WO2021238231A1 - 共享单车流动系统、基于子区划分的自动调度系统及方法 - Google Patents

共享单车流动系统、基于子区划分的自动调度系统及方法 Download PDF

Info

Publication number
WO2021238231A1
WO2021238231A1 PCT/CN2021/070687 CN2021070687W WO2021238231A1 WO 2021238231 A1 WO2021238231 A1 WO 2021238231A1 CN 2021070687 W CN2021070687 W CN 2021070687W WO 2021238231 A1 WO2021238231 A1 WO 2021238231A1
Authority
WO
WIPO (PCT)
Prior art keywords
bicycle
sub
demand
pick
shared
Prior art date
Application number
PCT/CN2021/070687
Other languages
English (en)
French (fr)
Inventor
赵亮
徐聪
吴云凤
白翰
崔娜
王修光
王国军
陈相融
陈启倪
Original Assignee
山东交通学院
山东正衢交通工程有限公司
山东正衢交通工程研究院
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 山东交通学院, 山东正衢交通工程有限公司, 山东正衢交通工程研究院 filed Critical 山东交通学院
Publication of WO2021238231A1 publication Critical patent/WO2021238231A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06315—Needs-based resource requirements planning or analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/20—Monitoring the location of vehicles belonging to a group, e.g. fleet of vehicles, countable or determined number of vehicles
    • G08G1/207—Monitoring the location of vehicles belonging to a group, e.g. fleet of vehicles, countable or determined number of vehicles with respect to certain areas, e.g. forbidden or allowed areas with possible alerting when inside or outside boundaries

Definitions

  • the present invention belongs to the technical field related to shared bicycles, and specifically relates to a shared bicycle flow system, an automatic dispatch system and method based on sub-zone division.
  • the existing technology mainly uses some algorithms of machine learning to analyze historical data of shared travel and build models for forecasting. It does not fully consider the actual needs of users. From the perspective of user needs, it can improve demand forecasting to a certain extent. Accuracy. Although the existing methods take into account the demand forecasting methods of users' reservations, they do not fully explore the real needs of users. Forecasting based on reservation data only lacks consideration of changing factors, making the forecast results inaccurate; at the same time, the factors are not considered. Comprehensive, did not fully consider the changing factors of the environment around the site, and ignored the influence of the attraction points around the site on the changes in demand.
  • the main consideration is the internal correlation between each site, and the internal characteristics of each site itself and the impact of the surrounding environment characteristics are not fully considered, and the effectiveness of the scheduling cannot be satisfied, resulting in the continuous ineffective scheduling work. , Resulting in a lot of waste of resources.
  • the purpose of the present invention is to provide a shared bicycle mobile system, an automatic dispatch system and method based on sub-zone division.
  • the proposed mobile system realizes the linkage of stations in a certain area and predicts through a comprehensive demand forecasting method
  • the demand of each site is then divided into dynamic sub-zones to form a demand scheduling plan for each site in the sub-zone.
  • the mobile system implements automatic transportation of shared bicycles according to the scheduling plan, and provides users with efficient and convenient Access car service.
  • the first object of the present invention is to provide a shared bicycle flow system, including an above-ground conveying device installed at each bicycle pick-and-place point, an underground conveying device connected to the ground conveying device of each bicycle pick-and-place point, and an over-ground conveying device or an underground conveying device.
  • a shared bicycle flow system including an above-ground conveying device installed at each bicycle pick-and-place point, an underground conveying device connected to the ground conveying device of each bicycle pick-and-place point, and an over-ground conveying device or an underground conveying device.
  • the second object of the present invention is to provide an automatic scheduling method based on sub-zone division, which includes the following steps:
  • the bicycle pick-and-place points are dynamically divided into sub-zones, and the scheduling plan is generated according to the division results;
  • the third object of the present invention is to provide an automatic dispatch system based on sub-zone division, including the above-mentioned shared bicycle flow system, and a control platform that sends dispatch instructions to the shared bicycle flow system;
  • the control platform includes a shared bicycle demand forecast system and dynamic sub-zones Divide the scheduling system;
  • the shared bicycle demand forecasting system is configured to execute the shared bicycle demand forecasting method in the above-mentioned automatic dispatch method based on subzone division;
  • the dynamic sub-zone division scheduling system is configured to execute the dynamic sub-zone division scheduling method in the above-mentioned automatic scheduling method based on sub-zone division.
  • the shared bicycle mobility system of the present invention combines the above ground and underground space to form a vehicle transportation network in conditional areas such as machine non-isolation and human non-isolation, and basically covers all shared bicycle demand locations in a certain area, and according to needs
  • the vehicles on the storage device or transportation track are transported to various bicycle pick-and-place points.
  • the bicycle pick-and-place points do not need to store a large number of bicycles for a long time.
  • a small area can meet the area requirements of the bicycle pick-and-place points.
  • users can access the bicycle conveniently and quickly in the area along the route; at the same time, the storage device is set in a multi-layer structure, which can reduce the footprint of bicycle storage. Breaking through the limitations of the original manpower involved in dispatching transportation, it can avoid the problems of limited dispatchers and untimely dispatching; it can effectively regulate user parking behaviors, avoid random parking problems, and efficiently dispatch vehicles, which can alleviate the difficulties of car use and parking .
  • the shared bicycle demand prediction method of the present disclosure combines user demand and shared travel attractiveness based on the random forest algorithm, can accurately obtain user demand, improve the accuracy of scheduling, reduce or avoid the execution of invalid scheduling, reduce the number of scheduling, and improve The system's scheduling execution efficiency.
  • the dynamic sub-area division scheduling method of the present disclosure is based on the tree-like branch principle to integrate internal and external factors for dynamic sub-area division scheduling, combined with the second aspect of the present disclosure based on the random forest algorithm to integrate user needs and shared travel attractiveness for shared bicycle demand
  • the prediction method supplements and optimizes the shared bicycle mobility system based on human non-isolation or machine non-isolation as described in the first aspect of the present disclosure, and can form a demand scheduling plan in each sub-area to maximize user scheduling requirements.
  • Figure 1 is a schematic diagram of the flow system structure of Embodiment 1 of the present disclosure
  • FIG. 2 is a schematic diagram of the structure of the storage device of Embodiment 1 of the present disclosure.
  • FIG. 3 is a schematic diagram of the structure of each storage layer in the storage device of Embodiment 1 of the present disclosure
  • FIG. 4 is a schematic diagram of the installation position of the mobile system device in the mobile system of Embodiment 1 of the present disclosure
  • FIG. 5 is a schematic diagram of the structure of the above-ground conveying device of Embodiment 1 of the present disclosure.
  • FIG. 6 is a schematic diagram of the structure of the underground conveying device of Embodiment 1 of the present disclosure.
  • FIG. 7 is a schematic diagram of the structure of the bicycle carrier of Embodiment 1 of the present disclosure.
  • FIG. 8 is a schematic diagram of the structure of the smart induction electronic lock in the bicycle carrier of the first embodiment of the present disclosure
  • FIG. 9 is a block diagram of an automatic scheduling system based on sub-region division according to Embodiment 2 of the present disclosure.
  • FIG. 10 is a flowchart of a method for forecasting demand for shared bicycles according to Embodiment 3 of the present disclosure.
  • FIG. 11 is a flowchart of a method for dynamic sub-area partitioning and scheduling according to Embodiment 4 of the present disclosure
  • FIG. 12 is a control method of execution scheduling according to Embodiment 5 of the present disclosure.
  • the present invention proposes a shared bicycle mobility system, an automatic dispatch system and method based on sub-zone division.
  • the shared bicycle mobile system includes an above-ground conveying device installed at each bicycle pick-and-place point, and an underground conveying device connected to the ground conveying device at each bicycle pick-and-place point , And the multi-layer storage device 17 that can provide bicycles for the above-ground conveying device or the underground conveying device; adjacent bicycle pick-and-place points and the above-ground storage device 17 are connected by the ground conveying device or the underground conveying device to form a mobile transportation network sharing bicycles.
  • the storage device 17 is used to store the bicycles, and the above-ground conveying device and the underground conveying device are used to transport the bicycles to various bicycle pick-and-place points according to the demand of the bicycles.
  • This embodiment combines the use of the ground and underground space to form a vehicle transportation network, and transports the vehicles in the storage device 17 or the transportation track to various bicycle pick-and-place points as needed.
  • the bicycle pick-and-place points do not need to store a large number of bicycles for a long time.
  • the area can meet the area requirements of bicycle pick-and-place points.
  • the storage device 17 is arranged in a multi-layer structure, which can reduce the floor space of bicycle storage. At the same time, the storage device 17 on the ground is used for unified storage. Place a certain number of bicycles.
  • the above-ground conveying device is installed at the mechanical non-isolated green belt 33, and the underground conveying device can be installed under the road 34. At the same time, the above-ground conveying device and the underground conveying device can be effectively avoided. The impact of mobile systems on road traffic.
  • the storage device 17 is used to store bicycles and can be set underground or on the ground.
  • the storage device 17 is set on the ground, which can effectively reduce construction costs.
  • the location of the storage device 17 can be set as required.
  • the storage device 17 can be set near a bicycle access point with a large amount of vehicles, and one or more bicycle access points share the same ground storage device 17. Specifically, it can be set in a suitable area near the machine-non-isolation or human-non-isolation, such as the green belts on both sides of the road.
  • FIG. 1 it is a schematic diagram of the structure of the above-ground conveying device connected to the storage device 17, and FIG. 5 is a schematic diagram of the structure of a separate above-ground conveying device.
  • the storage device 17 may have a structure as shown in Figs. 1 and 2, including a device housing that constitutes a bicycle storage space, a plurality of storage layers in the housing, and a transportation track that can move the bicycle between the storage layers. , The transportation track is provided with the entrance and exit of the transportation track to the storage layer in each storage layer.
  • three layers can be set as an example, and the first storage layer 20, the second storage layer 21, and the third storage layer 22 are provided.
  • the lowest storage layer is provided with a bicycle entrance 18 and a bicycle exit 19, which are respectively connected to an above-ground conveying device or an underground conveying device for moving the bicycle to the storage device 17, or transporting the bicycle of the storage device 17 to each bicycle for fetching. Put some.
  • the specific structure of the transportation track may be: a spirally ascending rotating track structure, which includes a column 36 arranged in the housing and a spirally ascending rail 23 fixed on the column.
  • each storage layer includes an entrance/exit rail 24 connected to the spirally ascending rail 23 and a vehicle storage area 25 connected to the entrance/exit rail 24 through the rail.
  • the vehicle storage area 25 is set as an inclined plane with a certain arc and angle, and a bicycle fixing device 37 is arranged on the inclined surface; optionally, the bicycle fixing device 37 can be set as
  • the clamping device is equipped with two clamping blocks, and the movement of the clamping blocks is controlled by a pneumatic or electric device.
  • a fixed device for placing the bicycle can be set in each device. It may be a bicycle carrier 8 that is free to move relative to each device.
  • the bicycle carrier 8 includes a bottom plate, a driving device and a bicycle fixing device arranged on the bottom plate to drive the bottom plate to move ,
  • the bicycle information recognition device, the main controller 39 and the wireless communication module, the main controller 39 is respectively connected with the driving device, the bicycle fixing device, the bicycle information recognition device and the wireless communication module, and the main controller 39 is wirelessly connected with the control module 35.
  • the driving device may be electrically driven, including a driving motor 12, a power supply battery connected to the driving motor 12, and a moving mechanism connected to the driving motor 12.
  • a solar panel 15 may also be provided. It is electrically connected with the power supply battery to provide electrical energy for the drive motor.
  • the moving mechanism can be wheels, crawler wheels or crawlers.
  • control module 35 can adopt a single-chip microcomputer.
  • the bicycle fixing device may include an inductive smart lock 9 and an inductive device 16 fixed on the bottom plate.
  • the inductive smart lock 9 and the inductive device 16 are respectively electrically connected to the main controller 39 for transmitting the inductive information to Control module 35.
  • the inductive smart lock 9 is automatically locked.
  • the inductive smart lock can be set to any shape, such as arc or polygon.
  • the sensing device 16 may be a pressure sensor, which is used to determine whether a bicycle is placed on the floor, so as to improve the reliability of the operation of the bicycle carrier 8 and the accuracy of providing bicycle placement information.
  • the information identification device includes a reed switch 10 arranged on a shared bicycle, a telescopic device 38 arranged on the bottom plate of the bicycle carrier 8 opposite to the reed switch 10, and a magnet 11 arranged on the top of the telescopic device 38,
  • the telescopic device 38 is electrically connected to the main controller 39, and the reed switch 10 is wirelessly connected to the main controller.
  • the sensing device 16 After sensing the signal, the sensing device 16 transmits the sensing signal to the main controller 39 of the transportation device, and controls the telescopic device 38 to pop up to a certain height, so that the magnet 11 on the top is close to the reed pipe 10 in the bottom shell near the pedal of the shared bicycle.
  • the tube 10 is closed and can be connected to the central control unit of the shared bicycle.
  • the central control unit transmits the lock signal to the main controller 39 through the wireless mobile communication module, and then controls the inductive smart lock 9 to close the lock, and the telescopic component automatically retracts into the transportation device .
  • the lock information of the inductive smart lock 9 on the bicycle carrier 8 of this embodiment is linked with the shared bicycle background system, and the shared bicycle background system receives the closing information of the shared bicycle on the lock and receives the inductive smart lock 9
  • the shared bicycle background system receives the closing information of the shared bicycle on the lock and receives the inductive smart lock 9
  • the car is returned successfully; when the user closes the lock on the bicycle, the operation of returning the car fails when the lock information of the inductive smart lock 9 is not received.
  • the shared bicycle After the shared bicycle is fixed, the shared bicycle is controlled to be locked to realize the operation of returning the bicycle, which can regulate the parking behavior of users, and make the shared bicycle parked on the transportation device in the mobile system to realize the circulation of shared bicycles.
  • the bicycle carrier 8 may further include a positioning module 13.
  • the positioning module 13 may be a GPS positioning module.
  • the above-ground transportation device can be configured as a track structure, as shown in Figure 1 or 5.
  • the above-ground transportation device includes a control module 35 and a ground transportation track 2 laid on the ground. Bar 1, and a bicycle access port provided on the protective barrier to provide bicycle access, and a vehicle access interactive device set at the bicycle access port for receiving interactive information.
  • the control module 35 is respectively carried by the bicycle through the communication module
  • the device 8 is in communication connection with the access car interactive device.
  • the control module 35 receives the information of the access vehicle interaction device and controls the bicycle carrier device 8 to carry the bicycle to the corresponding position.
  • the access car interactive device is used to receive the user's access to the bicycle information.
  • the protective fence 1 may be a fence or a fence.
  • the car access interaction device includes an automatic retractable door 4 provided at the bicycle access port, and a button for controlling the opening and closing of the retractable door, which may include a door open button 5, a car retrieval button 6, and a car storage button 7. .
  • the position sensor 3 is wirelessly connected to the main controller.
  • the controller controls the bicycle carrier 8 to stop.
  • the position sensor 3 may also use an RFID tag.
  • the bicycle carrier 8 is provided with an RFID reader. When the RFID reader detects the corresponding tag information, the vehicle stops.
  • the underground conveying device 26 is used to provide an underground transportation channel for the vehicle transfer between various bicycle pick-up and place points or the vehicle transfer between the storage device 17 and each bicycle pick-up and place point. Such as setting under the ground at some intersections.
  • the underground conveying device can realize the transportation of shared bicycles from the ground to the ground and from the ground to the ground.
  • the underground transportation device may include an underground transportation track 27, and a pressure expansion device 30 arranged at the connection port of the ground and underground.
  • the pressure expansion device 30 When the pressure expansion device 30 is extended to the first position, the pressure expansion and contraction The upper end surface of the device 30 is flush with the ground and butted.
  • the pressure expansion device 30 When the pressure expansion device 30 is compressed to the second position, the upper end surface of the pressure expansion device 30 is flush with the underground transportation track 27 and butted.
  • the pressure telescopic device 30 may include a bicycle carrying part 31 arranged from top to bottom, a telescopic mechanism fixedly connected to the bicycle carrying part 31 and a fixed platform 28, and a telescopic drive arranged on the fixed platform 28 electrically connected to the telescopic mechanism Powerplant 29.
  • the bicycle bearing portion 31 may be a bearing plate or a bearing plate with a track, and the track shape and structure of the track are matched with the underground transportation track 27.
  • the fixed platform 28 provides stable support.
  • a pressure sensor 32 may also be provided on the bicycle carrying part 31 to detect whether a bicycle or a bicycle carrying device 8 is placed on the bicycle carrying part 31.
  • the telescopic drive power device 29 may be a hydraulic drive device, and the telescopic mechanism presses a telescopic rod.
  • system may also include a control platform, which is communicatively connected with the control module 35 in the mobile system.
  • the user's car needs can be obtained, and the car can be accessed directly by pressing the door button.
  • control module 35 real-time receiving and analysis of operating information in the system, and output control
  • the bicycle and the bicycle carrier 8 in the entire mobile system are instructed and dispatched; the bicycle carrier 8 transports the bicycle in accordance with the control dispatch instruction of the control module 35.
  • the bicycle carrier 8 determines whether the parking is regulated.
  • control module 35 dispatches the shared bicycles or parked bicycle carriers of the nearby stations or the storage device 17 in real time according to the situation of each site and user needs. 8. To meet the needs of users for car storage and retrieval to the maximum extent.
  • This embodiment provides an automatic dispatch system based on the division of sub-zones, divides the shared bicycle sub-zones according to user needs, and automatically dispatches the shared bicycle mobile system described in Example 1 for dispatching each shared bicycle or bicycle in the mobile system.
  • the automatic dispatch system based on sub-zone division includes the shared bicycle mobility system described in Example 1, and a control platform that sends scheduling instructions to the shared bicycle mobility system.
  • the control platform includes a shared bicycle demand prediction system and dynamic sub-systems. District division dispatch system;
  • Shared bicycle demand forecasting system It is configured to predict the demand of users in different periods of each site, and obtain the forecast demand of each bicycle pick-and-place point; provide a theoretical basis for the number of vehicles distributed in the entire mobile system.
  • Dynamic sub-area division scheduling system It is configured to divide dynamic sub-areas according to the predicted demand of each bicycle pick-and-place point obtained, and dispatch according to the ratio of the bicycle demand of each bicycle pick-and-place point in the sub-area to generate scheduling
  • the solution is sent to the control module 35 so that the control module 35 controls the bicycle or the bicycle carrier 8 in the flow system.
  • the shared bicycle demand forecasting method is based on the random forest algorithm to integrate user needs and shared travel attractiveness, and can accurately obtain user needs.
  • This method can be implemented on the control platform connected to the control module, and the specific can be realized by the shared bicycle demand forecasting system, such as As shown in Figure 10, it includes the following steps:
  • Step 1 User demand statistics: Obtain user travel information and travel reservation information, and count the first bicycle demand X 1 at the bicycle pick-and-place point;
  • Step 2 Demand forecast based on the attractiveness of shared travel: determine the attraction points in the area near the bicycle pick-up and place point, and calculate the second bicycle demand X 2 of the bicycle pick-up and place point according to the attractiveness of each attraction point;
  • Step 3 Calculate the third bicycle demand X 3 at the pick-and-place point of the bicycle based on the random forest algorithm
  • Step 4 The weighted sum of the demand obtained in the above steps is used to obtain the demand of each bicycle pick-and-place point.
  • the user's travel information can be obtained through incentive feedback.
  • Incentive feedback can be the use of points incentives to send travel questionnaires, which include the start and end points of the main cycling route, travel time period, and user opinions. After receiving the questionnaire information, points are added to the account for filling in the questionnaire to ensure the reliability of the data. If the user’s riding information is seriously inconsistent with the content of the questionnaire, certain points will be deducted from the user.
  • Users mainly include fixed users with weekly, monthly or annual cards and some ordinary users.
  • step 2 the method of determining the attraction points in the area near the bicycle pick-and-place point, and calculating the second bicycle demand according to the attraction of each attraction point includes the following steps:
  • Step 21 Divide the attraction level of the attraction point
  • Attraction points are public places with a relatively large flow of people, such as hospitals, schools, parks, bus stations, subway stations, etc. Attraction point refers to a place that attracts people to share travel.
  • Optional can be divided according to the size of the flow of people, first level: bus station, subway station, second level: community, supermarket, school, third level: catering, park square, fourth level: other;
  • Step 22 Determine the attraction points in the setting area of the bicycle pick-and-place point, and determine the attraction reduction coefficient ⁇ x of each attraction point according to the attraction level;
  • the bicycle pick-and-place point setting area can be set to a range of one kilometer around the bicycle pick-and-place point.
  • the reduction factor ⁇ x can be set in proportion to the flow of people.
  • Step 23 Calculate the second bicycle demand X 2 at the pick-and-place point of shared bicycles according to the attractiveness reduction coefficient ⁇ x ;
  • X 2 is the demand for shared bicycles; S total is the area of one kilometer near the attraction point; Si suction is the area of a nearby attraction point, i is the i-th bicycle pick-and-place point; K is the attraction point for slow travel Proportion: For sites with obvious time and age characteristics, the travel ratio can be divided by age level; N is the number of attraction points; ⁇ x is the attraction reduction coefficient determined according to different levels of shared travel attraction points.
  • Step 3 The method of calculating the third bicycle demand X 3 at the pick-and-place point of the bicycle based on the random forest algorithm includes the following steps:
  • Step 31 Obtain a sample data set.
  • Step 32 Perform sample extraction on the sample data set to obtain training subsets of multiple decision trees.
  • the bootsrap re-sampling method is adopted to extract S training sample subsets from the total sample to construct S regression trees.
  • the extracted training samples are the training set, and the unextracted samples from the total sample are used as the test set.
  • Step 33 Decision tree construction: Based on the principle of loss minimization, each training subset is trained to obtain a decision tree. During the decision tree training process, a set number of feature variables with greater correlation are selected to participate in the node splitting of the decision tree. A random forest regression model is obtained through training of a training subset;
  • a decision tree is generated based on the principle of minimizing loss.
  • S training sample subsets generate a total of S decision trees to form a random forest.
  • select The feature variable of is set to not exceed log 2 M+1, where M is the number of associated feature variables.
  • the participating feature variables are selected according to the principle of correlation, sorted according to the size of the correlation, and the correlation G is selected Larger part of the characteristic variables participate in the decision tree node splitting process.
  • the method of determining the relevance can be as follows:
  • X is the demand variable
  • Y i is a certain characteristic variable
  • G is the correlation between the demand variable and a certain characteristic value
  • A is the sum of the number of data of all demand variables X and all characteristic variables Y
  • a i is a certain All data of the feature corresponds to the number of data in A.
  • the test set data is used for simulation, the error of the decision tree is estimated, and the parameters of the decision tree are optimized.
  • the error estimates of S decision trees are averaged to obtain the random forest generalization error estimates, and the model parameters are optimized.
  • Step 34 Random forest regression model prediction results: obtain the bicycle travel data and corresponding characteristic variable data at the bicycle pick-and-place point in real time, input them into the random forest regression model, obtain the voting results of each decision tree, and obtain the weighted random forest regression prediction results as The third bicycle demand X 3 at the bicycle pick-and-place point;
  • the prediction results output by the random forest regression prediction model are produced by the voting results of each decision tree.
  • the prediction results of random forest regression are as follows:
  • Y i wherein data related factors, H ik is a single decision trees prediction model, S is the total number of the constructed tree, X Y is a demand for motorcycles shared regression prediction result.
  • step 3 the number of feature variables for constructing the decision tree is limited according to the size of the correlation, and the random forest algorithm is optimized, so as to predict the demand X 3 more accurately.
  • ⁇ 1 , ⁇ 2 and ⁇ 3 represent the corresponding weights;
  • X is the total demand for the site;
  • X 1 is the demand for the first bicycle;
  • X 2 is the demand for the second bicycle based on the attractiveness of shared travel;
  • X 3 It is the third bicycle demand based on the prediction of the random forest algorithm.
  • This embodiment integrates user needs, shared travel attractiveness and random forest algorithm, and deeply mines user needs.
  • User) car demand and at the same time integrate some general user reservation information to improve the accuracy of demand forecasting; the introduction of the shared travel attractiveness index fully considers the impact of changes in demand caused by attraction points around the site; it can be analyzed more conveniently
  • a high-precision random forest algorithm is selected, through the correlation analysis of feature variables, the number of feature variables for constructing a decision tree is limited, and the accuracy of random forest prediction is increased, thereby improving the accuracy of demand forecasting.
  • the dynamic sub-area division scheduling method can adjust the sub-area range in real time through dynamic sub-area division, and improve the flexibility and timeliness of scheduling.
  • This method can be implemented on the control platform connected to the control module, and specifically can be implemented by a dynamic sub-area division scheduling system, as shown in Figure 11, including the following steps:
  • Step 1 Obtain bicycle access data and bicycle trajectory information of each bicycle pick-and-place point;
  • Step 2 According to the acquired data, classify the pick-and-place points of bicycles according to different characteristics
  • Step 3 According to the classification results, the adjacent bicycle pick-and-place points are dynamically divided according to the dynamic changes in demand to form multiple sub-areas;
  • Step 4 Dispatch each sub-area according to the result of sub-area division: If the scheduling cannot meet the bicycle demand of the sub-area, perform the above steps 1-3 to re-divide the sub-area.
  • step 1 collect historical data information and real-time dynamic access information of each station's access vehicles.
  • the data information of each bicycle pick-and-place point can be obtained more conveniently and quickly, so that the internal characteristics of each station's access car can be accurately analyzed.
  • step 2 according to the acquired data, the bicycle pick-and-place points are classified according to different characteristics
  • the classification of bicycle pick-and-place points according to different characteristics can include classification by time characteristics, classification by demand level, and so on.
  • Step 21 Classify according to time characteristics, which can be divided into:
  • Time-divided stations In certain specific time periods, the demand for bicycles is relatively large, such as morning peak and evening peak hours.
  • All-time stations stations with relatively large demand in all time periods
  • Step 22 Divide the ordinary stations according to the demand level: According to the predicted demand of each bicycle pick-up point, the demand is divided according to the demand level.
  • Step 3 the adjacent bicycle pick-and-place points are dynamically divided according to the dynamic changes of demand to form multiple dynamic sub-areas. Specifically, it can be:
  • Step 31 According to the time characteristics of time-sharing stations and full-time stations, combined with the dynamic demand of surrounding stations, form complementary sub-zones with surrounding stations and form a corresponding scheduling plan;
  • Step 32 Perform dynamic sub-area division and scheduling for common sites, which may include the following steps:
  • Step 321 Combine sites with high complementarity among ordinary sites as a pick-and-place point; complementarity means that high demand does not overlap in time, and the time periods of high demand are staggered.
  • Step 322 Dynamically select the pick-and-place point with the largest real-time demand and stable demand within the set area as the main site;
  • the pick-and-place points are geographically continuous. According to the principle of hierarchical sampling, a major site is selected every certain number of sites.
  • the selection principle can be: demand changes without obvious temporal characteristics; within a certain period of time The demand is higher than that of the surrounding sites, and the pick-and-place point of the bicycle with the greatest demand is dynamically selected according to the real-time demand;
  • Step 333 Divide the sub-areas according to the tree-like branching principle: use the principle of complementarity and the principle of shared travel attractiveness to select the most suitable surrounding stations and merge them with the main station as the center to form a sub-area;
  • the tree-like branch principle is to fully integrate the characteristics of the linkage of the mobile system sites. First select the main site as the root, and then grow according to the principle of one branch upwards. When the growth demand cannot be met, select other nodes in this branch to continue Grow.
  • the system of this embodiment is in the shape of a belt or road network. When dividing sub-zones, the main site is divided into a line according to the division conditions, and other sites on this line are selected for division when the division conditions are not met.
  • Q is the demand difference; Q is now the number of vehicles at the site; Q is also the number of vehicles returned within a certain period of time at the site; Q needs to be the number of vehicles required for the site within a certain period of time.
  • K is the reduction factor
  • Q 1 is the demand difference of the sub-area before the site is merged
  • Q 2 is the demand difference of the sub-area after the site is merged
  • H is the demand complementarity after the site is merged.
  • A represents the attractiveness of shared travel at a certain station
  • D i represents the distance from the station to an attractive point in the nearby attraction area
  • ⁇ x is the attractiveness reduction coefficient determined according to different levels of shared travel attractive points
  • R i attracts represents the radius of an attraction point in the attraction area
  • S total represents the total area of the attraction area.
  • step 4 scheduling is performed according to the results of sub-area division: for the obtained sub-area, the bicycle adjustments between the various pick-and-place points in the sub-area are carried out.
  • the sub-area level that is, between the sub-area
  • steps 1-4 to re-divide the sub-zones.
  • scheduling is performed in each sub-area first, and the degree of association in the sub-area is fully considered, so that the accuracy of the scheduling is improved, and the efficiency of the scheduling is improved.
  • Carrying out the scheduling between the pick-and-place points of bicycles in the sub-zone specifically: adjust the distribution of bicycle pick-and-place points in the sub-zone according to the demand ratio.
  • the scheduling between the sub-areas is carried out, and if it cannot be adjusted effectively, the current sub-area division is adjusted.
  • the current scheduling scheme cannot meet the needs of users, and the dynamic sub-area is re-divided to form a new scheduling adjustment scheme, which can realize the dynamic adjustment of the sub-area division and improve the flexibility of scheduling.
  • the dynamic sub-areas division method of this embodiment fully combines the characteristics of the mobile system site linkage, adopts the tree-like branch principle to select and divide the points, and the division conditions fully consider the influence of the site's internal and external factors, and use the principle of complementary demand and sharing
  • the selection of stations is based on the principle of travel attractiveness, considering not only the influence of surrounding changing factors, but also the different internal time characteristics of each station. It is divided into full-time points, time-specific points, and ordinary points, and the corresponding sub-zones are divided. Scheduling, each sub-area implements demand-proportional scheduling, which can better meet the user's scheduling needs.
  • This embodiment provides an automatic dispatching method based on sub-zone division.
  • the method is implemented in the control platform in the system described in Embodiment 2.
  • the shared bicycle sub-zone is divided according to user needs.
  • the system performs automatic dispatch, dispatching each shared bicycle or bicycle carrier 8 in the mobile system.
  • the automatic scheduling method based on sub-zone division includes the following steps:
  • the bicycle pick-and-place point is dynamically divided into sub-areas, and the scheduling plan is generated according to the division result;
  • Step 1 adopts the shared bicycle demand prediction method described in Embodiment 3, and step 2 adopts the dynamic sub-area division scheduling method described in Embodiment 4 to obtain a scheduling plan; it also includes a control method executed by the control module 35 according to the scheduling plan.
  • the control method executed by the control module 35, as shown in FIG. 12, includes the following:
  • Step 1 Transportation control between bicycle storage points: According to the scheduling plan, adjust the number of bicycles at each station by controlling the carrying device 8 between adjacent stations;
  • the carrier device 8 receives the signal sent by the control module 35, and moves to the designated position according to the above-ground conveying device and the underground conveying device 26.
  • the drive motor 12 of the carrier 8 When there is a demand for pick-up or parking, the drive motor 12 of the carrier 8 is activated to control the shared bicycle to be transported to the corresponding telescopic door 4; when the system has no demand for access to the car, the system detects that all doors are closed, control The module 35 realizes the equidistant emission of shared bicycles in each station according to the positioning data of the positioning device 3 at the bottom of each carrier 8;
  • the pressure sensor 32 When passing through the intersection, it is dispatched by the underground conveying device 26, and when the shared bicycle and its carrier 8 are transported to the pressure sensor 32 of the underground conveying device, the pressure sensor 32 transmits a signal to the telescopic driving power device 29 to activate the pressure telescopic rod 30, and then Control the bicycle carrying part 31 to transport the shared bicycle downwards. Finally, the pressure telescopic rods are all retracted into the fixed platform 28. After the shared bicycle is transported underground, it is transported through the underground transportation track 27. When it reaches the designated place, it passes and downwards. In the same transportation method, the shared bicycle is transported to the ground through the pressure telescopic device 30, and the scheduling task is continued.
  • Step 2 Access control of the access car: obtain user demand information of the access car interactive device;
  • the sensing devices (16) located at both ends of the carrying device are automatically locked when they sense that a vehicle is placed, and the parking is completed.
  • the nearest bicycle or carrier 8 may be at the current site, and there is no schedulable bicycle or carrier 8 at the current site, and may be in an adjacent site or in the storage device 17.
  • Step 3 Storage control of bicycles: Count the number of bicycles in the sub-zone, including normal vehicles and vehicles to be repaired, and place normal vehicles and repaired vehicles separately;
  • the surplus bicycles in the storage device of the sub-area are dispatched to the sub-area where the demand is not met.
  • the vehicle In order to facilitate maintenance by maintenance personnel, if the vehicle needs to be repaired, it is stored in the first storage layer 20 of the storage device. When the number of vehicles to be repaired exceeds the set value of the total number of stored vehicles in the storage device 17, it is sent Repair instructions to the terminal of the maintenance personnel.
  • a first storage layer 20, a second storage layer 21, and a third storage layer 22 are provided.
  • the shared bicycle When the vehicle does not need to be repaired, the shared bicycle will be transported and stored upward along the spirally ascending track 23. When storing, it will be transported upward through the spirally ascending track 23. When passing through the second storage layer 21, the shared bicycle will pass through the second storage layer 21 and the spiral. The entrance and exit 24 connected to the ascending track 23 enters the second floor.
  • the shared bicycle storage and fixing devices are arranged in a circle around the center, and the vehicles arriving from the spiral ascending track 23 are stored through continuous rotation.
  • the storage layer is 21
  • the first storage layer 20 and the third storage layer 22 are successively stored; when the shared bicycles are dispatched from the storage device 17, the vehicles at the first storage layer 20 are dispatched first, and when the underlying vehicles are not enough to dispatch, then Dispatching the shared bicycles of the upper storage device in turn.
  • the system performs timely scheduling adjustments in accordance with the above-mentioned steps and methods to meet user needs to the greatest extent.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Data Mining & Analysis (AREA)
  • Educational Administration (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种共享单车流动系统、基于子区划分的自动调度系统及方法，共享单车流动系统包括设置在各个单车取放点的地上输送装置，连接各个单车取放点的地上输送装置的地下输送装置,以及能够为地上输送装置或者地下输送装置提供单车的多层储存装置（17）；相邻的单车取放点、地上储存装置(17)通过地上运送装置或者地下运送装置连接，形成共享单车的流动运送网络。提出的流动系统，实现一定区域内站点的联动，通过综合的需求预测方法预测各个站点需求量，然后进行动态子区划分，形成子区内各个站点的需求调度方案，最后流动系统按照调度方案实现共享单车的自动运输，在用户有需求时，最大限度给用户提供高效便捷的存取车服务。

Description

共享单车流动系统、基于子区划分的自动调度系统及方法 技术领域
本发明属于涉及共享单车相关技术领域，具体的说，是涉及共享单车流动系统、基于子区划分的自动调度系统及方法。
背景技术
这里的陈述仅提供与本发明相关的背景技术，而不必然地构成现有技术。
近年来，共享单车的出现完善了公共交通系统，解决了市民出行“最后一公里”的难题，符合绿色环保的生活理念，但是随着共享单车数量的急剧增加，随之产生了一系列问题。在城市的大街小巷，单车乱停乱放的问题逐渐突出，“无路可走，无处可停”的现象引起社会各界的关注，规范停车点和推广电子围栏等一系列措施的实施，虽一定程度规范了停车问题，但是相对削弱了共享出行的便利性和随机性，出现找车难、停车难的问题，导致用户体验不佳；目前所有共享单车站点都是零散的，不利于协调控制，在集中管理方面有很大难度。
发明人发现，目前在共享单车运输领域，主要存在以下问题：
第一，全部需要依靠人力参与运输，共享单车覆盖区域较广，所以会耗费大量人力资源，并且存在调度不及时的问题；虽然有些发明提出一些共享单车运载车以及一些搬运装置，但是运载车在实现自动运输时，受周围环境的影响，存在很大的不安全因素。虽然有搬运装置，但还是需要人员实现共享单车站点与站点之间的运输，大大耗费人力并且存在调度人员有限，调度不及时的问题，不能很好的满足用户需求。
第二，各个站点共享单车投放数量与实际需求量不平衡问题较为普遍，不能很好的迎合用户需求。在共享单车需求预测领域，现有技术主要利用机器学习的一些算法分析共享出行的历史数据建立模型进行预测，没有充分考虑结合用户的实质需求，从用户需求的角度出发可以一定程度提高需求预测的准确性，虽然现有方法中考虑到用户预约的需求预测方法，但是没有充分挖掘用户的真正需求，仅仅依靠预约数据进行预测缺乏对变动因素的考虑，使得预测结果不准确；同时，考虑因素不全面，没有充分考虑站点周围环境的变动因素，忽略了站点周围吸引点对需求量的变动影响。
第三，在共享单车调度方法方面，主要考虑各个站点之间内在关联度，没有充分考虑各个站点本身的内在特征以及周围环境特征的影响，不能满足调度的有效性，导致无效的调度工作持续进行，造成大量的资源浪费。
发明内容
针对现有技术存在的不足，本发明的目的是提供共享单车流动系统、基于子区划分的自动调度系统及方法，提出的流动系统，实现一定区域内站点的联动，通过综合的需求预测方法预测各个站点需求量，然后进行动态子区划分，形成子区内各个站点的需求调度方案，最后流动系统按照调度方案实现共享单车的自动运输，在用户有需求时，最大限度给用户提供高效便捷的存取车服务。
为了实现上述目的，本发明是通过如下的技术方案来实现：
本发明的第一目的是提供共享单车流动系统，包括设置在各个单车取放点的地上输送装置，连接各个单车取放点地上输送装置的地下输送装置,以及能够为地上输送装置或者地下输送装置提供单车的多层储存装置；相邻的单车取放点、地上储存装置通过地上运送装置或者地下运送装置连接，形成共享单车的流动运送网络。
本发明的第二目的是提供基于子区划分的自动调度方法，包括如下步骤：
获取单车取放点的单车取放数据，基于随机森林算法融合用户需求和共享出行吸引力的方法，对各个单车取放点的需求量进行预测；
根据需求量的预测结果，基于树状分支结合内外因素，对单车取放点进行子区动态划分，根据划分结果生成调度方案；
按照调度方案执行调度的控制。
本发明的第三目的是提供基于子区划分的自动调度系统，包括上述的共享单车流动系统，以及向共享单车流动系统发送调度指令的控制平台；控制平台包括共享单车需求预测系统和动态子区划分调度系统；
共享单车需求预测系统被配置为用于执行上述的基于子区划分的自动调度方法中的共享单车需求预测方法；
或者，动态子区划分调度系统被配置为用于执行上述的基于子区划分的自动调度方法中的动态子区划分调度方法。
上述本发明的实施例的有益效果如下：
(1)本发明共享单车流动系统结合利用地上和地下的空间，在机非隔离、人非隔离等有条件的地带形成车辆运送网络，在一定区域内基本覆盖所有共享单车需求地，根据需要将储存装置或运输轨道上的车辆运送到各个单车取放点，单车取放点不用长时间储存大量的单车，较小的区域就可以满足单车取放点的面积要求，并且在流动系统内设置有多个连续的单车取放点，用户可以在沿线区域方便快捷的存取车；同时将储存装置设置为多层的结构，可以减少单车存放的占地面积。突破原有人力参与调度运输的局限，可以避免调度人员有限，调度不及时等问题；可有效规范用户停车行为，避免乱停乱放问题，同时高效调度车辆，可以缓解用车难、停车难问题。
(2)本公开的共享单车需求预测方法，基于随机森林算法融合用户需求和共享出行吸引力，能准确获得用户需求，提高调度的准确性，减少或避免无效调度的执行，减少调度次数，提高系统的调度执行效率。
(3)本公开的动态子区划分调度方法，基于树状分支原理综合内外因素进行动态子区划分调度，结合本公开第二方面基于随机森林算法融合用户需求和共享出行吸引力的共享单车需求预测方法，对本公开第一方面所述的基于人非隔离或机非隔离的共享单车流动系统进行补充优化，可以形成各个子区内的需求调度方案，最大限度满足用户调度需求。
附图说明
构成本发明的一部分的说明书附图用来提供对本发明的进一步理解，本发明的示意性实施例及其说明用于解释本发明，并不构成对本发明的不当限定。
图1是本公开实施例1的流动系统结构示意图；
图2是本公开实施例1的储存装置结构示意图；
图3是本公开实施例1的储存装置中各存储层的结构示意图；
图4是本公开实施例1的流动系统中流动系统装置设置位置示意图；
图5是本公开实施例1的地上输送装置的结构示意图；
图6是本公开实施例1的地下输送装置的结构示意图；
图7是本公开实施例1的单车运载装置的结构示意图；
图8是本公开实施例1的单车运载装置中智能感应电子锁结构示意图；
图9是本公开实施例2的基于子区划分的自动调度系统的框图；
图10是本公开实施例3的共享单车需求预测方法流程图；
图11是本公开实施例4的动态子区划分调度方法流程图；
图12是本公开实施例5的执行调度的控制方法；
其中：1、防护栏，2、地面运输轨道，3、位置传感器，4、自动伸缩门，5、开门按钮，6、取车按钮，7、存车按钮，8、单车运载装置，9、感应式智能锁，10、干簧管，11、磁铁，12、驱动电机，13、定位装置，14、电子锁固定底座，15、太阳能电池板，16、感应装置，17、储存装置，18、单车入口，19、单车出口，20、第一存储层，21、第二存储层，22、第三存储层，23、螺旋式上升轨道，24、出入口轨道，25、车辆存储区，26、地下输送装置，27、地下运输轨道，28、固定平台，29、伸缩驱动动力装置，30、压力伸缩装置，31、单车承载部分，32、压力传感器，33、机非隔离绿化带，34、道路，35、控制模块，36、立柱，37、车辆储存区内固定单车的装置，38、伸缩装置，39、主控制器。
为显示各部位位置而夸大了互相间间距或尺寸，示意图仅作示意使用。
具体实施方式
应该指出，以下详细说明都是例示性的，旨在对本发明提供进一步的说明。除非另有指明，本发明使用的所有技术和科学术语具有与本发明所属技术领域的普通技术人员通常理解的相同含义。
需要注意的是，这里所使用的术语仅是为了描述具体实施方式，而非意图限制根据本发明的示例性实施方式。如在这里所使用的，除非本发明另外明确指出，否则单数形式也意图包括复数形式，此外，还应当理解的是，当在本说明书中使用术语“包含”和/或“包括”时，其指明存在特征、步骤、操作、器件、组件和/或它们的组合；
为了方便叙述，本发明中如果出现“上”、“下”、“左”“右”字样，仅表示与附图本身的上、下、左、右方向一致，并不对结构起限定作用,仅仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的设备或元件必须具有特定的方位，以特定的方位构造和操作，因此不能理解为对本发明的限制。
术语解释部分:本发明中如出现术语“安装”、“相连”、“连接”、“固定”等，应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或为一体；可以是机械连接，也可以是电连接，可以是直接连接，也可以是通过中间媒介间接相连，可以是两个元件内部连接，或者两个元件的相互作用关系，对于本领域的普通技术人员而言，可以根据具体情况理解上述术语在本发明的具体含义。
正如背景技术所介绍的，现有技术中存在不足，为了解决如上的技术问题，本发明提出了共享单车流动系统、基于子区划分的自动调度系统及方法。
实施例1
本发明的一种典型的实施方式中，如图1-8所示，共享单车流动系统，包括设置在各个单车取放点的地上输送装置，连接各个单车取放点地上输送装置的地下输送装置，以及能够为地上输送装置或者地下输送装置提供单车的多层储存装置17；相邻的单车取放点、地上储存装置17通过地上运送装置或者地下运送装置连接，形成共享单车的流动运送网络。
储存装置17用于存放单车，地上输送装置和地下输送装置用于按照单车需求将单车运送至各个单车取放点。
本实施例结合利用地上和地下的空间，形成车辆运送网络，根据需要将储存装置17或运 输轨道中的车辆运送到各个单车取放点，单车取放点不用长时间储存大量的单车，较小的区域就可以满足单车取放点的面积要求。将储存装置17设置为多层的结构，可以减少单车存放的占地面积，同时通过地上储存装置17进行统一存储，当各个取放点有车辆需求，通过地上输送装置或者地下输送装置对各个取放点投放一定数量的单车。
如图4所示，以一个常规交叉口为例，地上输送装置设置在机非隔离绿化带33处，地下输送装置可以设置道路34的下方，同时设置地上输送装置和地下输送装置可以有效避免该流动系统对路面交通的影响。
可选的，储存装置17用于存放单车，可以设置在地下或者地上，优选的，储存装置设置在地上，可以有效减少建设成本。
可选的，储存装置17的设置位置可以根据需要进行设置，储存装置17可以设置在用车量较大的单车存取点附近，一个或者多个单车存取点共用一个地上储存装置17。具体的，可以设置在机非隔离或人非隔离附近合适地带，例如道路两边的绿化带。
如图1所示，为地上输送装置连接储存装置17的结构示意图，图5为单独的地上输送装置结构示意图。
储存装置17可以为如图1和图2所示的结构，包括构成单车容纳空间的装置壳体，设置壳体内的多个存储层，以及能够将单车在各个存储层之间移动运输的运输轨道，运输轨道在每一存储层分别设置运输轨道通向该存储层的入口和出口。
本实施例可以设置三层为例，设置了第一存储层20、第二存储层21和第三存储层22。
可以理解的，最低的存储层设置单车入口18和单车出口19，分别连接地上输送装置，或者地下输送装置，用于将单车移动至储存装置17，或者将储存装置17的单车运送至各个单车取放点。
在一些实施例中，可以设置运输轨道的具体结构为：螺旋式上升旋转轨道结构，包括设置在壳体内的立柱36及固定在立柱上的螺旋式上升轨道23。
可选的，如图3所示，每个存储层包括设置在该层与螺旋式上升轨道23相连接的出入口轨道24，以及与出入口轨道24通过轨道连接的车辆存储区25。
为了提高存储区车辆的存储面积，尽可能的多放置单车，车辆存储区25设置为一定弧度和角度的斜面，所述斜面上设置单车固定装置37；可选的，单车固定装置37可以设置为夹紧装置，设置两个夹紧块，通过气动或者电动装置控制夹紧块移动。
为实现单车在地上输送装置、地下输送装置和储存装置17之间运输传送，可以在各个装置内设置放置单车的固定装置，如可以在轨道上设置固定平台，可以为设置单独移动的运载设备，可以为相对于各个装置自由移动的单车运载装置8。
作为一种可以实现的结构，所有单车取放点还设置有控制模块35，如图7和8所示，单车运载装置8包括底板，设置在底板上的驱动底板移动的驱动装置和单车固定装置，单车信息识别装置，主控制器39以及无线通信模块，主控制器39分别与驱动装置、单车固定装置、单车信息识别装置以及无线通信模块分别连接，主控制器39与控制模块35无线连接。
具体的，驱动装置可以采用电驱动，包括驱动电机12、与驱动电机12连接的供电电池、与驱动电机12连接的移动机构，优选的，还可以设置太阳能电池板15，所述太阳能电池板15与供电电池电性连接，为驱动电机提供电能。移动机构可以为车轮、履带车轮或者履带。
可选的，控制模块35可以采用单片机。
可选的，单车固定装置可以包括固定在底板上的感应式智能锁9和感应装置16，所述感 应智能锁9和感应装置16分别与主控制器39电连接，用于将感应信息传输至控制模块35。当感应到有共享单车放置到运载装置8上时，感应式智能锁9自动锁上。
感应式智能锁可以设置为任意形状，如圆弧状或者多边形等。
感应装置16可以为压力传感器，用于确定该底板上是否放置了单车，提高单车运载装置8工作的可靠性和提供单车放置信息的准确性。
可以理解的，还包括电子锁固定底座14，用于固定感应式智能锁9的锁体。
可选的，信息识别装置包括设置在共享单车上的干簧管10，设置在单车运载装置8的底板上与干簧管10位置相对的伸缩装置38，设置在伸缩装置38顶端的磁铁11，所述伸缩装置38与主控制器39电连接，干簧管10与主控制器无线连接。
感应装置16感应到信号后，将感应信号传给运输装置的主控制器39，控制伸缩装置38弹出一定高度，使顶部的磁铁11接近共享单车踏板附近底部壳体内的干簧管10，干簧管10闭合，可以连通共享单车的中心控制单元，中心控制单元通过无线移动通信模块将关锁信号传给主控制器39，进而控制感应式智能锁9关锁，伸缩组件自动收缩回运输装置内。
进一步地，本实施例的单车运载装置8上的感应式智能锁9关锁信息与共享单车后台系统联动，共享单车后台系统接收到共享单车上车锁的关闭信息并且接收到感应式智能锁9关锁信息，还车成功；当用户关闭单车上的锁，在未接收到感应式智能锁9关锁信息时，还车操作失败。
共享单车被固定后，再控制共享单车上锁，实现还车操作，可以规范用户停车行为，使共享单车规范停放在流动系统中的运输装置上面，实现共享单车的流通。
可选的，为了确定单车运载装置8的具体位置，单车运载装置8还可以包括定位模块13。定位模块13可以为GPS定位模块。
在一些实施例中，地上输送装置可以设置为轨道结构，如图1或5所示，地上输送装置包括控制模块35和铺设在地面的地面运输轨道2，设置在地面运输轨道2两侧的防护栏1，以及设置在防护栏上提供单车出入通道的单车存取口，以及设置在单车存取口处的用于接收交互信息的存取车交互装置，控制模块35分别通过通信模块与单车运载装置8和存取车交互装置通信连接。
控制模块35接收存取车交互装置的信息控制单车运载装置8运载单车至相应的位置。存取车交互装置用于接收用户的存取单车的信息。
可选的，防护栏1可以为围墙或者为栅栏。
在一些实施例中，存取车交互装置包括单车存取口处设置的自动伸缩门4，以及用于控制伸缩门的开关的按钮，可以包括开门按钮5、取车按钮6和存车按钮7。
还包括设置在单车存取口处的位置传感器3，位置传感器3与主控制器无线连接，当单车运载装置8移动至位置传感器3处，位置传感器3将动作信号传输至主控制器39，主控制器控制单车运载装置8停车。位置传感器3也可以采用RFID标签，所述单车运载装置8上设置RFID阅读器，当RFID阅读器检测到相应的标签信息，停车。
地下输送装置26用于为各个单车取放点之间的车辆转运或者储存装置17和各个单车取放点之间的车辆转运提供地下运输通道，设置在不便于从路面上设置输送装置的位置，如设置在一些交叉口处的地面下。
地下输送装置可以实现共享单车从地面到地下以及从地下到地上的运输。在一些实施例中，如图6所示，地下输送装置可以包括地下运输轨道27，设置在地上地下连接口处的压力 伸缩装置30，当压力伸缩装置30伸长至第一位置时，压力伸缩装置30的上端面与地面平齐并对接，当压力伸缩装置30压缩至第二位置时，压力伸缩装置30的上端面与地下运输轨道27平齐并对接。
可选的，压力伸缩装置30可以包括从上到下依次设置的单车承载部分31、固定连接单车承载部分31的伸缩机构以及固定平台28、以及设置在固定平台28上电连接伸缩机构的伸缩驱动动力装置29。单车承载部分31可以为承载平板或者为有轨道的承载平板，其轨道的轨道形状结构与地下运输轨道27相匹配。固定平台28提供稳定支撑。单车承载部分31上面还可以设置压力传感器32，用于检测单车承载部分31上是否放置了单车或者单车运载装置8。
具体的，伸缩驱动动力装置29可以为液压驱动装置，伸缩机构压力伸缩杆。
进一步的，该系统还可以包括控制平台，所述控制平台与流动系统中的控制模块35通信连接。
上述流动系统的工作原理为：
通过存取车交互装置，获取用户的用车需求，按开门按钮可以直接存取车，当前站点没有车时可以按取车按钮、当前站点没有停车的运载装置8可以按存车按钮；储存装置17，包括正常车辆储存和待维修车辆储存，根据系统内共享单车数量以及需求情况，对系统内车辆进行有必要的补充和储存；控制模块35，实时接收并分析系统中的运行信息，输出控制指令调度整个流动系统内部的单车和单车运载装置8；单车运载装置8，按照控制模块35的控制调度指令运送单车。另外，通过单车运载装置8判断是否规范停车，当用户不能直接存取车时，控制模块35根据各个站点的情况和用户需求，实时调度附近站点或者储存装置17的共享单车或停放的单车运载装置8，最大限度满足用户存取车需求。
实施例2
本实施例提供基于子区划分的自动调度系统，根据用户需求进行共享单车子区划分，对实施例1所述的共享单车流动系统进行自动调度，用于调度流动系统中的每个共享单车或单车运载装置8。
基于子区划分自动调度系统，如图9所示，包括实施例1所述的共享单车流动系统，以及向共享单车流动系统发送调度指令的控制平台，控制平台包括共享单车需求预测系统和动态子区划分调度系统；
共享单车需求预测系统：被配置为用于对各个站点不同时段用户需求量进行预测，获得每个单车取放点的预测需求量；给整个流动系统配送车辆数提供理论依据。
动态子区划分调度系统：被配置为用于根据获得的每个单车取放点的预测需求量，进行动态子区的划分，按照子区内部各个单车取放点的单车需求比例调度，生成调度方案发送至控制模块35，以使控制模块35控制流动系统中的单车或单车运载装置8。
实施例3
共享单车需求预测方法，基于随机森林算法融合用户需求和共享出行吸引力，能准确获得用户需求，该方法可以在控制模块连接的控制平台上实现，具体的可以由共享单车需求预测系统实现，如图10所示，包括如下步骤：
步骤1、用户需求统计：获取用户出行信息及出行预订信息，统计单车取放点的第一单车需求量X 1；
步骤2、基于共享出行吸引力的需求预测：确定单车取放点附近区域的吸引点，根据每 个吸引点的吸引力计算获得单车取放点的第二单车需求量X 2；
步骤3、基于随机森林算法计算获得单车取放点的第三单车需求量X 3；
步骤4、对上述步骤中获得的需求量加权求和，获得每个单车取放点的需求量。
步骤1中，可以通过激励反馈的方式获取用户的出行信息。激励反馈可以为采用积分激励方式，发送出行问卷，所述问卷包括主要骑行路径起终点、出行时间段以及用户意见，接收到问卷调查信息，为填写问卷的账户增加积分，为保障数据的可靠性，如果用户骑行信息与问卷填写内容严重不符，也会给用户扣除一定积分。用户主要包括持有周卡、月卡或者年卡的固定用户以及一些普通用户。
步骤2中，确定单车取放点附近区域的吸引点，根据每个吸引点的吸引力计算获得第二单车需求量的方法，包括如下步骤：
步骤21、划分吸引点的吸引等级；
吸引点为人流量比较大的公共场所，如医院、学校、公园、公交站、地铁站等。吸引点指吸引人们共享出行的地点。
可选的，可按照人流量大小进行划分，一级：公交站点、地铁站点，二级：小区、超市、学校，三级：餐饮、公园广场，四级：其他；
步骤22、确定单车取放点设定区域内的吸引点，根据吸引等级确定每个吸引点的吸引力折减系数λ x；
单车取放点设定区域如可以设置为单车取放点周围一公里的范围区域，人流量越大，折减系数越大，需求量越大，可以按照人流比例设定折减系数λ x。
步骤23、根据吸引力折减系数λ x，计算获得共享单车取放点的第二单车需求量X 2；
第二单车需求量X 2求解，可以采用如下计算公式为：
Figure PCTCN2021070687-appb-000001
其中，X 2为共享单车需求量；S 总为吸引点附近一公里区域面积；S i吸为附近某吸引点占地面积，i为第i个单车取放点；K为吸引点慢行出行比例，对于有明显时间特征和年龄特征的站点，出行比例可按年龄层次划分；N为吸引点人数；λ x为根据不同等级共享出行吸引点确定的吸引力折减系数。
步骤3、基于随机森林算法计算获得单车取放点的第三单车需求量X 3的方法，包括如下步骤：
步骤31、获取样本数据集。
取整个区域系统内所有单车取放点的历史存取车数量的数据和相对应的相关特征数据，包括地理位置、时间、季节、节假日、工作日、天气、温度、湿度、风速等特征数据，以及区域内共享单车运行轨迹以及起终点数据作为原始数据集。
步骤32、样本数据集进行样本抽取，获得多个决策树的训练子集。
采取bootsrap重抽样方法，从总样本中抽取S个训练样本子集，用于构建S个回归树，抽取的训练样本为训练集，总样本中未抽取到的样本作为测试集。
步骤33、决策树构建：基于损失最小化原则，每个训练子集对应训练获得一个决策树，决策树训练过程中，选取相关性较大的设定数量的特征变量参与决策树节点分裂，多个训练子集训练获得随机森林回归模型；
每个训练样本子集，基于损失最小化原则，生成一个决策树，S个训练样本子集共生成S棵决策树，组成随机森林，为解决因特征变量过多造成的过拟合现象，选取的特征变量设定为不超过log 2M+1,其中，M为表示相关联的特征变量的个数，对参与的特征变量根据相关性原则进行选取，根据相关性大小排序，选取相关性G较大的部分特征变量参与决策树节点分裂过程。
相关性判定方法，可以如下：
Figure PCTCN2021070687-appb-000002
其中，X为需求变量，Y i为某个特征变量，G为需求变量与某特征值的相关性，A为所有需求变量X与所有特征变量Y的数据个数的和，A i为某个特征所有数据对应A中数据的个数。
S个决策树构建完成后，利用测试集数据进行仿真，对决策树误差进行估计，优化决策树参数。将S棵决策树误差估计取平均，得到随机森林泛化误差估计值，对模型参数进行优化。
步骤34、随机森林回归模型预测结果：实时获取单车取放点的单车出行数据和对应的特征变量数据，输入至随机森林回归模型，获得各个决策树投票结果，加权获得随机森林回归预测结果即为单车取放点的第三单车需求量X 3；
随机森林回归预测模型输出的预测结果由各棵决策树投票结果产生。随机森林回归预测结果如下：
Figure PCTCN2021070687-appb-000003
其中，Y i为相关特征因素数据，H ik为单棵决策树预测模型，S为构建的总的决策树数目，X Y为共享单车需求量回归预测结果。
步骤3中，根据相关性大小限定构建决策树的特征变量数目，优化随机森林算法，从而 更精确的预测需求量X 3。
通过步骤1-3，获得整个需求预测的结果由三部分组成，对上述步骤中获得的需求量加权求和，具体如下：
X＝λ 1X 1+λ 2X 2+λ 3X 3
其中，λ 1、λ 2和λ 3代表相应的权重；X为站点总的需求量；X 1为第一单车需求量；X 2为基于共享出行吸引力获得的第二单车需求量；X 3为基于随机森林算法的预测获得的第三单车需求量。
本实施例融合用户需求、共享出行吸引力和随机森林算法，深度挖掘用户需求，从用户自身经济、便利的角度入手，提供激励反馈服务，主要挖掘一定时间内固定用户(周、月、年卡用户)用车需求，同时综合一些普通用户的预定信息，提高需求预测的精度；引入共享出行吸引力这一指标，充分考虑了站点周边吸引点对需求量造成的变动影响；可以更加便捷的分析各个站点历史数据，选用高精度的随机森林算法，通过特征变量相关性分析，限定构建决策树的特征变量个数，增加随机森林预测的准确性，从而提高需求量预测的准确性。
实施例4
动态子区划分调度方法，通过动态子区划分可以实时调整子区范围，提高调度的灵活性和时效性。该方法可以在控制模块连接的控制平台上实现，具体的可以由动态子区划分调度系统实现，如图11所示，包括如下步骤：
步骤1、获取各个单车取放点的单车存取车数据及单车的轨迹信息；
步骤2、根据获取的数据，按照不同的特征对单车取放点进行分类；
步骤3、根据分类结果，对相邻的单车取放点按照需求量的动态变化进行动态划分，形成多个子区；
步骤4、根据子区划分结果对每个子区进行调度：调度如果不能满足子区的单车需求，执行上述步骤1-3重新进行子区划分。
步骤1中，采集各个站点存取车的历史数据信息和实时的动态存取信息。采用实施例1的流动系统，可以更加方便快捷的获取每个单车取放点的数据信息，从而可以精确分析每个站点的存取车的内在特征。
步骤2中，根据获取的数据，按照不同的特征对单车取放点进行分类；
按照不同的特征对单车取放点进行分类可以包括按时间特征、按需求等级划分等。
步骤21、按照时间特征进行分类，可以划分为包括:
分时段站点：在某些特定时间段内单车需求量相对比较大，如早高峰和晚高峰时段。
全时段站点：所有时间段内需求量都相对比较大的站点；
普通站点：没有明显时间特征。
步骤22、对普通站点按需求等级划分：根据各个单车取放点预测的需求量，按照需求量等级进行划分。
步骤3、根据分类结果，对相邻的单车取放点按照需求的动态变化进行动态划分，形成多个动态子区，具体的，可以为：
步骤31、针对分时段站点和全时段站点的时间特征，结合周围站点的动态需求情况，与周围站点形成互补子区并形成相应的调度方案；
步骤32、针对普通站点进行动态子区的划分调度，可以包括如下步骤：
步骤321、对普通站点中的互补性高的站点进行合并，作为一个取放点；互补性为高需求量在时间上不重叠，高需求量的时间段是错开的。
步骤322、动态选择设定区域范围内实时的需求量最大且需求量稳定的取放点作为主站点；
取放点在地理位置上都是连续性的，根据分层取样的原理，每隔一定数量的站点选取一个主要站点，选取原则可以为：需求量变化没有明显的时间特征；在某段时间内需求量相对周围站点较高，根据实时的需求量动态选择需求最大的单车取放点；
步骤333、根据树状分支原理进行子区划分：以主站点为中心采用互补性原则和共享出行吸引力原则选取周围最合适站点进行合并，形成一个子区；
所述树状分支原理，是充分结合流动系统站点联动的特点，先选取主要站点作为根部，然后按照一根枝向上的原理生长，当不能满足生长需求时，再选择此枝中的其他节点继续生长。本实施例的系统呈带状或路网状，划分子区时，从主站点开始根据划分条件呈一条线进行划分，当不满足划分条件时再选取此条线上其他站点进行划分。
当前站点继续向下合并，不满足划分原则，则结束此站点的合并，然后检索之前合并的站点是否还存在继续合并的可能，当子区内合并的站点数达到设定数量(如8个时)结束当前子区的划分，具体的，可以如下：
计算站点一定时间内需求差值，具体方法如下：
Q＝Q 现+Q 还-Q 需
其中，Q为需求差值；Q 现为站点现有车辆数；Q 还为站点一定时间内还车数；Q 需为站点一定时间内需求车辆数。
满足互补性原则：某站点与主要站点合并后，总的需求差值Q呈减小趋势。需求互补性的判定具体方法如下：
Figure PCTCN2021070687-appb-000004
其中，K为折减系数；Q 1为站点合并之前子区的需求差值；Q 2为站点合并之后子区的需求差值；H为合并站点后的需求互补性。
满足共享出行吸引力原则：如果当前子区需求差值为负，说明供不应求，则合并站点时选择共享出行吸引力小的站点，如果需求差值为正，说明供过于求，则选取共享出行吸引力大的站点。
共享出行吸引力的判定方法具体如下：
Figure PCTCN2021070687-appb-000005
其中，A表示某站点共享出行吸引力的大小；D i代表站点到附近吸引区域内某一吸引点的距离；λ x为根据不同等级共享出行吸引点确定的吸引力折减系数；R i吸代表吸引区域内某吸引点半径；S 总代表吸引区域总面积。
步骤4中，根据子区划分结果进行调度：针对获得的子区，进行子区内部各个取放点之间的单车调整，当内部调整不能满足需求量要求，进行子区层面即子区之间的单车调整，当子区层面不能满足需求量要求，执行步骤1-4重新进行子区划分。
本实施例通过在每个子区内部先进行调度，充分考虑子区内的关联度，提高了调度的准确性，提升了调度的效率。
进行子区内部单车取放点之间的调度，具体为：按照需求比例调整单车在子区内部各个单车取放点的分布。
如果某个子区需求差值超过某一阈值，则进行子区之间的调度，若不能有效调整，则调整当前子区的划分。
当前调度方案不能满足用户需求，重新进行动态子区的划分，形成新的调度调整方案，可以实现子区划分的动态调整，提高了调度的灵活性。
本实施例的动态子区划分方法，充分结合流动系统站点联动的特点，采用树状分支原理进行点位的选取划分，划分条件充分考虑站点内、外部因素的影响，利用需求互补性原则和共享出行吸引力原则进行站点选取，既考虑周围变动因素的影响，同时也考虑每个站点不同的内在时间特征，分为全时段点位、分时段点位和普通点位，进行相应的子区划分调度，各个子区内实行需求比例调度，可以更好的满足用户调度需求。
实施例5
本实施例提供基于子区划分的自动调度方法，该方法在实施例2所述的系统中的控制平台中实现，根据用户需求进行共享单车子区划分，对实施例1所述的共享单车流动系统进行自动调度，调度流动系统中的每个共享单车或单车运载装置8。
基于子区划分的自动调度方法，包括如下步骤：
S1、获取单车取放点的单车取放数据，基于随机森林算法融合用户需求和共享出行吸引力的方法，对各个单车取放点的需求量进行预测；
S2、根据需求量的预测结果，基于树状分支原理结合内外因素，对单车取放点进行子区动态划分，根据划分结果生成调度方案；
S3、按照调度方案执行调度的控制。
步骤1采用实施例3所述的共享单车需求预测方法，步骤2采用实施例4所述动态子区划分调度方法，获得调度方案；还包括控制模块35按照调度方案执行的控制方法。
控制模块35执行的控制方法，如图12所示，包括如下：
步骤1、单车存放点之间的输送控制：按照调度方案，在相邻站点之间通过控制运载装 置8调整各个站点的单车数量；
运载装置8接收控制模块35发出的信号，按照地上输送装置和地下输送装置26，移动至指定位置。
当有取车或者停车需求时，启动运载装置8的驱动电机12，进而控制共享单车运输至相应的伸缩门4处；当系统没有存取车需求时，系统检测到所有门都关闭时，控制模块35根据每个运载装置8底部定位装置3的定位数据，实现每个站点内共享单车的等距排放；
当通过交叉口时通过地下输送装置26进行调度，共享单车以及其运载装置8运输到地下输送装置的压力传感器32时，压力传感器32将信号传给伸缩驱动动力装置29启动压力伸缩杆30，进而控制单车承载部分31将共享单车向下运送，最后压力伸缩杆全部收缩到固定平台28内，共享单车运送到地下之后，再通过地下运输轨道27进行运输，到指定地点时，再通过与向下运输相同的方法，通过压力伸缩装置30将共享单车运送到地面上，继续进行调度任务。
步骤2、存取车的存取控制：获取存取车交互装置的用户需求信息；
当用户需要取车时，接收到取车信息，控制伸缩门打开；调度距离最近的单车调度至该伸缩门位置；
当用户需要停车，接收到停车信息，控制伸缩门打开；调度距离最近的运载装置8调度至该伸缩门位置；
用户停车后，位于运载装置两端的感应装置(16)感应到有车辆放置时自动锁上，完成停车。
其中，距离最近的单车或者运载装置8，可以在当前站点，当前站点没有可以调度的单车或者运载装置8，可以在相邻站点或者在储存装置17中。
步骤3、单车的存储控制：统计子区内的单车数量数据，包括正常车辆和待维修车辆，并将正常车辆和维修车辆分开放置；
当数量不满足需求量要求时，向该子区的储存装置17补充单车；
当单车数量大于需求量要求，将该子区的储存装置中多余的单车调度至需求量没有满足的子区。
为便于维修人员维修，若车辆需要维修，则将其储存到储存装置的第一存储层20的位置上，当需维修的车辆数超过此储存装置17的总储存车辆数的设定数值，发送维修指令至维修人员的终端。
设置了第一存储层20、第二存储层21和第三存储层22。
当车辆不需维修时，共享单车随着螺旋式上升轨道23向上运输储存，储存时，先通过螺旋式上升轨道23向上运输，在经过第二存储层21时，通过第二存储层21与螺旋式上升轨道23相连接的出入口24进入第二层，共享单车存放、固定装置是围绕着中心旋转一周排列开的，通过不断的旋转储存从螺旋式上升轨道23到达的车辆，当储存满第二存储层21时，然后再相继储存第一存储层20、第三存储层22；当从储存装置17调度共享单车时，先调度第一存储层20处的车辆，当底层车辆不够调度时，再依次调度上层储存装置的共享单车。
系统根据实时的用户需求，按照上述步骤方法及时进行调度调整，最大限度满足用户需求。
以上所述仅为本公开的优选实施例而已，并不用于限制本公开，对于本领域的技术人员来说，本公开可以有各种更改和变化。凡在本公开的精神和原则之内，所作的任何修改、等 同替换、改进等，均应包含在本公开的保护范围之内。
上述虽然结合附图对本公开的具体实施方式进行了描述，但并非对本公开保护范围的限制，所属领域技术人员应该明白，在本公开的技术方案的基础上，本领域技术人员不需要付出创造性劳动即可做出的各种修改或变形仍在本公开的保护范围以内。

Claims (10)

  1. 共享单车流动系统，其特征是：包括设置在各个单车取放点的地上输送装置，连接各个单车取放点的地上输送装置的地下输送装置,以及能够为地上输送装置或者地下输送装置提供单车的多层储存装置；相邻的单车取放点、地上储存装置通过地上运送装置或者地下运送装置连接，形成共享单车的流动运送网络。
  2. 如权利要求1所述的共享单车流动系统，其特征是：储存装置包括构成单车容纳空间的装置壳体，设置壳体内的多个存储层，以及能够将单车在各个存储层之间移动运输的运输轨道，运输轨道在每一存储层分别设置运输轨道通向该存储层的入口和出口；
    或者
    储存装置的运输轨道包括设置在壳体内的立柱及固定在立柱上的螺旋式上升轨道。
  3. 如权利要求1所述的共享单车流动系统，其特征是：流动系统还包括相对于各个装置自由移动的单车运载装置，各个单车存放点还设置有控制模块，单车运载装置包括底板，设置在底板上的驱动底板移动的驱动装置和单车固定装置，单车信息识别装置，主控制器以及无线通信模块，主控制器分别与驱动装置、单车固定装置、单车信息识别装置以及无线通信模块分别连接，主控制器与控制模块无线连接；
    或者
    单车固定装置包括固定在底板上的感应式智能锁和感应装置，所述感应智能锁和感应装置分别与主控制器电连接，用于将感应信息传输至控制模块；
    或者
    信息识别装置包括设置在共享单车上的干簧管，设置在单车运载装置的底板上与干簧管位置相对的伸缩装置，设置在伸缩装置顶端的磁铁，所述伸缩装置与主控制器电连接，干簧管与主控制器无线连接。
  4. 如权利要求1所述的共享单车流动系统，其特征是：地上输送装置包括控制模块和铺设在地面的地面运输轨道，设置在地面运输轨道两侧的防护栏，以及设置在防护栏上提供单车出入通道的单车存取口，以及设置在单车存取口处的用于接收交互信息的存取车交互装置，控制模块分别通过通信模块与单车运载装置和存取车交互装置通信连接；
    或者
    存取车交互装置包括单车存取口处设置的自动伸缩门，以及用于控制伸缩门的开关的按钮，可以包括开门按钮、取车按钮和存车按钮；
    或者
    地下输送装置包括地下运输轨道，设置在地上地下连接口处的压力伸缩装置，当压力伸缩装置伸长至第一位置时，压力伸缩装置的上端面与地面平齐并对接，当压力伸缩装置压缩至第二位置时，压力伸缩装置的上端面与地下运输轨道平齐并对接；
    压力伸缩装置包括从上到下依次设置的单车承载部分、固定连接单车承载部分的伸缩机构以及固定平台、以及设置在固定平台上电连接伸缩机构的伸缩驱动动力装置。
  5. 基于子区划分的自动调度方法，其特征是，包括如下步骤：
    获取单车取放点的单车取放数据，基于随机森林算法融合用户需求和共享出行吸引力的方法，对各个单车取放点的需求量进行预测；
    根据需求量的预测结果，基于树状分支结合内外因素，对单车取放点进行子区动态划分，根据划分结果生成调度方案；
    按照调度方案执行调度的控制。
  6. 如权利要求5所述的基于子区划分的自动调度方法，其特征是：基于随机森林算法融合用户需求和共享出行吸引力的方法，对各个单车取放点的需求量进行预测，即为共享单车需求预测方法，包括如下步骤；
    获取用户出行信息及出行预订信息，统计单车取放点的第一单车需求量；
    确定单车取放点附近区域的吸引点，根据每个吸引点的吸引力计算获得单车取放点的第二单车需求量；
    基于随机森林算法计算获得单车取放点的第三单车需求量；
    对上述步骤中获得的需求量加权求和，获得每个单车取放点的需求量。
  7. 如权利要求6所述的基于子区划分的自动调度方法，其特征是：共享单车需求预测方法中，确定单车取放点附近区域的吸引点，根据每个吸引点的吸引力计算获得第二单车需求量的方法，包括如下步骤：
    划分吸引点的吸引等级；
    确定单车取放点设定区域内的吸引点，根据吸引等级确定每个吸引点的吸引力折减系数；
    根据吸引力折减系数，计算获得共享单车取放点的第二单车需求量；
    或者
    基于随机森林算法计算获得单车取放点的第三单车需求量的方法，包括如下步骤：
    获取样本数据集；
    样本数据集进行样本抽取，获得多个决策树的训练子集；
    决策树构建：基于最小化原则，每个训练子集对应训练获得一个决策树，决策树训练过程中，选取相关性较大的设定数量的特征变量参与决策树节点分裂，多个训练子集训练获得随机森林回归模型；
    随机森林回归模型预测结果：实时获取单车取放点的单车出行数据和对应的特征变量数据，输入至随机森林回归模型，获得各个决策树投票结果，加权获得随机森林回归预测结果即为单车取放点的第三单车需求量。
  8. 如权利要求5所述的基于子区划分的自动调度方法，其特征是：
    按照调度方案执行调度的控制包括单车存放点之间的输送控制、存取车的存取控制以及单车的存储控制；
    或者
    基于树状分支结合内外因素，对单车取放点进行子区动态划分，根据划分结果生成调度方案的方法，即为动态子区划分调度方法，包括如下步骤：
    获取各个单车取放点的单车存取车数据及单车的轨迹信息；
    根据获取的数据，按照不同的特征对单车取放点进行分类；
    根据分类结果，对相邻的单车取放点按照需求量的动态变化进行动态划分，形成多个子区；
    根据子区划分结果进行调度：调度如果不能满足子区的单车需求，执行上述方法重新进行子区划分。
  9. 如权利要求8所述的基于子区划分的自动调度方法，其特征是：所述动态子区划分调度方法中，按照不同的特征对单车取放点进行分类可以包括按时间特征、按需求等级划分，获得分时段站点、全时段站点和普通站点；
    或者
    根据分类结果，对相邻的单车取放点按照需求的动态变化进行动态划分，形成多个动态子区，具体为：
    针对分时段站点和全时段站点的时间特征，结合周围站点的动态需求情况，与周围站点形成子区；
    对普通站点中的互补性高的站点进行合并，作为一个取放点；
    动态选择设定区域范围内实时的需求量最大且需求量稳定的取放点作为主站点；
    根据树状分支原理进行子区划分：以主站点为中心采用互补性原则和共享出行吸引力原则选取周围最合适站点进行合并，形成一个子区；
    或者
    根据子区划分结果进行调度，具体为：针对获得的子区，进行子区内部各个取放点之间的单车调整，当内部调整不能满足需求量要求，进行子区层面即子区之间的单车调整，当子区层面不能满足需求量要求，重新进行子区划分。
  10. 基于子区划分的自动调度系统，其特征是：包括如权利要求1-4任一项所述的共享单车流动系统，以及向共享单车流动系统发送调度指令的控制平台；控制平台包括共享单车需求预测系统和动态子区划分调度系统；
    共享单车需求预测系统被配置为用于执行如权利要求5-9任一项所述的基于子区划分的自动调度方法中的共享单车需求预测方法；
    或者，动态子区划分调度系统被配置为用于执行如权利要求5-9任一项所述的基于子区划分的自动调度方法中的动态子区划分调度方法。
PCT/CN2021/070687 2020-05-26 2021-01-07 共享单车流动系统、基于子区划分的自动调度系统及方法 WO2021238231A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010455032.6 2020-05-26
CN202010455032.6A CN111598481B (zh) 2020-05-26 2020-05-26 共享单车流动系统、基于子区划分的自动调度系统及方法

Publications (1)

Publication Number Publication Date
WO2021238231A1 true WO2021238231A1 (zh) 2021-12-02

Family

ID=72188712

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/070687 WO2021238231A1 (zh) 2020-05-26 2021-01-07 共享单车流动系统、基于子区划分的自动调度系统及方法

Country Status (2)

Country Link
CN (1) CN111598481B (zh)
WO (1) WO2021238231A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114596667A (zh) * 2022-03-08 2022-06-07 北京物资学院 一种数据共享方法、系统及计算机存储介质
CN114626766A (zh) * 2022-05-12 2022-06-14 深圳市伟创高科电子有限公司 基于大数据的共享电动车调度方法、装置、设备及介质
CN114841610A (zh) * 2022-05-26 2022-08-02 北京理工大学 共享单车调度方法、装置、设备及可读存储介质
CN115936409A (zh) * 2023-02-27 2023-04-07 北京阿帕科蓝科技有限公司 工单信息的生成方法、装置、计算机设备
CN116611565A (zh) * 2023-05-23 2023-08-18 岭南师范学院 基于数据分析的用户行为监控系统
CN117808377A (zh) * 2024-03-01 2024-04-02 交通运输部公路科学研究所 一种基于指标折减及样本量折减的公路工程质量评估系统

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111598481B (zh) * 2020-05-26 2023-09-01 山东交通学院 共享单车流动系统、基于子区划分的自动调度系统及方法
CN112836951B (zh) * 2021-01-26 2023-10-24 深圳市泰比特科技有限公司 一种基于大数据的共享单车云平台智能调度方法及系统
CN112766591A (zh) * 2021-01-27 2021-05-07 同济大学 共享单车调度方法
CN115936240B (zh) * 2022-12-23 2024-03-15 暨南大学 共享单车需求预测与投放调度方法
CN116824861B (zh) * 2023-08-24 2023-12-05 北京亦庄智能城市研究院集团有限公司 基于城市大脑平台多维数据的共享单车调度方法和系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108038578A (zh) * 2017-12-28 2018-05-15 东南大学 基于需求预测和中心辐射网络的公共自行车静态调度方法
CN108416513A (zh) * 2018-02-12 2018-08-17 杭州电子科技大学 一种基于混合启发式算法的公共自行车动态调度方法
CN108460537A (zh) * 2018-03-23 2018-08-28 华南理工大学 一种基于大数据与用户心理预测共享自行车需求量的方法
CN109208961A (zh) * 2018-10-15 2019-01-15 上海市城市建设设计研究总院（集团）有限公司 共享单车自动运输系统及其使用方法
US20190279235A1 (en) * 2018-03-12 2019-09-12 Toyota Jidosha Kabushiki Kaisha Shared vehicle management server and non-transitory storage medium storing shared vehicle management program
CN111598481A (zh) * 2020-05-26 2020-08-28 山东交通学院 共享单车流动系统、基于子区划分的自动调度系统及方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10460375B2 (en) * 2013-12-06 2019-10-29 International Business Machines Corporation Device and method for implementing a vehicle sharing reward program
CN107767659B (zh) * 2017-10-13 2021-01-26 东南大学 基于arima模型的共享单车吸引量和发生量预测方法
CN109102590B (zh) * 2018-08-27 2020-11-10 浪潮电子信息产业股份有限公司 一种基于云计算的地下共享单车停车场系统及运行方法
CN109555340A (zh) * 2018-11-01 2019-04-02 齐齐哈尔大学 共享单车智能地下停放系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108038578A (zh) * 2017-12-28 2018-05-15 东南大学 基于需求预测和中心辐射网络的公共自行车静态调度方法
CN108416513A (zh) * 2018-02-12 2018-08-17 杭州电子科技大学 一种基于混合启发式算法的公共自行车动态调度方法
US20190279235A1 (en) * 2018-03-12 2019-09-12 Toyota Jidosha Kabushiki Kaisha Shared vehicle management server and non-transitory storage medium storing shared vehicle management program
CN108460537A (zh) * 2018-03-23 2018-08-28 华南理工大学 一种基于大数据与用户心理预测共享自行车需求量的方法
CN109208961A (zh) * 2018-10-15 2019-01-15 上海市城市建设设计研究总院（集团）有限公司 共享单车自动运输系统及其使用方法
CN111598481A (zh) * 2020-05-26 2020-08-28 山东交通学院 共享单车流动系统、基于子区划分的自动调度系统及方法

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114596667A (zh) * 2022-03-08 2022-06-07 北京物资学院 一种数据共享方法、系统及计算机存储介质
CN114596667B (zh) * 2022-03-08 2023-08-11 北京物资学院 一种数据共享方法、系统及计算机存储介质
CN114626766A (zh) * 2022-05-12 2022-06-14 深圳市伟创高科电子有限公司 基于大数据的共享电动车调度方法、装置、设备及介质
CN114841610A (zh) * 2022-05-26 2022-08-02 北京理工大学 共享单车调度方法、装置、设备及可读存储介质
CN115936409A (zh) * 2023-02-27 2023-04-07 北京阿帕科蓝科技有限公司 工单信息的生成方法、装置、计算机设备
CN116611565A (zh) * 2023-05-23 2023-08-18 岭南师范学院 基于数据分析的用户行为监控系统
CN116611565B (zh) * 2023-05-23 2023-12-29 岭南师范学院 基于数据分析的用户行为监控系统
CN117808377A (zh) * 2024-03-01 2024-04-02 交通运输部公路科学研究所 一种基于指标折减及样本量折减的公路工程质量评估系统
CN117808377B (zh) * 2024-03-01 2024-04-26 交通运输部公路科学研究所 一种基于指标折减及样本量折减的公路工程质量评估系统

Also Published As

Publication number Publication date
CN111598481B (zh) 2023-09-01
CN111598481A (zh) 2020-08-28

Similar Documents

Publication Publication Date Title
WO2021238231A1 (zh) 共享单车流动系统、基于子区划分的自动调度系统及方法
CN104260724B (zh) 一种车辆智能预测控制系统及其方法
CN105303882B (zh) 基于时空维度动态停车收费策略的停车管理系统及方法
CN109584603A (zh) 一种停车位智能识别和引导方法及系统
CN107230385A (zh) 共享车位方法及系统、车位锁
CN108716173A (zh) 一种自动驾驶汽车专用道路及其使用方法
CN103592926A (zh) 基于机械式立体停车场与agv相结合的智能控制系统及方法
CN108149980A (zh) 充电桩、停车库以及充电控制方法和装置
CN101944205A (zh) 一种工厂物料配送车辆调度系统
CN113033959A (zh) 基于tod发展模式的轨道交通站点周边建成环境评价方法
CA3088914C (en) Navigation method and system for electric vehicles based on electricity quantity guidance of energy-storage charging piles
CN110910671B (zh) 一种机器人停车场车位动态规划方法
CN103150924A (zh) 地下车位引导停车系统
Zhu et al. Solar photovoltaic generation for charging shared electric scooters
CN110458456A (zh) 基于人工智能的需求响应式公交系统的调度方法及系统
CN112233443A (zh) 一种基于agv的无人停车场自动泊车管理系统
CN102800212A (zh) 一种物联网环境下的智能泊车系统
CN101038685A (zh) 城市自行车共用管理系统
CN109523064A (zh) 一种基于多网融合的智能微枢纽
CN113192357A (zh) 停车场管理方法、边缘服务器及区域管理服务器
CN206363535U (zh) 一种高速公路主动发光轮廓标交通诱导装置
WO2023143647A2 (zh) 区域电动非机动车能源控制系统
CN110758133B (zh) 自动驾驶车辆智能充电方法
Newman et al. Introducing the 21st Century Boulevard: A post-COVID response to urban regeneration of main road corridors
Yang et al. Research on the Construction Framework of Smart Park: A Case Study of Intelligent Renovation of Beijing Haidian Park

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: 21812560

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 24.03.2023)

122 Ep: pct application non-entry in european phase

Ref document number: 21812560

Country of ref document: EP

Kind code of ref document: A1