A method for generating dual-travel mode carpooling stations based on travel reservation demand

By collecting passenger travel information, time-space clustering analysis and Dijkstra algorithm to plan driving routes, a dual-travel mode ride-sharing solution is formed, which solves the problems of traffic congestion and resource waste, and improves vehicle usage efficiency and passenger experience.

CN114358386BActive Publication Date: 2024-11-15ENG TECH RES INST OF ZHENJIANG JIANGSU UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111487717.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-07
Publication Date
2024-11-15
Estimated Expiration
2041-12-07

AI Technical Summary

Technical Problem

The existing technology failed to effectively use the dual mode of private cars and buses to ride together when solving traffic congestion problems, and lacked attention to time factors, resulting in waste of resources and poor passenger travel experience.

Method used

By collecting passenger reservation information, selecting passenger groups with similar travel paths, conducting time-space clustering analysis, forming an initial plan for engaging in a dual-travel mode, determining whether to drive a private car, using the Dijkstra algorithm to plan driving routes, generating information on the combined site, and improving Vehicle usage efficiency and passenger experience.

Benefits of technology

It improves vehicle usage efficiency, alleviates road congestion, improves passenger travel experience, and optimizes the allocation of transportation resources through the initial plan for engaging in the dual-travel mode.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114358386B_ABST
    Figure CN114358386B_ABST
Patent Text Reader

Abstract

The present invention provides a method for generating a dual travel mode carpooling station based on the reservation travel demand. First, the passenger reservation travel information is collected. According to the passenger travel demand, passengers with similar travel paths are selected as the initial passenger group; then the boarding location and travel time of the initial passenger group are subjected to spatiotemporal clustering analysis to determine the initial carpooling station information; then, according to the passenger's commuting mode selection, an initial carpooling plan of the dual travel mode with two means of transportation is formed; finally, the initial carpooling plan is released to the passengers to confirm whether to participate in the carpooling and driving information, and the vehicle type, number of vehicles and route information are arranged according to the passenger feedback information, the carpooling station is generated, and the carpooling station information is released to the passengers. The present invention can improve the utilization efficiency of vehicles and can improve the utilization efficiency of vehicles.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent public transportation, and in particular relates to a method for generating a dual travel mode carpooling site based on reservation travel demand. Background Art

[0002] In recent years, with the continuous acceleration of my country's urbanization process, the number of cars in my country has increased rapidly, and urban infrastructure has been unable to meet the travel needs during peak travel periods, especially during the morning and evening peaks on weekdays. In my country's large and medium-sized cities, traffic congestion is becoming increasingly serious, causing inconvenience to people's lives, increasing social production costs, and having an adverse impact on economic development and my country's urbanization. At the same time, many commuter private cars have only one driver, and during non-peak travel periods, public transportation vehicles have a serious phenomenon of being empty, resulting in a waste of resources. The above phenomenon shows that there is an urgent need to use a method to make reasonable use of traffic resources and alleviate the current situation of urban traffic congestion.

[0003] With the rise of customized public transportation and online car-hailing industries, carpooling is becoming more and more popular among travelers. So far, 29 provinces including Beijing, Hebei, and Shanxi have opened customized public transportation services, with more than 5,400 operating routes and an annual passenger volume of nearly 180 million. Online car-hailing, ride-sharing and other travel modes have also become one of the main ways for people to travel. Foreign scholars mostly focus on analyzing the spatial location of carpooling stations and use clustering algorithms to determine the location of customized public transportation stations. At present, the research on synthetic stations is mostly based on hierarchical clustering, DBSCAN algorithm, and K-means algorithm to perform cluster analysis of spatial geographic location, obtain coordinate points, analyze and use GPS data, SP, RP survey data, and online car-hailing data to carry out research on the generation of customized public transportation stations, and have conducted extensive research in path planning and carpooling information matching. However, considering the characteristics of domestic travel commuting, the above research results only consider the use of buses as a single means of transportation, and do not discuss the carpooling problem under the dual mode of private cars and buses. In addition, the selection of carpooling stations lacks attention to time, and mostly focuses on cluster analysis of spatial location. Summary of the invention

[0004] In view of the shortcomings in the prior art, the present invention provides a method for generating dual travel mode carpooling stations based on reservation travel needs, which can improve the utilization efficiency of vehicles, alleviate road congestion, and improve passengers' travel experience.

[0005] The present invention achieves the above technical objectives through the following technical means.

[0006] A method for generating a dual travel mode carpooling station based on reservation travel demand, specifically:

[0007] Collect passengers' travel reservation information and select passengers with similar travel routes as the initial passenger group based on their travel needs;

[0008] Conduct spatiotemporal clustering analysis on the boarding locations and travel times of the initial passenger group to determine the initial carpooling site information; form an initial dual-travel mode carpooling plan with two means of transportation based on the passengers' commuting mode choices; publish the initial carpooling plan to passengers to confirm whether they participate in carpooling and driving information; arrange vehicle types, number of vehicles and route information based on passenger feedback, generate carpooling sites, and publish carpooling site information to passengers.

[0009] Furthermore, the passenger travel demand includes the latitude and longitude coordinates of the departure point of each passenger i. i , the latitude and longitude coordinates of the end point i , time to reach the destination Td i , whether they are willing to drive, and if so, the standard passenger capacity of the vehicle they intend to drive Q i .

[0010] Furthermore, the calculation formula for the similarity of the travel paths is:

[0011]

[0012] Where: p represents the last stop of the first passenger's initial path, q represents the last stop of the second passenger's initial path, Head(tr i ) represents the first stop in the initial travel path of passenger i, Rest(tr i ) represents the subsequence of all stations except the first point on the initial travel path of passenger i;

[0013] The initial travel path is:

[0014] Where: p i =(lat, lng, t), indicating that the passenger is located at the geographic coordinate position {lat, lng) at time t, where lat and lng represent the latitude and longitude of the bus stop respectively.

[0015] Furthermore, the initial plan for dual travel mode carpooling is specifically as follows: determine whether the carpooler drives a private car based on the carpooler's information; if the carpooler drives a private car, count the number of passengers excluding the driver to obtain the number of empty seats C in the private car; if the carpooler does not drive a private car, count the total number of non-driving passengers V; determine whether C is greater than V. When C is greater than V, do not add buses, and plan the private car route based on the preliminary travel route using the Dijkstra algorithm; otherwise, select the bus model based on the difference, and plan the bus route based on the preliminary travel route using the Dijkstra algorithm.

[0016] Furthermore, the commuting modes include private cars and buses.

[0017] Furthermore, the carpooling station information includes location information, departure time, and information about the vehicle that each passenger takes.

[0018] Further, cluster information is extracted from the initial group of passengers, and the lo i , travel time ti, generate passenger information set p i :p i = {lo i ,ti};

[0019] Construct the loss function J(u, r): Among them: uk is the cluster center;

[0020] When passengers are divided into K clusters, the corresponding loss function is recorded as D K , and calculate: Gap(K)=E(logD k )-logD k , where E(logD k ) is logD k expectations;

[0021] When Gap(K) takes the maximum value, the corresponding K value is the optimal number of classification clusters. K-means clustering is used according to the K value to generate the initial ride-sharing station set {u i}; Each initial ride-sharing station contains the geographical coordinate cluster center l si , travel time demand cluster center t si And the information of the passengers si , where x si ∈X, X is the overall passenger travel demand set in the region.

[0022] The beneficial effects of the present invention are:

[0023] (1) The present invention forms an initial plan for dual-travel mode carpooling with two means of transportation based on the passenger's commuting mode selection. The initial plan for dual-travel mode carpooling is specifically as follows: determine whether the carpooler drives a private car based on the carpooler's information. If the carpooler drives a private car, count the number of passengers other than the driver to obtain the number of empty seats C in the private car. If the carpooler does not drive a private car, count the total number of non-driving passengers V; determine whether C is greater than V. When C is greater than V, do not add buses, and plan the private car route based on the initial travel route using the Dijkstra algorithm. Otherwise, select a bus model based on the difference, and plan the bus route based on the initial travel route using the Dijkstra algorithm. The initial plan for dual-travel mode carpooling of the present invention improves the utilization efficiency of vehicles and alleviates road congestion;

[0024] (2) The present invention extracts cluster information from the initial group of carpooling passengers and determines the initial carpooling station information. The carpooling passengers have the shortest walking distance to the bus station and the shortest waiting time, thereby improving the travel experience of passengers. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a flow chart of the method for generating a dual travel mode carpooling station based on reservation travel demand according to the present invention;

[0026] Figure 2 This is the vehicle type matching flow chart of the present invention. DETAILED DESCRIPTION

[0027] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0028] like Figure 1 As shown, a method for generating a dual travel mode carpooling station based on a reservation travel demand specifically includes the following steps:

[0029] Step (1) divides the road network of the selected area into nodes and collects travel reservation information from passengers in the selected area.

[0030] Step (1.1), divide the road network of the selected area into nodes, establish a road network layer based on the existing roads, and set up road network nodes based on the existing bus stops on the road network, including the latitude and longitude coordinates of the bus stops and the lane driving direction; mark the road network nodes to generate a road network node set {D};

[0031] Step (1.2), calculate the latitude and longitude coordinates of the departure point of each passenger i in the selected area. i , the latitude and longitude coordinates of the end point i , time to reach the destination Td i, whether they are willing to drive, and if so, the standard passenger capacity of the vehicle they intend to drive Q i , and conduct statistics;

[0032] Step (1.3), construct the overall passenger travel demand set in the region

[0033] X={x1,x2,···,x n} (1)

[0034] Where: n is the number of passengers with travel needs;

[0035] The demand x of a certain passenger in the passenger travel demand set i The sample feature vector of (i∈(1, 2, ...n)) is:

[0036] x i = {lo i ,ld i ,ti} (2)

[0037] Sample features include the spatial coordinates of the start and end points (including lo i andld i ) and travel time ti (based on the time to reach the destination Td i Reverse deduction).

[0038] Step (1.4), according to the Dijkstra algorithm, based on the starting point and destination coordinates lo of a passenger i i ,ld i , preliminarily plan the path of passenger i and calculate the initial travel path tr of passenger i i , initial travel path tr i It is composed of nodes in the road network node set {D}; the initial travel path of a passenger i where p i =(lat, lng, t), which represents the geographic coordinate position of passenger i at time t, where lat and lng represent the latitude and longitude of the bus stop respectively.

[0039] Step (2): Select passengers with a high initial travel path overlap as the initial group of passengers to share the ride

[0040] The statistical method for the initial group of carpooling passengers is as follows: based on the preliminary planned initial travel paths of the passengers, the similarity of the initial travel paths of the passengers is calculated using the DTW (Dynamic Time Warping) algorithm, and the passengers with relatively similar travel paths are selected to form the initial group of carpooling passengers.

[0041] The similarity calculation process is shown in formula (3):

[0042]

[0043] Where: p represents the last stop of one passenger’s initial path, q represents the last stop of another passenger’s initial path, Head(tr i ) represents the first stop in the initial travel path of passenger i, Rest(tr i ) represents the subsequence of all stations except the first point on the initial travel path of passenger i;

[0044] Step (3): Perform spatiotemporal cluster analysis on the boarding locations and travel times of the initial group of carpool passengers to determine the initial carpool station information.

[0045] Step (3.1) extracts cluster information from the initial group of passengers and selects the starting point coordinates lo of each passenger i i , travel time ti, generate the passenger information set p required for K-means cluster analysis i , as shown in formula (4).

[0046] p i ={loi,ti} (4)

[0047] Step (3.2), assuming u1, u2, ..., uk are K cluster centers, r nk ∈0, 1 means p i Whether it belongs to the K cluster with uk as the cluster center, construct the loss function J(u, r), as shown in formula (5).

[0048]

[0049] When passengers are divided into K clusters, the corresponding loss function is recorded as D K According to Gap statistic, when Gap(K) reaches the maximum value, the corresponding K value is the optimal number of classification clusters. The calculation method is shown in formula (6):

[0050] Gap(K)=E(logD k )-logD k (6)

[0051] Among them, E(logD k ) is logD k The expectation is usually generated using Monte Carlo simulation.

[0052] Step (3.3), use K-means clustering according to the K value in step (3.2) to generate the initial ride-sharing site set {u i}; Each initial ride-sharing station contains the geographical coordinate cluster center lsi , travel time demand cluster center t si And the information of the passengers si , where x si ∈X.

[0053] Step (4), based on the passenger's commuting mode selection, an initial dual-travel mode carpooling plan with two means of transportation is formed.

[0054] like Figure 2 As shown in the figure, it is determined whether the carpooler drives a private car based on the carpooler information. If the carpooler drives a private car, the number of passengers except the driver is counted to obtain the number of empty seats C in the private car. The calculation method of C is shown in formula (7). If the carpooler does not drive a private car, the total number of non-driving passengers (those who have travel needs but do not drive a private car) V is counted.

[0055]

[0056] Where m is the number of people who choose to travel by private car.

[0057] Determine whether C is greater than V. When C is greater than V, do not add buses, and use the Dijkstra algorithm to plan private car routes based on the preliminary travel routes. Otherwise, select bus models based on the difference, and use the Dijkstra algorithm to plan bus routes based on the preliminary travel routes.

[0058] Step (5): publish carpooling information to passengers, collect passenger feedback, and generate carpooling sites.

[0059] The initial carpooling plan is released to passengers to confirm whether they participate in the carpooling and driving information. The car model, number of vehicles and route information are arranged according to the passenger feedback information, and the carpooling station is generated; the carpooling station information is released to passengers, including location information, departure time, and information on the vehicle each passenger takes, and the final carpooling station is generated.

[0060] The embodiments are preferred implementations of the present invention, but the present invention is not limited to the above-mentioned implementations. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essential content of the present invention belong to the protection scope of the present invention.

Claims

1. A method for generating dual travel mode carpooling stations based on reservation travel demand, Features: Collect passengers' travel reservation information and select passengers with similar travel routes as the initial passenger group based on their travel needs; Conduct spatiotemporal clustering analysis on the boarding locations and travel times of the initial passenger group to determine the initial carpooling site information; form an initial dual-travel mode carpooling plan with two modes of transportation based on the passengers' commuting mode choices; Publish the initial carpooling plan to passengers, confirm whether to participate in the carpooling and driving information, arrange the type of vehicle, number of vehicles and route information according to the passenger feedback information, generate the carpooling station, and publish the carpooling station information to passengers; the commuting mode includes private car and bus; the carpooling station information includes location information, departure time, and vehicle information of each passenger; The statistical method of the initial passenger group is: according to the initial travel paths of the passengers preliminarily planned, the similarity of the initial travel paths of the passengers is calculated using DTW, and the passengers with relatively similar travel paths are selected to form the initial passenger group; The calculation formula for the similarity of the travel paths is: Where: p represents the last stop of the first passenger's initial path, q represents the last stop of the second passenger's initial path, Head(tr i ) represents the first stop in the initial travel path of passenger i, Rest(tr i ) represents the subsequence of all stations except the first point on the initial travel path of passenger i; The spatiotemporal cluster analysis is specifically as follows: extracting cluster information from the initial passenger group, selecting the latitude and longitude coordinates of the starting point of each passenger i i , travel time ti, generate passenger information set p i :p i = {lo i ,ti}; construct loss function Where: uk is the cluster center; when the passengers are divided into K clusters, the corresponding loss function is recorded as D K , and calculate: Gap(K)=E(logD k )-logD k , where E(logD k ) is logD k When Gap(K) takes the maximum value, the corresponding K value is the optimal number of classification clusters. K-means clustering is used according to the K value to generate the initial ride-sharing site set {u i }; Each initial ride-sharing station contains the geographical coordinate cluster center l si , travel time demand cluster center t si And the information of the passengers si , where x si ∈X, X is the overall passenger travel demand set in the region; The dual travel mode carpooling initial plan is specifically as follows: determine whether the carpooler drives a private car based on the carpooler's information; if the carpooler drives a private car, count the number of passengers except the driver to obtain the number of empty seats C in the private car; if the carpooler does not drive a private car, count the total number of non-driving passengers V; determine whether C is greater than V, when C is greater than V, do not add buses, and plan the private car route based on the preliminary travel route using the Dijkstra algorithm; otherwise, select the bus model based on the difference, and plan the bus route based on the preliminary travel route using the Dijkstra algorithm.

2. The method for generating a dual travel mode carpooling station based on reservation travel demand according to claim 1, It is characterized in that The passenger travel demand includes the latitude and longitude coordinates of the starting point of each passenger i, lo i , the latitude and longitude coordinates of the end point i , time to reach the destination Td i , whether they are willing to drive, and if so, the standard passenger capacity of the vehicle they intend to drive Q i .

3. The method for generating a dual travel mode carpooling station based on reservation travel demand according to claim 1, It is characterized in that The initial travel path is: Where: p i =(lat,lng,t), indicating that the passenger is located at the geographic coordinate position (lat,lng) at time t, where lat and lng represent the latitude and longitude of the bus stop respectively.