data import module is used to import user information, activity time, activity location, and traffic mode driving route into the knowledge map after data integration, so that the knowledge map has actual user travel information content.
  • the retrieval demand input module is used to receive inquiries after a case appears, obtain the user ID of the inquiry/query and the query time range, and input them into the knowledge graph.
  • the activity chain acquisition module that is, the travel chain acquisition module, is used to retrieve the queried user ID, obtain its travel location and all user activity chains, exclude data within the non-query time range, and obtain a valid set of user activity chains.
  • the epidemic prevention analysis module is used to sequentially read and query each travel location of the user from the effective user activity chain set, and for each travel location, obtain its location type, and if it is a public transportation vehicle, obtain the starting and ending points of the train number, and obtain The relevant personnel at the start and end of the train will be used as the target personnel, and medical observation will be carried out on the target personnel. If it is a city facility such as a station, the facility and the user's stay time in the facility will be read, and the relevant personnel will be obtained as the target personnel. Conduct medical observation and monitor the facilities for epidemic prevention.
  • the user's mobile phone records can be used to extract all travel information during the incubation period of the disease from the database, through which all passengers or persons in the same city facility at the same time can be extracted As the object of infection screening and the key protection object of epidemic prevention.
  • an individual travel chain is constructed based on multi-source data fusion, combined with traffic travel surveys, combined with existing travel chains for verification, and sample expansion is performed according to the results of traffic travel surveys; Vehicles, using software such as MATSim for multi-modal traffic simulation.
  • the embodiment of the present invention is applied to a resident travel chain generation device based on multi-source data fusion, including:
  • Mobile phone signaling data analysis module which is used to obtain mobile phone signaling data, perform skip data cleaning processing, drift position confirmation processing, different operators and user identification processing and dwell time analysis processing on the mobile phone signaling data, and obtain user travel information, wherein, the travel information includes the residence point, residence time and travel trajectory; based on the travel information, the user's itinerary is cut, and the starting and ending points of the user's itinerary, the time corresponding to the starting and ending point of the itinerary, and the itinerary trajectory are obtained; the traversal For each itinerary of the user, the travel mode is matched for each itinerary of the user, and the travel mode identification of each itinerary is obtained;
  • a public transportation data matching module which is used to spatially match the itinerary trajectory of the public transportation itinerary with the bus line trajectory when the user's itinerary includes a public transportation itinerary, and then based on the starting and ending points of the itinerary and the itinerary of the public transportation itinerary At the time corresponding to the start-destination point, the bus travel itinerary and the bus number information are spatiotemporally matched to generate the first matching result; Carry out space-time matching on the line trajectory, carry out space-time matching with the start-end point of the itinerary of the track traffic travel itinerary, the corresponding time of the start-end point of the itinerary and the rail transit train number information, and generate the second matching result; when the user's itinerary includes a car travel itinerary, Based on the itinerary trajectory of the car travel itinerary, the start and end points of the itinerary, and the corresponding time of the start and end points of the itinerary, the time and space matching of the car travel itinerary and the cars in the preset database is performed to generate a third matching result;
  • a travel behavior integration and modeling module which is used to obtain the number of trains taken by the user for each trip according to the first matching result, the second matching result, and the third matching result, combined with the start and end of each trip of the user point to generate the user's travel chain.
  • the device for generating travel chains based on multi-source data fusion in the present invention has the same beneficial effects as the above-mentioned method for generating travel chains based on multi-source data fusion, which will not be repeated here.
  • the embodiment of the present invention also proposes a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is read and run by a processor, the above-mentioned multi-source data fusion based Resident travel chain generation method or shared ride query method based on knowledge graph and resident travel chain.
  • the beneficial effect of the computer-readable storage medium of the present invention is consistent with the aforementioned method for generating travel chains based on multi-source data fusion, and will not be described here.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种基于多源数据融合的居民出行链生成方法及共乘查询方法，居民出行链生成方法包括：对手机信令数据执行相关处理得到用户的出行信息；基于出行信息切割用户行程；对用户的每个行程进行出行方式匹配；将公交出行行程与公交线路轨迹、公交车次信息进行时空匹配，生成第一匹配结果；将轨迹交通出行行程与轨道交通线路轨迹、轨道交通车次信息进行时空匹配，生成第二匹配结果；将小汽车出行行程与预置数据库中的小汽车进行时空匹配，生成第三匹配结果；根据匹配结果，得到用户每个行程的乘坐车次，结合用户每个行程的行程起讫点，生成用户的出行链，从而实现人-公共交通车辆的联合建模，获得更精准的出行分析结果。

Description

基于多源数据融合的居民出行链生成方法及共乘查询方法 技术领域
本发明涉及交通数据处理技术领域，具体涉及一种基于多源数据融合的居民出行链生成方法及共乘查询方法。
背景技术
早期的出行链模型研究是以出行调查为基础的。随着交通大数据平台的成熟，涌现了大量基于手机信令的城市居民出行的研究。2017年，大连理工大学的戴宇心进行了基于LTE信令数据的移动定位算法研究，而2018年，浙江大学的蔡正义博士基于手机信令数据，对杭州的手机信令数据进行了处理并基于其个体数据进行了建模，更加准确的分析了各区域的交通情况。同年，西南交通大学的卢泰宇基于手机信令数据分析了不同交通方式的识别敏感度。手机信令数据可以做到持续长期大量供给，但手机信令数据仅有轨迹，对出行方式、车辆、具体出发地目的地等分析不足，若仅仅使用手机信令数据建立出行链模型，则因为信息类型过于单一，模型准确性不足。
发明内容
本发明解决的问题是现有出行链模型仅使用手机信令数据建立，信息类型单一，模型准确性不足。
本发明提出一种基于多源数据融合的居民出行链生成方法，包括：
获取手机信令数据，对所述手机信令数据执行跳跃数据清理处理、漂移位置确认处理、不同运营商同用户识别处理及驻留时间分析处理，得到用户的出行信息，其中，所述出行信息包含驻留点、驻留时间和出行轨迹；
基于所述出行信息切割用户行程，获得用户所有行程的行程起讫点、行程起讫点对应时刻及行程轨迹；
遍历用户的每个行程，对用户的每个行程进行出行方式匹配，得到每个行程的出行方式标识；
当所述用户行程包含公交出行行程时，将所述公交出行行程的行程轨迹与公交线路轨迹进行空间匹配，再基于所述公交出行行程的行程起讫点与行 程起讫点对应时刻，将所述公交出行行程与公交车次信息进行时空匹配，生成第一匹配结果；
当所述用户行程包含轨迹交通出行行程时，将所述轨迹交通出行行程的行程轨迹与轨道交通线路轨迹进行空间匹配、将所述轨迹交通出行行程的行程起讫点、行程起讫点对应时刻与轨道交通车次信息进行时空匹配，生成第二匹配结果；
当所述用户行程包含小汽车出行行程时，基于所述小汽车出行行程的行程轨迹、行程起讫点、行程起讫点对应时刻，将所述小汽车出行行程与预置数据库中的小汽车进行时空匹配，生成第三匹配结果；
根据所述第一匹配结果、所述第二匹配结果、所述第三匹配结果，得到用户每个行程的乘坐车次，结合用户每个行程的行程起讫点，生成用户的出行链。
可选地，所述将所述公交出行行程的行程轨迹与公交线路轨迹进行空间匹配包括：
将所述公交出行行程的行程轨迹拟合到道路上，获得所述行程轨迹对应的第一路径集合及其路径顺序；
获取公交线路轨迹对应的第二路径集合及其路径顺序，将所述第一路径集合与所述第二路径集合按照路径顺序进行全路径匹配，其中，所述全路径匹配指将所述第一路径集合中的所有路径按照路径顺序与所述第二路径集合进行匹配；
判断是否存在至少一条与所述第一路径集合全路径匹配的公交线路轨迹；
若是，则将与所述第一路径集合全路径匹配的公交线路轨迹作为所述公交出行行程对应的候选公交线路轨迹；
若否，则将所述第一路径集合拆分成至少两个路径子集，其中，当所述路径子集包含两个或两个以上的路径时，所述两个或两个以上的路径均为路径顺序相邻的路径；将至少两个所述路径子集分别与所述第二路径集合按照路径顺序进行全路径匹配；当每个所述路径子集均匹配到至少一条公交线路轨迹时，判定发生换乘，将所述公交出行行程的行程轨迹按照各所述路径子 集进行行程分割，得到与各所述路径子集对应的行程，将各所述路径子集匹配的公交线路轨迹分别作为相应行程对应的候选公交线路轨迹。
可选地，所述基于所述公交出行行程的行程起讫点与行程起讫点对应时刻，将所述公交出行行程与公交车次信息进行时空匹配，生成第一匹配结果包括：
获得所述候选公交线路轨迹对应的公交车次信息；
将与所述候选公交线路轨迹对应行程的行程起讫点匹配至所述候选公交线路轨迹，得到与所述行程起讫点对应的起/讫公交站；
从所述公交车次信息中，获取所有公交车次到达所述起/讫公交站的到站时刻，基于所述行程起讫点对应时刻及所述到站时刻，分别计算所述行程与各所述公交车次的时间匹配度；
当所述行程与所有公交车次的时间匹配度均小于第二预设阈值时，判定匹配失败；
当所述行程与至少一个公交车次的时间匹配度大于或等于所述第二预设阈值时，将与所述行程时间匹配度最大的公交车次作为与用户匹配的车次。
可选地，所述将与所述候选公交线路轨迹对应行程时间匹配度最大的公交车次作为与用户匹配的车次之后，还包括：
获取目标车次的公交刷卡数据，根据所述公交刷卡数据统计所述目标车次在所述起/讫公交站的逗留时间对应的第一上车人数或第一下车人数，其中，所述目标车次为与所述候选公交线路轨迹对应行程时间匹配度最大的公交车次；
统计在所述起/讫公交站及所述逗留时间，匹配至所述目标车次的第二上车人数或第二下车人数；
计算所述第一上车人数与所述第二上车人数的差值，或者计算所述第一下车人数与所述第二下车人数的差值；
当所述差值小于预设差值时，保留所述目标车次对应的匹配结果；
当所述差值大于或等于所述预设差值时，基于所述差值更新所述目标车次对应的匹配结果，根据更新后的所述目标车次对应的匹配结果，更新用户匹配的车次。
可选地，所述生成用户的出行链之后，还包括：
获取预设时间内与第一用户乘坐车次相同的第二用户；
获取所述第一用户和所述第二用户共同的乘坐车次的空座数量及所述第二用户的出行特征参数，基于第一预设公式计算所述第一用户和所述第二用户的共乘概率，其中，所述第一预设公式包括：
P＝P(α)·f(x)/∑f(x)，
Figure PCTCN2022070988-appb-000001
其中，P指共乘概率，x为所述乘坐车次的空座数量，M 0、σ +、σ -为正实数的预设值，P(α)为出行特征参数，f(x)为上车概率，f(x)/∑f(x)为对分布概率的归一化处理。
可选地，所述基于所述小汽车出行行程的行程轨迹、行程起讫点、行程起讫点对应时刻，将所述小汽车出行行程与预置数据库中的小汽车进行时空匹配，生成第三匹配结果包括：
遍历所述预置数据库中每条小汽车出行路径信息，获取所述小汽车出行的起终点位置，结合所述行程起讫点，判断所述小汽车出行行程与所述小汽车出行路径信息的起终点位置是否均满足预设空间误差；
若是，则根据所述行程起讫点对应时刻，判断所述小汽车出行行程与所述小汽车出行路径信息的起终点时间是否均满足预设时间误差；
若是，则计算所述小汽车出行行程与所述小汽车出行路径信息的轨迹重合度，判断所述轨迹重合度是否大于第三预设阈值；
若是，则确定所述轨迹重合度最高的小汽车出行路径信息，将所述轨迹重合度最高的小汽车出行路径信息对应的车辆作为与用户匹配的车次。
可选地，所述根据所述第一匹配结果、所述第二匹配结果、所述第三匹 配结果，得到用户每个行程的乘坐车次，结合用户每个行程的行程起讫点，生成用户的出行链包括：
当用户没有任何公共交通线路匹配时，判定该用户使用私家车出行；
当用户仅有一条公共交通线路匹配成功时，判定该用户使用私家车或相应的公共交通出行，其中，所述公共交通出行包含公交出行、轨迹交通出行及小汽车出行；
当用户有多于一条公共交通线路匹配成功时，将用户随机分配到其中一个车次上，通过迭代计算，以扩样后的公交刷卡数据、轨道交通刷卡数据和/或小汽车付费数据为目标数据，调整匹配至各出行方式及各车辆的人数，在符合所述目标数据后，从未能匹配的用户中抽取一定比例的用户作为轨道交通地面车站边门入闸或公交车上投币乘客，其余用户归入私家车出行；
根据获得的匹配结果，结合出行时间、地块，对用户个体建立离散选择模型，通过所述离散选择模型选择出发地和目的地的，以生成居民出行链。
可选地，所述对用户的每个行程进行出行方式匹配，得到每个行程的出行方式标识包括：
遍历每个行程，判断所述行程的所述行程轨迹的基站归属与轨道交通基站是否匹配；
若是，则用户的出行方式为轨道交通出行；
若否，则获取所述行程起讫点间的实际旅行时间，并通过路径规划API获取所述行程起讫点的公交时间和小汽车驾车时间，基于第二预设公式计算所述行程为公交出行行程、小汽车出行行程的概率，其中，所述第二预设公式包括：
Figure PCTCN2022070988-appb-000002
Figure PCTCN2022070988-appb-000003
其中，P(公交)指所述行程为公交出行行程的概率，P(驾车)指所述行 程为小汽车出行行程的概率，T 手机指所述行程起讫点间的实际旅行时间，T 驾车指所述小汽车驾车时间，T 公交指所述公交时间。
本发明还提出一种基于知识图谱和居民出行链的共乘查询方法，包括：
获取查询用户和查询时间范围，在基于居民出行链生成的知识图谱中获取所述查询用户在所述查询时间范围内的出行链，其中，所述居民出行链基于如上所述的基于多源数据融合的居民出行链生成方法构建而成，所述知识图谱包含用户实体、停留位置实体、非交通设施实体、车次实体、线路实体、车辆实体、交通基础设施实体，所述用户实体与所述居民出行链的用户个体对应，所述停留位置实体、所述非交通设施实体与所述居民出行链的活动对应，所述车次实体、所述线路实体、所述车辆实体、所述交通基础设施实体与所述居民出行链的出行链对应；
从所述出行链中读取出行位置；
当所述出行位置为公共交通工具时，获取所述出行位置的车次和所述查询用户乘坐该车次的起终点，基于所述车次和所述起终点从所述知识图谱中获取所述查询用户的共乘人员。
本发明还提出一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器读取并运行时，实现如上所述的基于多源数据融合的居民出行链生成方法，或如上所述的基于知识图谱和居民出行链的共乘查询方法。
本发明的有益效果为：
1)通过多源交通数据融合建立出行链模型。本发明通过对多源数据的融合，充分发挥了各类数据的优势，从而获得信息量更丰富的可以长期优化的出行链模型。
2)通过基于手机信令数据对用户每个行程进行出行方式粗匹配，再将手机信令数据及公共交通运营数据、GPS数据等多源数据融合，对用户每个行程进行公共交通线路、公共交通车辆的细匹配，弥补了手机信令数据误差大的缺陷，同时分析出用户个体的出行方式及搭乘的具体车辆，将人、公共交通车辆综合考虑，实现了人-公共交通车辆的联合建模，一方面，因公共 交通车辆有公交刷卡数据、轨道交通AFC数据等实际数据可对匹配结果进行校核，可保证一定的匹配精度，另一方面，将用户每个行程匹配到具体的车辆，可扩大出行链应用场景及适应范围，如基于出行链获得共乘情况，为疫情防控、案件侦破、出行特征分析提供基础，有利于本发明的推广应用。
附图说明
图1为本发明基于多源数据融合的居民出行链生成方法一实施例示意图；
图2为本发明基于多源数据融合的居民出行链生成方法另一实施例示意图；
图3为本发明基于多源数据融合的居民出行链生成方法中确定用户出行方式的一实施例示意图；
图4为本发明基于多源数据融合的居民出行链生成方法中用户出行轨迹与公交数据匹配的一实施例示意图；
图5为本发明基于知识图谱和居民出行链的共乘查询方法中知识图谱图数据库实体和关系整体结构实施例示意图；
图6为本发明基于知识图谱和居民出行链的共乘查询方法一实施例示意图。
具体实施方式
为使本发明的上述目的、特征和优点能够更为明显易懂，下面结合附图对本发明的具体实施例做详细的说明。
为便于理解本发明，首先简要介绍现有技术中存在的问题。
1)仅基于手机信令数据建立出行链模型限制较大。手机信令数据仅有轨迹，对出行方式、车辆、具体出发地目的地等分析不足。职住分布、遥感等数据仅能提供宏观数据，对具体出行链仅有指导和辅助作用，因而使用职住分布、遥感数据结合手机信令数据建立出行链模型，对出行方式、车辆、具体出发地目的地等具体信息分析不足。若仅仅使用手机信令数据建立出行链模型，则因为手机信令数据只有轨迹，难以基于手机信令数据分析出行方 式、车辆、具体出发目的地等，而只考虑个体的位置、出行轨迹，不考虑出行方式、车辆等其他交通信息，就交通分析而言，与交通车辆相关的部分未建模，基于此得到的出行链模型准确性不足。同时，仅使用手机信令数据，使用的数据类型单一，手机信令数据带来的准确性偏差难以被弥补。此外，手机信令的点位是基站GPS数据，而一个基站覆盖范围下的出行地点类型较多，在没有其他数据支撑的情况下，无法实现具体分析。
2)现有的出行链模型中，人和车辆是分离的。人的出行链仅关注其活动地点及出行方式，对其具体乘坐的车辆并无建模，对具体车辆、共乘等方面皆无相应信息，这限制了出行链的应用范围。
需要说明的是，下文中的“公共交通”包含公交、轨道交通、出租车及网约车(本文中小汽车出行包含出租车及网约车)。
图1和图2为本发明基于多源数据融合的居民出行链生成方法一实施例示意图。参照图1和图2，所述基于多源数据融合的居民出行链生成方法包括：
步骤S100，获取手机信令数据，对所述手机信令数据执行跳跃数据清理处理、漂移位置确认处理、不同运营商同用户识别处理及驻留时间分析处理，得到用户的出行信息，其中，所述出行信息包含驻留点、驻留时间和出行轨迹。
手机信令数据，是一组由手机信令基站坐标组成的点位数组，手机信令数据可以做到持续长期大量供给。通过手机信令数据，分析相应个体的驻留点、驻留时间和出行轨迹，具体而言，在对手机信令数据执行跳跃数据清理、漂移位置确认、不同运营商同用户识别、驻留时间分析后，可获得相应个体的驻留点、驻留时间和出行轨迹。其中，有关跳跃数据清理/漂移位置确认：由于手机使用基站具有不稳定性，可能出现两种情况：第一种是用户实际没有出行，但由于其处于多个基站之间，因此其定位会在几个基站之间来回跳跃；第二种是用户的一组连续出行点之间有一个不合常理的跳跃点。基于此，需要对这两类点位数据进行跳跃数据清理及漂移位置确认。
有关不同运营商同用户识别：当使用多运营商数据时，会出现一人多号的情况。根据手机信令的点位情形，一人多号的数据会出现行程高度吻合等 特征，根据这些特征，可排除重复数据；同时，若有集成数据中心，则可根据数据库内部不可导出的用户登记信息进行排查，以排除重复的手机信令数据。
基于手机信令数据，获取用户每日的全部位置点位，通过点位的变化关系，如驻留时间、间隔距离等信息，判断点位动静状态。对于无运动的点位，认为是用户在该地点驻留，即用户的驻留点。两个驻留点之间的点位记录即用户的行程轨迹信息。通过对用户点位的分析，将用户的一天的点位切分为一系列的出行活动，并通过点位位置获得每个出行活动的出发地和到达地等行程特征。用户的出行信息，包含驻留点、驻留时间、行程轨迹、行程出发地、到达地等。
由于各个运营商的用户在整个城市中占比不同，不同年龄段的手机占有率不同，因此可对手机信令数据进行统计学扩样，基于扩样后的手机信令数据生成用户的出行信息。
步骤S200，基于所述出行信息切割用户行程，获得用户所有行程的行程起讫点、行程起讫点对应时刻及行程轨迹。
可通过手机信令点位聚类等方式切割用户行程，具体切割方式为现有技术，此处不赘述。通过对各驻留点的驻留时间进行分析，可得到行程起讫点，例如将驻留时间大于预设时长的驻留点作为行程起讫点。确定行程起讫点后，可直接读取行程起讫点对应时刻，即行程起点对应时刻和行程终点对应时刻。
步骤S300，遍历用户的每个行程，对用户的每个行程进行出行方式匹配，得到每个行程的出行方式标识。
所述出行方式包含公交出行、轨道交通出行、小汽车出行。其中，小汽车出行，包含出租车出行、网约车出行等。
用户的出行信息中，可能包含多个行程，这些行程可能均使用同一种出行方式，如均使用公交出行，也可能使用不同出行方式，如行程AB使用公交出行，行程BC使用轨道交通出行。以单个行程为单位匹配用户的出行方式及后续与公共交通线路、车辆的匹配。
可选地，步骤S200包括：
遍历每个行程，判断所述行程的所述行程轨迹的基站归属与轨道交通基 站是否匹配；若是，则用户的出行方式为轨道交通出行；若否，则获取所述行程起讫点间的实际旅行时间，并通过路径规划API获取所述行程起讫点的公交时间和小汽车驾车时间，基于第二预设公式计算所述行程为公交出行行程、小汽车出行行程的概率。
将基于手机信令数据获得的行程起讫点的实际旅行时间记为T 手机，将通过路径规划API获取的行程起讫点间的公交时间记为T 公交，将通过路径规划API获取的行程起讫点间的小汽车驾车时间记为T 驾车，P(公交)指公交出行的概率，P(驾车)指小汽车出行的概率。以第二预设公式计算置信度，第二预设公式为：
Figure PCTCN2022070988-appb-000004
Figure PCTCN2022070988-appb-000005
当P(公交)>P(驾车)时，判定行程为公交出行，当P(公交)<P(驾车)时，判定行程为小汽车出行。当P(公交)＝P(驾车)，判定行程为公交出行和小汽车出行，在后续匹配流程中，进一步精准匹配。
由上述第二预设公式可以看出：
当起讫点间公交时间T 公交大于驾车时间T 驾车，此时为长距离出行：若T 手机/T 驾车大于1而T 手机/T 公交接近1，则用户为公交出行是大概率事件。
当起讫点间公交时间T 公交大于驾车时间T 驾车，此时为长距离出行：T 手机/T 驾车接近1，而T 手机/T 公交小于1。判定小汽车出行为大概率事件。
当起讫点间公交时间约等于驾车时间，此时为短距离出行：较难判别哪种交通方式，置信度接近0.5，可将公交出行和小汽车出行均作为用户的出行方式，在后续基于公共交通车辆运营数据及GPS数据等精确匹配，确定用户最终的出行方式。
在一实施方式中，如图3，首先通过手机信令数据的基站归属判断用户行程是否为地下轨道交通，若是，则基于地下基站的轨迹初步确定用户行程的出行方式和可能的路线，例如，用户的出行方式/交通方式为地铁出行，可能路线为1号线；若用户行程并非地下轨道交通，则通过路径规划API抓取用户行程起讫点的公交时间和驾车时间，并获得用户行程起讫点间的实际旅行实际，通过上述第二预设公式计算出公交出行概率P(公交)和小汽车出行概率P(驾车)，判断P(公交)/P(驾车)是否大于预设值N，若小于预设值N，则可能为小汽车出行，因小汽车无固定线路，因此将行程轨迹与小汽车出行轨迹进行道路级轨迹拟合，确定用户行程的出行方式和可能的路线；若P(公交)/P(驾车)大于预设值N，则可能为公交出行，将公交线路的轨迹点数据，与手机信令的点位数组进行向量上的相似度比较(可使用两个向量的欧氏距离大小表征相似度)，筛选出可能的线路记录，即对出行线路进行初次筛选，在后续基于公共交通车辆运营数据及GPS数据等进行匹配时，在初次筛选的出行线路的基础上，进一步精准筛选。进一步地，根据初次筛选出的多条线路与手机信令点位数组的相似度长度等信息，结合路径规划API返回的推荐线路，判定是否有换乘，若有，则将公交线路轨迹拆分，确定用户行程可能的所有线路。
将手机信令分析的用户行程与公共交通运行数据，通过交通方式、时空的匹配，分析每个用户行程采用的公共交通方式及出行车次。
步骤S400，当所述用户行程包含公交出行行程时，将所述公交出行行程的行程轨迹与公交线路轨迹进行空间匹配，再基于所述公交出行行程的行程起讫点与行程起讫点对应时刻，将所述公交出行行程与公交车次信息进行时空匹配，生成第一匹配结果。
其中，可通过获取公交GPS数据并进行地图匹配获得公交线路轨迹，包含站点位置、出行路径等信息。通过获取公交运营数据以获得公交车次信息， 公交线路的运营数据包含沿线停靠站点、到站时刻、车牌、车次等信息，车次信息包含沿线停靠站点、到站时刻等信息。用于与公交线路轨迹、公交车次信息进行时空匹配的用户出行信息，除了行程轨迹、行程起讫点和行程起讫点对应时刻外，还可包含换乘站点、初次筛选出的线路等。
第一匹配结果，即：是否存在与用户匹配的公交线路和公交车次，以及若是存在，与用户匹配的公交线路和车次。
步骤S500，当所述用户行程包含轨迹交通出行行程时，将所述轨迹交通出行行程的行程轨迹与轨道交通线路轨迹进行空间匹配、将所述轨迹交通出行行程的行程起讫点、行程起讫点对应时刻与轨道交通车次信息进行时空匹配，生成第二匹配结果。
其中，基于手机信令数据，提取轨道交通地下站的旅客数据，并与轨道交通AFC数据进行匹配；基于经历两个以上乘客的地下站手机信令数据，推测各列车抵达各车站的时间，从而获取列车车次；将手机信令数据的用户匹配到地下站个体乘坐的轨道交通线路和车次。对于地上轨道站，结合轨道交通AFC刷卡数据及基于手机信令数据获得的用户出行信息，将用户分配到各车次上，并与公交卡对应。通过长期的分析，用户使用的公交卡和手机信令用户ID的对应关系会进一步明晰，通过两种数据的匹配情况，结合同一张公交卡的全出行链数据，将同一用户的轨道-公交数据进行统合，从而获得用户较准确的出行链。
第二匹配结果，即：是否存在与用户匹配的轨道交通线路和轨道交通车次，以及若是存在，则与用户匹配的轨道交通线路和车次。当存在与用户匹配的轨道交通线路和车次时，使用轨道交通AFC刷卡数据对匹配结果进行校验，或者使用轨道交通AFC刷卡数据与轨道交通ATS列车运行数据结合起来对匹配结果进行校验，以保证匹配结果的准确性。
步骤S600，当所述用户行程包含小汽车出行行程时，基于所述小汽车出行行程的行程轨迹、行程起讫点、行程起讫点对应时刻，将所述小汽车出行行程与预置数据库中的小汽车进行时空匹配，生成第三匹配结果。
其中，预置数据库中存储有小汽车出行的起终点位置、时间、出行路径、载客状态等信息。小汽车包含出租车、网约车，基于出租车/网约车的GPS 轨迹数据，进行地图匹配并获取出租车/网约车全天完整的出行行程。
步骤S700，根据所述第一匹配结果、所述第二匹配结果、所述第三匹配结果，得到用户每个行程的乘坐车次，结合用户每个行程的行程起讫点，生成用户的出行链。
在得到第一匹配结果、第二匹配结果及第三匹配结果后，即可确定用户每个行程是否匹配成功，以及若是成功，则匹配到的线路与车次信息，结合每个行程的行程起讫点，生成用户的出行链。
需说明的是：步骤S700中的乘坐车次，当车辆为公交车或轨道交通车辆时，其具有相对固定的营运班次，乘坐车次包含车辆标识(如车牌号或几号车)及车辆班次，当车辆为小汽车时，乘坐车次包含车辆标识(如车牌号)，一实施方式中，为小汽车的每段行程标号，乘坐车次除包含车辆标识外，还包含行程标号。
通过基于手机信令数据对用户每个行程进行出行方式粗匹配，再将手机信令数据及公共交通运营数据、GPS数据等多源数据融合，对用户每个行程进行公共交通线路、公共交通车辆的细匹配，弥补了手机信令数据误差大的缺陷，同时分析出用户个体的出行方式及搭乘的具体车辆，将人、公共交通车辆综合考虑，实现了人-公共交通车辆的联合建模，一方面，因公共交通车辆有公交刷卡数据、轨道交通AFC数据等实际数据可对匹配结果进行校核，可保证一定的匹配精度，另一方面，将用户每个行程匹配到具体的车辆，可扩大出行链应用场景及适应范围，如基于出行链获得共乘情况，有利于本发明的推广应用。
可选地，步骤S400中所述将所述公交出行行程的行程轨迹与公交线路轨迹进行空间匹配包括：
将公交出行行程的行程轨迹拟合到道路上，获得行程轨迹对应的第一路径集合及其路径顺序。
获取公交线路轨迹对应的第二路径集合及其路径顺序，将第一路径集合与第二路径集合按照路径顺序进行全路径匹配，其中，全路径匹配指将第一路径集合中的所有路径按照路径顺序与第二路径集合进行匹配。其中，一条公交线路轨迹对应一个第二路径集合，获取将要与公交出行行程匹配的公交 线路轨迹对应的第二路径集合及其路径顺序。
例如，第一路径集合有A、B、C、D、E、F等6条路径，其路径顺序为A->B->C->D->E->F，将第一路径集合与第二路径集合进行全路径匹配，即将A、B、C、D、E、F等6条路径与第二路径集合中的路径进行匹配。
判断是否存在至少一条与第一路径集合全路径匹配的公交线路轨迹。一实施方式中，当第二路径集合包含第一路径集合的所有路径，且该所有路径在第二路径集合中的路径顺序与在第一路径集合中的路径顺序相同时，判定该第二路径集合与第一路径集合全路径匹配，即该第二路径集合对应的公交线路轨迹与第一路径集合全路径匹配。
若是，则将与第一路径集合全路径匹配的公交线路轨迹作为公交出行行程对应的候选公交线路轨迹。
若否，则将第一路径集合拆分成至少两个路径子集，其中，当路径子集包含两个或两个以上的路径时，两个或两个以上的路径均为路径顺序相邻的路径；将至少两个路径子集分别与第二路径集合按照路径顺序进行全路径匹配；当每个路径子集均匹配到至少一条公交线路轨迹时，判定发生换乘，将公交出行行程的行程轨迹按照各路径子集进行行程分割，得到与各路径子集对应的行程，将各路径子集匹配的公交线路轨迹分别作为相应行程对应的候选公交线路轨迹。
例如，第一路径集合有A、B、C、D、E、F等6条路径，其路径顺序为A->B->C->D->E->F，当不存在与第一路径集合全路径匹配的公交线路轨迹时，将第一路径集合拆分成“A->B->C”、“D->E->F”两个路径子集，将“A->B->C”、“D->E->F”分别与公交线路轨迹对应的路径集合进行匹配，若“A->B->C”匹配到公交线路轨迹1，“D->E->F”匹配到公交线路轨迹2，则判定发生换乘，即发生从公交线路轨迹1到公交线路轨迹2的换乘。将公交出行行程的行程轨迹按照各路径子集进行行程分割，得到“A->B->C”、“D->E->F”两段行程，将公交线路轨迹1作为“A->B->C”对应的候选公交线路轨迹，将公交线路轨迹2作为“D->E->F”对应的候选公交线路轨迹。
一实施例中，将第一路径集合拆分成两个路径子集，该两个路径子集中均为路径顺序相邻的路径，为便于描述，引入拆分点的概念，例如，将A、 B、C、D、E、F等6条路径拆分成“A->B->C”、“D->E->F”两个路径子集，则D为拆分点。因为将第一路径集合拆分成两个路径子集有多种不同拆分结果，不同拆分点能得到不同的拆分结果，可预先设置多个拆分点，首先，根据第一个拆分点拆分第一路径集合得到两个路径子集，将该两个路径子集分别与第二路径集合按照路径顺序进行全路径匹配，若每个路径子集均匹配到至少一条公交线路轨迹，则将各路径子集匹配的公交线路轨迹分别作为相应行程对应的候选公交线路轨迹，第一路径集合的匹配结束，若有路径子集未匹配到至少一条公交线路轨迹，则换下一个拆分点，直至每个路径子集均匹配到至少一条公交线路轨迹或所有拆分点均匹配完，当所有拆分点都匹配完，且还有路径子集未匹配到公交线路轨迹时，判定公交线路匹配失败，将公交出行行程归入小汽车出行或私家车出行。
可选地，参见图4，步骤S400中所述基于所述公交出行行程的行程起讫点与行程起讫点对应时刻，将所述公交出行行程与公交车次信息进行时空匹配，生成第一匹配结果包括：
步骤S401，获得所述候选公交线路轨迹对应的公交车次信息。
公交的特点在于车辆轨迹确定，但时空重复度高，且发车频率较高。因此需要在时间与空间上对所有可能的线路及车次进行寻找，最终找到符合阈值要求的匹配方案。
当候选公交线路轨迹有多条时，对每条候选公交线路轨迹执行步骤S401-步骤S405。
步骤S402，将与所述候选公交线路轨迹对应行程的行程起讫点匹配至所述候选公交线路轨迹，得到与所述行程起讫点对应的起/讫公交站。
行程起讫点即个体出行起止位置。将行程起点、行程终点分别匹配至候选公交线路轨迹，得到与行程起讫点对应的起/讫公交站，即起公交站和讫公交站。
步骤S403，从所述公交车次信息中，获取所有公交车次到达所述起/讫公交站的到站时刻，基于所述行程起讫点对应时刻及所述到站时刻，分别计算所述行程与各所述公交车次的时间匹配度。
获取候选公交线路轨迹的公交运营数据，读取候选公交线路轨迹所有公 交车次到达起/讫公交站的到站时刻。
计算行程轨迹与候选公交线路轨迹各车次的时间匹配度，可通过用户行程起点对应时刻与各车次到达起公交站的到站时刻，计算用户到达起公交站后的等车时间，基于该等车时间确定时间匹配度，其中，该等车时间越小，时间匹配度越大，相应车次与用户的匹配度越高，等车时间越大，时间匹配度越小，相应车次与用户的匹配度越低。其中，可将用户行程起点对应时刻与各车次到达起公交站的到站时刻的差值作为等车时间。
步骤S404，当所述行程轨迹与所有公交车次的时间匹配度均小于第二预设阈值时，判定匹配失败，判定用户使用私家车出行。
通过时间匹配度，可剔除轨迹与点位符合但相应时间没有车次的公交线路。若行程轨迹与候选公交线路轨迹所有车次的时间匹配度均小于第二预设阈值，说明候选公交线路轨迹所有车次与用户的匹配度都过低，判定匹配失败，判定用户使用私家车出行。
步骤S405，当所述行程轨迹与至少一个公交车次的时间匹配度大于或等于所述第二预设阈值时，将与所述候选公交线路轨迹对应行程时间匹配度最大的公交车次作为与用户匹配的车次。
其中，为便于描述，将与行程轨迹的时间匹配度最大的车次称为目标车次。
可选地，步骤S405之后，还包括：
步骤S406，获取目标车次的公交刷卡数据，根据所述公交刷卡数据统计所述目标车次在所述起/讫公交站的逗留时间对应的第一上车人数或第一下车人数，其中，所述目标车次为与所述候选公交线路轨迹对应行程时间匹配度最大的公交车次。
在执行完步骤S405之后，获得了初步的匹配结果，为保证匹配结果符合实际，通过公交刷卡数据对匹配结果进行校核。
其中，目标车次在起/讫公交站的逗留时间，指从目标车次到达起/讫公交站至离开起/讫公交站的时间，与用户的行程起讫点对应时刻较为接近，用户在所述逗留时间在起/讫公交站上目标车次、下目标车次。
目标车次在所述起/讫公交站的逗留时间对应的第一上车人数，指目标 车次在起公交站的逗留时间内的第一上车人数，第一下车人数指目标车次在讫公交站的逗留时间内的第一下车人数。
步骤S407，统计在所述起/讫公交站及所述逗留时间，匹配至所述目标车次的第二上车人数或第二下车人数。
步骤S408，计算所述第一上车人数与所述第二上车人数的差值，或者计算所述第一下车人数与所述第二下车人数的差值。
步骤S409，当所述差值小于预设差值时，保留所述目标车次对应的匹配结果。
当差值小于预设差值时，说明初步的匹配结果符合实际情况，因而可以保留目标车次对应的匹配结果，即保留目标车次与用户之间的匹配关系，将目标车次及其线路作为与用户匹配的车次和公交线路。
步骤S410，当所述差值大于或等于所述预设差值时，基于所述差值更新所述目标车次对应的匹配结果，根据更新后的所述目标车次对应的匹配结果，更新用户匹配的车次。
当差值大于或等于预设差值时，说明初步的匹配结果不符合实际情况，因而需要对匹配结果进行调整，基于差值更新目标车次对应的匹配结果，具体而言，可将与差值相等数量的用户的出行方式判定为私家车出行，将剩余的用户的匹配结果保留，其中，将判定为私家车出行的用户从需要分析的用户中排除，不做进一步分析。通过公交刷卡数据对匹配结果进行校核，可以保证匹配结果符合实际情况，进而保证建立的居民出行链的准确性。
给出一如图4所示的实施例：
输入用户一次出行的手机信令数据，即用户单个行程的手机信令数据；
读取用户该行程可能乘坐的线路集，输入公交GPS轨迹数据，读取线路集的轨迹数据；
计算用户出行链轨迹(即该此行程的行程轨迹)与公交线路轨迹的重合度；
判断重合度是否大于阈值；
若是，则更新可能的线路集，即：将重合度大于阈值的公交线路轨迹作为用户可能乘坐的线路；若否，则读取另一条公交线路轨迹，返回执行所述 计算用户出行链轨迹与公交线路轨迹的重合度的步骤；
在更新可能的线路集后，读取用户行程起讫点，将起讫点分别匹配至对应线路(即可能的线路集中的线路)的公交站，从公交运营数据中读取对应线路的公交运营数据，读取对应线路所有车次的沿线到站时刻，基于用户行程起讫点对应时刻与所有车次的沿线到站时刻，计算用户到达公交站后的等车时间；
判断等车时间是否小于阈值；
若等车时间大于或等于阈值，则更新可能的线路集(将等车时间大于或等于阈值的车次及线路去除)；
若等车时间小于阈值，则寻找等车时间最少的车次，并根据公交刷卡数据，统计该车次在相应时间在起/讫公交站的上/下车人数，统计同一时间段内匹配至各站点的上车人数与基于公交刷卡数据的上车人数是否一致，若否，则匹配失败，若是，则将等车时间最少的车次及其线路作为与用户匹配的车次和公交线路。
可选地，所述生成用户的出行链之后，还包括：
获取预设时间内与第一用户乘坐车次相同的第二用户；
获取所述第一用户和所述第二用户共同的乘坐车次的空座数量及所述第二用户的出行特征参数，基于第一预设公式计算所述第一用户和所述第二用户的共乘概率，其中，所述第一预设公式包括：
P＝P(α)·f(x)/∑f(x)，
Figure PCTCN2022070988-appb-000006
其中，P指共乘概率，x为所述乘坐车次的空座数量，M 0、σ +、σ -为正实数的预设值，P(α)为出行特征参数，f(x)为上车概率，f(x)/∑f(x) 为对分布概率的归一化处理。
基于上文所述的步骤，通过手机信令数据和公共交通数据将用户出行的点位信息与公共交通线路/车辆进行匹配，得到第一匹配结果、第二匹配结果、第三匹配结果，得到用户的乘坐车次。而通过上述步骤得出的用户的乘坐车次，是用户最有可能搭乘的公共交通线路/车辆，而用户实际上是否乘坐匹配的公共交通车辆前往终点存在一定的概率。
已知第一用户实际乘坐的车次，获取通过上述步骤确定的出行链，进而确定与第一用户乘坐车次相同的第二用户，计算第一用户与第二用户的共乘概率。第一用户和第二用户共同的乘坐车次的空座数量，通过上下车人数获得，其中，上车人数可通过该公共交通的刷卡数或者上车的概率获得，下车的人数可以通过手机信令或者大样本统计的OD获得。
获取空座数量后，即可通过上述第一预设公式，计算共乘概率。
通过将人-车关联，计算出两两用户之间的共乘概率，进而实现人-人关联，有利于建立疫情防控、案件侦破、出行特征分析等的数据基础。
可选地，所述基于所述小汽车出行行程的行程轨迹、行程起讫点、行程起讫点对应时刻，将所述小汽车出行行程与预置数据库中的小汽车进行时空匹配，生成第三匹配结果包括：
步骤S420，遍历所述预置数据库中每条小汽车出行路径信息，获取所述小汽车出行的起终点位置，结合所述行程起讫点，判断所述小汽车出行行程与所述小汽车出行路径信息的起终点位置是否均满足预设空间误差。
步骤S421，若是，即小汽车出行行程与小汽车出行路径信息的起终点位置均满足预设空间误差，则根据所述行程起讫点对应时刻，判断所述小汽车出行行程与所述小汽车出行路径信息的起终点时间是否均满足预设时间误差。
小汽车出行路径信息包含起终点位置和时间，在执行步骤S420-步骤S421时，可直接从小汽车出行路径信息中获取起终点位置和时间，以分别结合用户的行程起讫点、行程起讫点对应时刻，判断起终点位置是否均满足预设空间误差、起终点时间是否均满足预设时间误差。
若用户的行程轨迹与小汽车出行路径信息的起终点位置不均满足预设 空间误差，则直接判定匹配失败，判定用户为私家车出行，不执行步骤S421及其之后的步骤。
步骤S422，若是，即小汽车出行行程与小汽车出行路径信息的起终点时间均满足预设时间误差，则计算所述小汽车出行行程与所述小汽车出行路径信息的轨迹重合度，判断所述轨迹重合度是否大于第三预设阈值。
若用户的行程轨迹与小汽车出行路径信息的起终点位置均满足预设空间误差、起终点时间均满足预设时间误差，则进一步判断轨迹重合度，可选地，获取用户的行程轨迹的路径集合以及小汽车出行路径的路径集合，计算两种不同数据源轨迹的路径集合之间的轨迹重合度。
若小汽车出行行程与小汽车出行路径信息的起终点时间存在任一时间不满足预设时间误差，则判定匹配失败，判定用户为私家车出行。
若轨迹重合度小于或等于第三预设阈值，则判定匹配失败，判定用户为私家车出行。
步骤S423，若是，即若轨迹重合度大于第三预设阈值，则确定所述轨迹重合度最高的小汽车出行路径信息，将所述轨迹重合度最高的小汽车出行路径信息对应的车辆作为与用户匹配的车次。
若轨迹重合度大于第三预设阈值，则在计算完所有的小汽车出行路径信息与行程轨迹的轨迹重合度后，选取轨迹重合度最高的小汽车出行路径信息，该轨迹重合度最高的小汽车出行路径信息中的小汽车车次作为用户匹配的车次。可选地，遍历每条小汽车出行路径信息，将与用户的行程轨迹的轨迹重合度大于第三预设阈值的小汽车出行路径信息存储，在计算完所有的小汽车出行路径信息与行程轨迹的轨迹重合度后，寻找轨迹重合度最高的小汽车出行路径信息。
可选地，所述生成用户的出行链之后，还包括：
当用户没有任何公共交通线路匹配时，判定该用户使用私家车出行，将使用私家车出行的用户从需要分析的用户中排除，不做进一步分析，即不作为居民出行链的分析对象。
当用户仅有一条公共交通线路匹配成功时，判定该用户使用私家车或相应的公共交通出行。
当用户有多于一条公共交通线路匹配成功时，将用户随机分配到其中一条公共交通线路上，通过迭代计算，以扩样后的公交刷卡数据、轨道交通刷卡数据和/或小汽车付费数据为目标数据，调整匹配至各出行方式及各车辆的人数，在符合所述目标数据后，从未能匹配的用户中抽取一定比例的用户作为轨道交通地面车站边门入闸或公交车上投币乘客，其余用户归入私家车出行。例如：扩样后的个体有100个人，其中20个人匹配到公共交通，剩余80人，按照预设比例将该80个人分配到各种出行方式，例如10个人乘坐公交车，5个人乘坐地铁，其余65人认为采用私家车出行。
根据获得的匹配结果，结合出行时间、地块，对用户个体建立离散选择模型，通过所述离散选择模型选择符合实际情况的出发地和目的地的设施，以生成居民出行链。
其中的出发地和目的地的设施，包括住宅、公司等。
获取匹配结果，基于用户时间、出发地到达地的地点属性，使用长期数据对用户的出行行为进行分析，获得出行特征，如长期在家(待业/家中上班等)、每日定时往返(普通上班族)、每日定时往返且中间有多个出行(商务、公关人士)等特征，以及实际出行方式、出行时间段等，再基于出行特征对不同类型的用户进行聚类分析，获取其出行目的的特征函数，推测用户的出行目的，因具体的通过聚类分析获得用户的出行目的的相关内容为现有技术，此处不赘述。
通过对匹配结果进行分类，将未成功匹配到一条公共交通线路的用户排除，不作为分析对象，将成功匹配到多条公共交通线路的用户，通过扩样后的公交刷卡数据和轨道交通刷卡数据再次调整匹配，以最终获得较为准确的匹配结果，基于此生成较为准确的出行链。
一实施例中，本发明基于知识图谱和居民出行链的共乘查询方法包括：
获取查询用户ID和查询时间范围，在基于居民出行链生成的知识图谱中获取所述查询用户ID在所述查询时间范围内的出行链，其中，所述居民出行链基于上所述的基于多源数据融合的居民出行链生成方法构建而成；从所述出行链中读取出行位置；当所述出行位置为公共交通工具时，获取所述出行位置的车次和所述查询用户ID乘坐该车次的起终点，基于所述车次和 所述起终点从所述知识图谱中获取所述查询用户ID的共乘人员。
其中，查询用户ID指已知的查询依据，通过上述步骤，查询出与该用户ID共乘的人员。查询时间范围，指在该查询时间范围内，与查询用户ID共乘的人员。
获取查询用户ID查询时间范围内的出行链，即获取该用户ID在一定时间范围内的活动链，包含了其出行时间和出行位置。
出行位置可分为公共交通工具和城市设施，当出行位置为城市设施时，获取查询用户ID在该城市设施的逗留时间，进而筛选出在该逗留时间内该城市设施内的人员，即与查询用户ID可能在同一时间处于同一空间的人员。
其中，在通过上述基于多源数据融合的居民出行链生成方法获得居民出行链后，可基于居民出行链构建知识图谱。知识图谱图数据库的实体对应出行链模型中的的个体、交通工具、城市设施；知识图谱图数据库中的关系对应出行链模型中，个体出行全出行中每一环之间的连接。如图5为图数据库实体和关系整体结构示意图。出行链模型(ABM)各主要模块的内容、需求与知识图谱(KG)对应关系见下表：
Figure PCTCN2022070988-appb-000007
Figure PCTCN2022070988-appb-000008
其中，知识图谱包含用户实体、停留位置实体、非交通设施实体、车次实体、线路实体、车辆实体、交通基础设施实体，所述用户实体与所述居民出行链的用户个体对应，所述停留位置实体、所述非交通设施实体与所述居民出行链的活动对应，所述车次实体、所述线路实体、所述车辆实体、所述交通基础设施实体与所述居民出行链的出行链对应。
如图6所示，给出将基于知识图谱和居民出行链的共乘查询方法应用于防疫分析，如图6示出了一种基于知识图谱和居民出行链的共乘查询方法的防疫分析系统，包含知识图谱构建模块、检索需求输入模块、数据导入模块、活动链获取模块以及防疫分析模块。其中，知识图谱构建模块用于通过基础数据和用户出行链信息结构构建知识图谱，基础数据包含设施数据、公共交通车次线路、公共交通车辆信息，用户出行链信息结构包括用户结构、城市设施结构和交通设施结构。数据导入模块，用于将用户信息、活动时间活动地点、交通模式行驶路线经数据整合后，导入知识图谱，使知识图谱中具有实际的用户出行信息内容。检索需求输入模块，用于在出现病例后，接收询问，获取询问/查询的用户ID和查询时间范围，将其输入知识图谱。活动链获取模块，即出行链获取模块，用于检索查询的用户ID，获取其出行位置以及用户全部活动链，排除非查询时间范围内的数据，获得有效的用户活动链集合。防疫分析模块，用于从有效的用户活动链集合中依次读取查询用户的各个出行位置，对于每个出行位置，获取其位置类型，若其为公共交通工具，则获取车次乘坐起终点，获取该车次该起终点的相关人员作为目标人员，对目标人员进行医学观察，若其为车站等城市设施，则读取设施及用户在该设施的逗留时间，获取相关人员作为目标人员，对目标人员进行医学观察， 并对设施进行防疫监控。
通过分析感染人员全链条的出行活动轨迹和交通方式，可以使用用户的手机记录从数据库中提取疾病潜伏期内全部出行信息，通过这些信息可提取全部同乘人员或者同时间在同一城市设施内的人员作为感染排查对象和防疫的重点保护对象。输入确诊案例的手机信令ID、感染起始时间(必选)以及出行时间、方式与车次(可选信息)信息，通过时空匹配，在知识图谱中检索出符合时间要求采用公共交通方式的感染出行链，并输出与该用户相关的所有共乘人员的分析结果。通过对用户全出行链的建模分析，将有利于对用户及用户活动区域进行精准检测和管理，挖掘高风险人群的活动规律，通过基于城市设施、交通工具的图论算法分析，研判高风险的POI和交通车次，有利于对重点POI和公共交通进行预防。
在另一实施例中，基于多源数据融合构建个体出行链，结合交通出行调查，结合现有的出行链进行校核，并根据交通出行调查的结果进行扩样；根据扩样结果构建个体和车辆，使用MATSim等软件进行多模式交通仿真。
本发明实施例应用于基于多源数据融合的居民出行链生成装置，包括：
手机信令数据分析模块，其用于获取手机信令数据，对所述手机信令数据执行跳跃数据清理处理、漂移位置确认处理、不同运营商同用户识别处理及驻留时间分析处理，得到用户的出行信息，其中，所述出行信息包含驻留点、驻留时间和出行轨迹；基于所述出行信息切割用户行程，获得用户所有行程的行程起讫点、行程起讫点对应时刻及行程轨迹；遍历用户的每个行程，对用户的每个行程进行出行方式匹配，得到每个行程的出行方式标识；
公共交通数据匹配模块，其用于当所述用户行程包含公交出行行程时，将所述公交出行行程的行程轨迹与公交线路轨迹进行空间匹配，再基于所述公交出行行程的行程起讫点与行程起讫点对应时刻，将所述公交出行行程与公交车次信息进行时空匹配，生成第一匹配结果；当所述用户行程包含轨迹交通出行行程时，将所述轨迹交通出行行程的行程轨迹与轨道交通线路轨迹进行空间匹配、将所述轨迹交通出行行程的行程起讫点、行程起讫点对应时刻与轨道交通车次信息进行时空匹配，生成第二匹配结果；当所述用户行程包含小汽车出行行程时，基于所述小汽车出行行程的行程轨迹、行程起讫点、 行程起讫点对应时刻，将所述小汽车出行行程与预置数据库中的小汽车进行时空匹配，生成第三匹配结果；
出行行为整合、建模模块，其用于根据所述第一匹配结果、所述第二匹配结果、所述第三匹配结果，得到用户每个行程的乘坐车次，结合用户每个行程的行程起讫点，生成用户的出行链。
本发明基于多源数据融合的居民出行链生成装置相对于现有技术所具有的有益效果与上述基于多源数据融合的居民出行链生成方法一致，此处不赘述。
本发明实施例还提出一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器读取并运行时，实现如上所述的基于多源数据融合的居民出行链生成方法或基于知识图谱和居民出行链的共乘查询方法。
本发明计算机可读存储介质相对于现有技术所具有的有益效果与上述基于多源数据融合的居民出行链生成方法一致，此处不赘述。
虽然本发明公开披露如上，但本发明公开的保护范围并非仅限于此。本领域技术人员在不脱离本发明公开的精神和范围的前提下，可进行各种变更与修改，这些变更与修改均将落入本发明的保护范围。

Claims (10)

  1. 一种基于多源数据融合的居民出行链生成方法，其特征在于，包括：
    获取手机信令数据，对所述手机信令数据执行跳跃数据清理处理、漂移位置确认处理、不同运营商同用户识别处理及驻留时间分析处理，得到用户的出行信息，其中，所述出行信息包含驻留点、驻留时间和出行轨迹；
    基于所述出行信息切割用户行程，获得用户所有行程的行程起讫点、行程起讫点对应时刻及行程轨迹；
    遍历用户的每个行程，对用户的每个行程进行出行方式匹配，得到每个行程的出行方式标识；
    当所述用户行程包含公交出行行程时，将所述公交出行行程的行程轨迹与公交线路轨迹进行空间匹配，再基于所述公交出行行程的行程起讫点与行程起讫点对应时刻，将所述公交出行行程与公交车次信息进行时空匹配，生成第一匹配结果；
    当所述用户行程包含轨迹交通出行行程时，将所述轨迹交通出行行程的行程轨迹与轨道交通线路轨迹进行空间匹配、将所述轨迹交通出行行程的行程起讫点、行程起讫点对应时刻与轨道交通车次信息进行时空匹配，生成第二匹配结果；
    当所述用户行程包含小汽车出行行程时，基于所述小汽车出行行程的行程轨迹、行程起讫点、行程起讫点对应时刻，将所述小汽车出行行程与预置数据库中的小汽车进行时空匹配，生成第三匹配结果；
    根据所述第一匹配结果、所述第二匹配结果、所述第三匹配结果，得到用户每个行程的乘坐车次，结合用户每个行程的行程起讫点，生成用户的出行链。
  2. 如权利要求1所述的基于多源数据融合的居民出行链生成方法，其特征在于，所述将所述公交出行行程的行程轨迹与公交线路轨迹进行空间匹配包括：
    将所述公交出行行程的行程轨迹拟合到道路上，获得所述行程轨迹对应的第一路径集合及其路径顺序；
    获取公交线路轨迹对应的第二路径集合及其路径顺序，将所述第一路径 集合与所述第二路径集合按照路径顺序进行全路径匹配，其中，所述全路径匹配指将所述第一路径集合中的所有路径按照路径顺序与所述第二路径集合进行匹配；
    判断是否存在至少一条与所述第一路径集合全路径匹配的公交线路轨迹；
    若是，则将与所述第一路径集合全路径匹配的公交线路轨迹作为所述公交出行行程对应的候选公交线路轨迹；
    若否，则将所述第一路径集合拆分成至少两个路径子集，其中，当所述路径子集包含两个或两个以上的路径时，所述两个或两个以上的路径均为路径顺序相邻的路径；将至少两个所述路径子集分别与所述第二路径集合按照路径顺序进行全路径匹配；当每个所述路径子集均匹配到至少一条公交线路轨迹时，判定发生换乘，将所述公交出行行程的行程轨迹按照各所述路径子集进行行程分割，得到与各所述路径子集对应的行程，将各所述路径子集匹配的公交线路轨迹分别作为相应行程对应的候选公交线路轨迹。
  3. 如权利要求2所述的基于多源数据融合的居民出行链生成方法，其特征在于，所述基于所述公交出行行程的行程起讫点与行程起讫点对应时刻，将所述公交出行行程与公交车次信息进行时空匹配，生成第一匹配结果包括：
    获得所述候选公交线路轨迹对应的公交车次信息；
    将与所述候选公交线路轨迹对应行程的行程起讫点匹配至所述候选公交线路轨迹，得到与所述行程起讫点对应的起/讫公交站；
    从所述公交车次信息中，获取所有公交车次到达所述起/讫公交站的到站时刻，基于所述行程起讫点对应时刻及所述到站时刻，分别计算所述候选公交线路轨迹对应行程与各所述公交车次的时间匹配度；
    当所述候选公交线路轨迹对应行程与所有公交车次的时间匹配度均小于第二预设阈值时，判定匹配失败；
    当所述候选公交线路轨迹对应行程与至少一个公交车次的时间匹配度大于或等于所述第二预设阈值时，将与所述候选公交线路轨迹对应行程时间匹配度最大的公交车次作为与用户匹配的车次。
  4. 如权利要求3所述的基于多源数据融合的居民出行链生成方法，其特 征在于，所述将与所述候选公交线路轨迹对应行程时间匹配度最大的公交车次作为与用户匹配的车次之后，还包括：
    获取目标车次的公交刷卡数据，根据所述公交刷卡数据统计所述目标车次在所述起/讫公交站的逗留时间对应的第一上车人数或第一下车人数，其中，所述目标车次为与所述候选公交线路轨迹对应行程时间匹配度最大的公交车次；
    统计在所述起/讫公交站及所述逗留时间，匹配至所述目标车次的第二上车人数或第二下车人数；
    计算所述第一上车人数与所述第二上车人数的差值，或者计算所述第一下车人数与所述第二下车人数的差值；
    当所述差值小于预设差值时，保留所述目标车次对应的匹配结果；
    当所述差值大于或等于所述预设差值时，基于所述差值更新所述目标车次对应的匹配结果，根据更新后的所述目标车次对应的匹配结果，更新用户匹配的车次。
  5. 如权利要求1所述的基于多源数据融合的居民出行链生成方法，其特征在于，所述生成用户的出行链之后，还包括：
    获取预设时间内与第一用户乘坐车次相同的第二用户；
    获取所述第一用户和所述第二用户共同的乘坐车次的空座数量及所述第二用户的出行特征参数，基于第一预设公式计算所述第一用户和所述第二用户的共乘概率，其中，所述第一预设公式包括：
    P＝P(α)·f(x)/∑f(x)，
    Figure PCTCN2022070988-appb-100001
    其中，P指共乘概率，x为所述乘坐车次的空座数量，M 0、σ +、σ - 为正实数的预设值，P(α)为出行特征参数，f(x)为上车概率，f(x)/∑f(x)为对分布概率的归一化处理。
  6. 如权利要求1所述的基于多源数据融合的居民出行链生成方法，其特征在于，所述基于所述小汽车出行行程的行程轨迹、行程起讫点、行程起讫点对应时刻，将所述小汽车出行行程与预置数据库中的小汽车进行时空匹配，生成第三匹配结果包括：
    遍历所述预置数据库中每条小汽车出行路径信息，获取所述小汽车出行的起终点位置，结合所述行程起讫点，判断所述小汽车出行行程与所述小汽车出行路径信息的起终点位置是否均满足预设空间误差；
    若是，则根据所述行程起讫点对应时刻，判断所述小汽车出行行程与所述小汽车出行路径信息的起终点时间是否均满足预设时间误差；
    若是，则计算所述小汽车出行行程与所述小汽车出行路径信息的轨迹重合度，判断所述轨迹重合度是否大于第三预设阈值；
    若是，则确定所述轨迹重合度最高的小汽车出行路径信息，将所述轨迹重合度最高的小汽车出行路径信息对应的车辆作为与用户匹配的车次。
  7. 如权利要求1所述的基于多源数据融合的居民出行链生成方法，其特征在于，所述根据所述第一匹配结果、所述第二匹配结果、所述第三匹配结果，得到用户每个行程的乘坐车次，结合用户每个行程的行程起讫点，生成用户的出行链包括：
    当用户没有任何公共交通线路匹配时，判定该用户使用私家车出行；
    当用户仅有一条公共交通线路匹配成功时，判定该用户使用私家车或相应的公共交通出行，其中，所述公共交通出行包含公交出行、轨迹交通出行及小汽车出行；
    当用户有多于一条公共交通线路匹配成功时，将用户随机分配到其中一个车次上，通过迭代计算，以扩样后的公交刷卡数据、轨道交通刷卡数据和/或小汽车付费数据为目标数据，调整匹配至各出行方式及各车辆的人数，在符合所述目标数据后，从未能匹配的用户中抽取一定比例的用户作为轨道交通地面车站边门入闸或公交车上投币乘客，其余用户归入私家车出行；
    根据获得的匹配结果，结合出行时间、地块，对用户个体建立离散选择 模型，通过所述离散选择模型选择出发地和目的地的，以生成居民出行链。
  8. 如权利要求1所述的基于多源数据融合的居民出行链生成方法，其特征在于，所述对用户的每个行程进行出行方式匹配，得到每个行程的出行方式标识包括：
    遍历每个行程，判断所述行程的所述行程轨迹的基站归属与轨道交通基站是否匹配；
    若是，则用户的出行方式为轨道交通出行；
    若否，则获取所述行程起讫点间的实际旅行时间，并通过路径规划API获取所述行程起讫点的公交时间和小汽车驾车时间，基于第二预设公式计算所述行程为公交出行行程、小汽车出行行程的概率，其中，所述第二预设公式包括：
    Figure PCTCN2022070988-appb-100002
    Figure PCTCN2022070988-appb-100003
    其中，P(公交)指所述行程为公交出行行程的概率，P(驾车)指所述行程为小汽车出行行程的概率，T 手机指所述行程起讫点间的实际旅行时间，T 驾车指所述小汽车驾车时间，T 公交指所述公交时间。
  9. 一种基于知识图谱和居民出行链的共乘查询方法，其特征在于，包括：
    获取查询用户和查询时间范围，在基于居民出行链生成的知识图谱中获取所述查询用户在所述查询时间范围内的出行链，其中，所述居民出行链基于权利要求1至8中任一项所述的基于多源数据融合的居民出行链生成方法构建而成，所述知识图谱包含用户实体、停留位置实体、非交通设施实体、车次实体、线路实体、车辆实体、交通基础设施实体，所述用户实体与所述居民出行链的用户个体对应，所述停留位置实体、所述非交通设施实体与所述居民出行链的活动对应，所述车次实体、所述线路实体、所述车辆实体、 所述交通基础设施实体与所述居民出行链的出行链对应；
    从所述出行链中读取出行位置；
    当所述出行位置为公共交通工具时，获取所述出行位置的车次和所述查询用户乘坐该车次的起终点，基于所述车次和所述起终点从所述知识图谱中获取所述查询用户的共乘人员。
  10. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器读取并运行时，实现如权利要求1-8任一项所述的基于多源数据融合的居民出行链生成方法，或如权利要求9所述的基于知识图谱和居民出行链的共乘查询方法。
PCT/CN2022/070988 2021-06-30 2022-01-10 基于多源数据融合的居民出行链生成方法及共乘查询方法 WO2023273292A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110730818.9 2021-06-30
CN202110730818.9A CN113256987B (zh) 2021-06-30 2021-06-30 基于多源数据融合的居民出行链生成方法及共乘查询方法

Publications (1)

Publication Number Publication Date
WO2023273292A1 true WO2023273292A1 (zh) 2023-01-05

Family

ID=77190176

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/070988 WO2023273292A1 (zh) 2021-06-30 2022-01-10 基于多源数据融合的居民出行链生成方法及共乘查询方法

Country Status (2)

Country Link
CN (1) CN113256987B (zh)
WO (1) WO2023273292A1 (zh)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116170761A (zh) * 2023-04-25 2023-05-26 北京城建交通设计研究院有限公司 一种手机信令数据综合扩样校核方法及系统
CN116206452A (zh) * 2023-05-04 2023-06-02 北京城建交通设计研究院有限公司 一种城市交通出行的稀疏数据特征分析方法和系统
CN116233759A (zh) * 2023-04-25 2023-06-06 北京城建交通设计研究院有限公司 一种居民出行轨迹调查方法及系统
CN116628527A (zh) * 2023-07-24 2023-08-22 浙江大学 一种集成出行策略的设计方法及系统
CN117407606A (zh) * 2023-12-14 2024-01-16 青岛理工大学 一种基于大语言模型和知识图谱的旅游线路推荐方法
CN118152831A (zh) * 2024-01-25 2024-06-07 哈尔滨工业大学 一种基于个体的公交出行时空聚类方法
CN118427448A (zh) * 2024-07-05 2024-08-02 西北大学 一种多模式出行路径推荐方法与装置

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113256987B (zh) * 2021-06-30 2021-10-29 深圳市城市交通规划设计研究中心股份有限公司 基于多源数据融合的居民出行链生成方法及共乘查询方法
CN113889284B (zh) * 2021-09-16 2022-07-05 同济大学 一种基于公共交通知识图谱的传染病接触目标追踪方法
CN113891252B (zh) * 2021-09-18 2024-02-02 苏州规划设计研究院股份有限公司 基于手机信令数据的轨道客流全程od提取方法及系统
CN114416710B (zh) * 2021-12-29 2023-04-07 苏州大学 一种快速路车辆od位置提取方法及系统
CN115062244B (zh) * 2022-08-18 2023-02-03 深圳市城市交通规划设计研究中心股份有限公司 一种基于多源数据的时空伴随者及同职住人员搜索方法
CN115934701B (zh) * 2023-01-06 2023-08-01 深圳市城市交通规划设计研究中心股份有限公司 基于知识图谱的个体出行补全方法、电子设备及存储介质
CN116128128A (zh) * 2023-01-17 2023-05-16 北京融信数联科技有限公司 基于智能体图谱的城市职住平衡预测方法、系统和介质
CN118036853B (zh) * 2024-04-11 2024-07-05 深圳市城市交通规划设计研究中心股份有限公司 一种个体出行链生成方法、装置及存储介质
CN118378860B (zh) * 2024-06-24 2024-10-18 民航机场规划设计研究总院有限公司 一种基于航班波的机场陆侧交通需求分配方法及系统

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009104169A1 (en) * 2008-02-21 2009-08-27 Nxp B.V. In-vehicle traffic navigation terminal
WO2014060786A1 (en) * 2012-10-15 2014-04-24 Freescale Semiconductor, Inc. Method and apparatus for reporting traffic information
CN103810851A (zh) * 2014-01-23 2014-05-21 广州地理研究所 一种基于手机定位的交通出行模式识别方法
CN108055645A (zh) * 2018-01-19 2018-05-18 深圳技术大学（筹） 一种基于手机信令数据的路径识别方法及系统
CN108320501A (zh) * 2017-12-21 2018-07-24 江苏欣网视讯软件技术有限公司 基于用户手机信令的公交线路识别方法
CN110874362A (zh) * 2019-10-29 2020-03-10 青岛海信网络科技股份有限公司 一种数据关联分析方法及装置
CN111653097A (zh) * 2020-05-29 2020-09-11 南京瑞栖智能交通技术产业研究院有限公司 一种基于手机信令数据且含个人属性修正的城市出行方式综合识别方法
CN111653099A (zh) * 2020-06-10 2020-09-11 南京瑞栖智能交通技术产业研究院有限公司 基于手机信令数据的公交客流od获取方法
CN112530166A (zh) * 2020-12-01 2021-03-19 江苏欣网视讯软件技术有限公司 基于信令数据与大数据分析识别公交出行上下车站点的方法与系统
CN113256987A (zh) * 2021-06-30 2021-08-13 深圳市城市交通规划设计研究中心股份有限公司 基于多源数据融合的居民出行链生成方法及共乘查询方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7003316B1 (en) * 2002-02-22 2006-02-21 Virtual Fonlink, Inc. System and method for wireless transactions
RU2580597C2 (ru) * 2014-05-13 2016-04-10 Федеральное государственное казенное военное образовательное учреждение высшего профессионального образования "Военная академия материально-технического обеспечения имени генерала армии А.В. Хрулёва" Система контроля соблюдения правил дорожного движения
CN111653093B (zh) * 2020-05-29 2022-06-17 南京瑞栖智能交通技术产业研究院有限公司 一种基于手机信令数据的城市出行方式综合识别方法

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009104169A1 (en) * 2008-02-21 2009-08-27 Nxp B.V. In-vehicle traffic navigation terminal
WO2014060786A1 (en) * 2012-10-15 2014-04-24 Freescale Semiconductor, Inc. Method and apparatus for reporting traffic information
CN103810851A (zh) * 2014-01-23 2014-05-21 广州地理研究所 一种基于手机定位的交通出行模式识别方法
CN108320501A (zh) * 2017-12-21 2018-07-24 江苏欣网视讯软件技术有限公司 基于用户手机信令的公交线路识别方法
CN108055645A (zh) * 2018-01-19 2018-05-18 深圳技术大学（筹） 一种基于手机信令数据的路径识别方法及系统
CN110874362A (zh) * 2019-10-29 2020-03-10 青岛海信网络科技股份有限公司 一种数据关联分析方法及装置
CN111653097A (zh) * 2020-05-29 2020-09-11 南京瑞栖智能交通技术产业研究院有限公司 一种基于手机信令数据且含个人属性修正的城市出行方式综合识别方法
CN111653099A (zh) * 2020-06-10 2020-09-11 南京瑞栖智能交通技术产业研究院有限公司 基于手机信令数据的公交客流od获取方法
CN112530166A (zh) * 2020-12-01 2021-03-19 江苏欣网视讯软件技术有限公司 基于信令数据与大数据分析识别公交出行上下车站点的方法与系统
CN113256987A (zh) * 2021-06-30 2021-08-13 深圳市城市交通规划设计研究中心股份有限公司 基于多源数据融合的居民出行链生成方法及共乘查询方法

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116170761A (zh) * 2023-04-25 2023-05-26 北京城建交通设计研究院有限公司 一种手机信令数据综合扩样校核方法及系统
CN116233759A (zh) * 2023-04-25 2023-06-06 北京城建交通设计研究院有限公司 一种居民出行轨迹调查方法及系统
CN116206452A (zh) * 2023-05-04 2023-06-02 北京城建交通设计研究院有限公司 一种城市交通出行的稀疏数据特征分析方法和系统
CN116206452B (zh) * 2023-05-04 2023-08-15 北京城建交通设计研究院有限公司 一种城市交通出行的稀疏数据特征分析方法和系统
CN116628527A (zh) * 2023-07-24 2023-08-22 浙江大学 一种集成出行策略的设计方法及系统
CN116628527B (zh) * 2023-07-24 2023-11-10 浙江大学 一种集成出行策略的设计方法及系统
CN117407606A (zh) * 2023-12-14 2024-01-16 青岛理工大学 一种基于大语言模型和知识图谱的旅游线路推荐方法
CN117407606B (zh) * 2023-12-14 2024-03-05 青岛理工大学 一种基于大语言模型和知识图谱的旅游线路推荐方法
CN118152831A (zh) * 2024-01-25 2024-06-07 哈尔滨工业大学 一种基于个体的公交出行时空聚类方法
CN118427448A (zh) * 2024-07-05 2024-08-02 西北大学 一种多模式出行路径推荐方法与装置

Also Published As

Publication number Publication date
CN113256987A (zh) 2021-08-13
CN113256987B (zh) 2021-10-29

Similar Documents

Publication Publication Date Title
WO2023273292A1 (zh) 基于多源数据融合的居民出行链生成方法及共乘查询方法
Kondor et al. Estimating savings in parking demand using shared vehicles for home–work commuting
Gurumurthy et al. Analyzing the dynamic ride-sharing potential for shared autonomous vehicle fleets using cellphone data from Orlando, Florida
CN110111574B (zh) 一种基于流量树分析的城市交通不平衡评价方法
Nielsen et al. Relevance of detailed transfer attributes in large-scale multimodal route choice models for metropolitan public transport passengers
Huang et al. A method for bus OD matrix estimation using multisource data
JP6307376B2 (ja) 交通分析システム、交通分析プログラムおよび交通分析方法
Patlins et al. The new approach for passenger counting in public transport system
Chen et al. An analysis of movement patterns between zones using taxi GPS data
Wei et al. Optimal routing design of feeder transit with stop selection using aggregated cell phone data and open source GIS tool
CN113573238B (zh) 一种基于手机信令的联程旅客出行链识别方法
CN113344268B (zh) 一种城市交通出行数据分析方法
Zahabi et al. Transit trip itinerary inference with GTFS and smartphone data
Yao et al. Trip end identification based on spatial-temporal clustering algorithm using smartphone positioning data
Tian et al. Identifying residential and workplace locations from transit smart card data
Kostakos Using Bluetooth to capture passenger trips on public transport buses
CN111931968B (zh) 一种应用手机gps数据来优化公共自行车站点布局的方法
Crawford et al. Analysing spatial intrapersonal variability of road users using point-to-point sensor data
CN111177294B (zh) 基于车辆轨迹数据求解交叉口中介中心性的方法
Lan et al. Inferring alighting bus stops from smart card data combined with cellular signaling data
Tang et al. Spatio-temporal reachable area calculation based on urban traffic data
Kumar Transit Origin Destination Estimation using Automated Data
CN109508815B (zh) 基于地铁ic卡数据的通学活动空间测度分析方法
Fadeev et al. Using Automated Fare Collection System Data to Determine Transport Demand
Kong et al. Research on OD Estimation of Public Transit Passenger Flow Based on Multi-source 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: 22831129

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 13.05.2024)