CN110146102B - Path planning method, device, equipment and storage medium - Google Patents
Path planning method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN110146102B CN110146102B CN201910482413.0A CN201910482413A CN110146102B CN 110146102 B CN110146102 B CN 110146102B CN 201910482413 A CN201910482413 A CN 201910482413A CN 110146102 B CN110146102 B CN 110146102B
- Authority
- CN
- China
- Prior art keywords
- task
- sequence
- edge
- served
- matrix
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3407—Route searching; Route guidance specially adapted for specific applications
- G01C21/343—Calculating itineraries, i.e. routes leading from a starting point to a series of categorical destinations using a global route restraint, round trips, touristic trips
Abstract
The invention discloses a path planning method, a path planning device, a path planning equipment and a storage medium. The method comprises the following steps: determining a task edge to be served according to the service request; determining a task matrix based on the task edge to be served; and determining a planning path according to the task matrix and a preset path planning model. According to the technical scheme of the embodiment of the invention, the path planning efficiency can be improved by taking the task side as the minimum unit for providing the service, the conveying efficiency of the unmanned express delivery vehicle is improved, and the calculation cost can be reduced.
Description
Technical Field
The embodiment of the invention relates to the technical field of computer application, in particular to a path planning method, a path planning device, a path planning equipment and a storage medium.
Background
Along with the development of modern life, the appearance of express delivery makes human life more convenient, and unmanned express delivery car becomes the research focus in the express delivery field at present, and unmanned express delivery car need go to customer's department automatically and receive and dispatch the express delivery, but because unmanned express delivery car capacity is fixed, once fill or empty, unmanned express delivery car need return the warehouse and load and unload the express delivery, in order to make unmanned express delivery car reach best transport efficiency, every street only serves once and total distance of traveling is minimum, need plan the service route of unmanned express delivery car.
In order to solve the above problems, in the prior art, each street to be served is taken as a unit for path planning, and since there are many serving streets and the scale of path planning is very large, the path planning process takes too long, and huge computing resources need to be consumed.
Disclosure of Invention
The invention provides a path planning method, a path planning device and a storage medium, which are used for realizing efficient planning of paths and improving service efficiency.
In a first aspect, an embodiment of the present invention provides a path planning method, including:
determining a task edge to be served according to the service request;
determining a task matrix based on the task edge to be served;
and determining a planning path according to the task matrix and a preset path planning model.
In a second aspect, an embodiment of the present invention provides a path planning apparatus, including:
the task edge determining module is used for determining the edge of the task to be served according to the service request;
the matrix determining module is used for determining a task matrix based on the task edge to be served;
and the path generation module is used for determining a planned path according to the task matrix and a preset path planning model.
In a third aspect, an embodiment of the present invention provides an apparatus, including:
one or more processors;
a memory for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement a path planning method as in any of the embodiments of the invention.
In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the path planning method according to any embodiment of the present invention.
According to the technical scheme of the embodiment of the invention, the task edge to be served is determined according to the service request, the task matrix is generated based on the task edge to be served, the planning path is determined according to the task matrix and the preset path planning model, and the task edge is taken as the minimum unit for providing the service, so that the path planning efficiency is improved, the conveying efficiency of the unmanned express delivery vehicle for providing the service is improved, and the calculation cost and the cost are reduced.
Drawings
Fig. 1 is a flowchart of a path planning method according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating a graph structure transformation according to an embodiment of the present invention;
fig. 3 is a flowchart of a path planning method according to a second embodiment of the present invention;
fig. 4 is a diagram illustrating an example of a path planning method according to a second embodiment of the present invention;
fig. 5 is a schematic structural diagram of a path planning apparatus according to a third embodiment of the present invention;
fig. 6 is a schematic structural diagram of an apparatus according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be noted that, for convenience of description, only a part of the structures related to the present invention, not all of the structures, are shown in the drawings, and furthermore, embodiments of the present invention and features of the embodiments may be combined with each other without conflict.
Example one
Fig. 1 is a flowchart of a path planning method according to an embodiment of the present invention, where the present embodiment is applicable to a situation where an unmanned express delivery vehicle provides a service to a door, and the method may be executed by a path planning device, and the device may be implemented in a hardware and/or software manner, where the method according to the embodiment of the present invention specifically includes the following steps:
The service request may be a request sent when a user needs a service, the service request may include a service location and a workload, for example, the location where the user needs to send an express and the weight of the express, the task edge to be served may be a minimum unit for providing the service, and may have attributes such as a length, a task amount, and a location, a vertex in the task edge to be served may be a warehouse, that is, the warehouse may be located in the task edge to be served, the task edge to be served may be an actual street edge, or may be a virtual edge determined by the service request, for example, a task edge to be served may be formed by connecting places corresponding to locations closest to the service request.
Specifically, the service request may be obtained in a manner of obtaining the service request at a fixed time interval, for example, obtaining the service request every 24 hours, extracting a position and a workload in the service request, corresponding the service request to a street according to a mapping relationship between the position and a street in an actual map, changing a task amount of the street according to the task amount, and using the street mapped with the service request as a task edge to be served; the service requests can be connected into the task edge to be served according to the positions of the requests.
And 102, determining a task matrix based on the task edge to be served.
The task matrix can be a matrix used for representing the edge relation and the attribute of the task to be served, and the task matrix can be generated by the task to be served through the graph convolution neural network, so that the information of the task to be served is stored through the task matrix.
Specifically, the to-be-serviced task edges may be connected to generate a to-be-serviced graph, because the information of the to-be-serviced task edges exists in the edges of the to-be-serviced graph, in order to keep the information of the to-be-serviced task edges through a graph convolution neural network, the to-be-serviced graph needs to be converted, fig. 2 is a diagram of graph structure conversion provided in an embodiment of the present invention, referring to fig. 2, the edges of the to-be-serviced graph may be converted into vertices, if the edges in the to-be-serviced graph share a vertex with other edges, the vertices in the converted graph may be correspondingly connected to generate edges, in a process of converting the to-be-serviced graph, since a warehouse exists in the to-be-serviced graph in the form of a vertex, in a process of converting an edge into a point, a warehouse existing in the form of a point may lose an attribute carried by itself, a virtual edge may be constructed for the warehouse in the to-be-serviced graph, that is, it may be considered that a, after the graph to be served is converted, the converted graph to be served can be used as the input of the graph convolution neural network, the output result can be used as a task matrix, and it can be understood that the edge of the task to be served can be directly used as a vertex without actually connecting the edge of the task to be served, and the length, the position and the like of the edge of the task to be served can be used as the attributes of the vertex.
And 103, determining a planning path according to the task matrix and a preset path planning model.
The path planning model can preset a neural network model, can take a task matrix as input, and can output a sequence containing task edges to be served as a planning path, the preset path planning model can comprise a coding sub-model and a decoding sub-model, the coding sub-model can process the task matrix to obtain high-level attribute information related to the task edges to be served, such as the distance between the task edges to be served and a warehouse, the distance between the task edges to be served, the correlation degree between the task edges to be served, the length of the task edges to be served, the road surface condition, the oil consumption, the speed of service processing and other attributes, and the decoding layer can calculate based on the obtained attribute information to obtain the feasibility probabilities corresponding to different sequence combinations of the task edges served by the unmanned express delivery vehicle.
In the embodiment of the present invention, a task matrix may be used as an input of a path planning model, the path planning model may be processed and calculated based on the task matrix, service probabilities may be sequentially calculated according to task vectors corresponding to task edges to be served in each task matrix, and a sequence of servicing the task edges to be served may be determined based on the probabilities so as to determine a planned path, for example, task edges to be served A, B, C and D, the path planning model first determines that the probability that a first edge to be serviced is A, B, C and the probability that the first edge to be serviced is 0.1, 0.2, and 0.5, respectively, and then the first edge to be serviced is task edge D; then the path planning model determines that the probability that the second task to be served is A, B and the probability of C are 0.3, 0.4 and 0.3 respectively, and the task edge of the second service is B; then, the path planning model determines a third edge needing service, wherein the probabilities of A and C are 0.3 and 0.7 respectively, so that the third edge needing service is C; finally, edge A is served. Therefore, the output sequence of the tasks to be served in the path planning model can be D, B, C, A in sequence, and the tasks to be served can be connected in the map according to the output sequence of the tasks to be served to generate the planned path. It is worth noting here that the above example assumes that the truck could load A, B, C and all the cargo on task side D at once, but in practice this is not guaranteed. Therefore, when the amount of cargo in the remaining task edge is larger than the remaining cargo space of the unmanned vehicle, the path planning model predicts that the probability of returning to the warehouse for unloading is the largest (in practice, the warehouse is assumed to be a task edge E with a length of 0, and which can be serviced for multiple times, and the task edge E is added to the above-mentioned sequence of the edge to be serviced). When all the task edges not in the warehouse are in the service order, the path planning model is finished.
According to the technical scheme of the embodiment of the invention, the task side to be served is determined according to the service request, the task matrix is determined based on the task side to be served, and the planning path is determined based on the task matrix and the preset path planning model, so that the task side becomes the minimum unit for carrying out service, the unmanned express delivery vehicle can carry out service in a task side mode, the path planning efficiency is improved, the calculation cost is reduced, and the conveying efficiency of the unmanned express delivery vehicle can be improved.
Example two
Fig. 3 is a flowchart of a path planning method provided in the second embodiment of the present invention, and the second embodiment of the present invention is optimization based on the above embodiments, where the path planning method in the second embodiment of the present invention includes:
The request location may be a location where the user requests to perform service, for example, may be a receiving location or a receiving location of express delivery, and may be generated according to order information in the service request.
In the embodiment of the invention, the service request can be acquired regularly, the request position needing service in the service request can be extracted, the request position can be the receiving position information or the delivery position information of express delivery, and the specific form of the request position information can be latitude and longitude data or address information.
The street to be served may be actually existing street information, for example, northwest wang east road in hai lake area of beijing, the street to be served may be stored in the form of a map, and the street to be served may include location information and a workload required to perform a service.
Specifically, the request position may be mapped to the street to be serviced according to the specific position information of the request position, and if the request position is north-west highway 10 of the haihui district of beijing city, the request position may be mapped to the street to be serviced of the north-west highway of the haihui district of beijing city according to the specific address information, and the street to be serviced to which the service request is mapped may be used as a task edge to be serviced.
In this embodiment of the present invention, determining the task edge to be serviced according to the service request may further include: extracting a request position of the service request; and connecting the service requests within the threshold distance according to the request position to generate a task edge to be served.
The threshold range may be a maximum distance for connecting the service requests, and the service requests within the threshold range may be sequentially connected into one edge according to the request position.
Specifically, the service requests may be acquired periodically, the request positions that need to be serviced in the service requests may be extracted, the request positions may be express delivery position information or delivery position information, the specific form of the request position information may be latitude and longitude data, or may be address information, after the request positions of the service requests are acquired, task edges to be serviced may be generated according to the request positions and the service requests within a threshold range, it may be understood that the service requests may be classified according to the request positions based on the threshold range, and then the service requests within each classification may be sequentially connected to generate task edges to be serviced.
And 203, generating a structural matrix according to the distance information of the task edge to be served.
The distance information may be distance information between the task sides to be served, for example, the distance between the task side a to be served and the task side B to be served may be a distance between the task side a to be served and the task side B to be served, since the task side to be served is an actual street and is not a location point, the distance between the task side a to be served and the task side B to be served may be determined by distances between four end points a1, a2, B1 and B2 of two task sides to be served, and the calculation formula may be
Where dc (x, y) represents the distance between x and y, the structural matrix may be a matrix storing information about the distance between task edges to be served, and the elements in the structural matrix represent information about the distance between two task edges corresponding to the positions of the elements, for example, element a23 in the structural matrix may beIs the distance information between the second to-be-serviced task edge and the third to-be-serviced task edge.
Specifically, the distances between the task sides to be served can be calculated pairwise according to the appearance sequence, the distance calculation method can be used for calculating the distances between the end points of the task sides to be served, the calculated distances can be stored in a matrix form to generate a structural matrix, and it can be understood that the method for calculating the distances between the task sides to be served is not limited to calculating the distances between the end points, and fixed position points can be randomly selected from the task sides to be served, and the distances between the position points can be calculated respectively.
And 204, generating a characteristic matrix according to the attribute information of the task edge to be served.
The attribute information may be information for representing attributes of the to-be-served task, for example, attribute information such as traveling time of the unmanned express vehicle at the to-be-served task, road condition parameters of the to-be-served task, oil consumption of the unmanned express vehicle at the to-be-served task, time of loading and unloading goods at the to-be-served task, task amount at the to-be-served task, and current capacity of the unmanned express vehicle at the to-be-served task may be included in the attribute matrix.
Specifically, attribute information such as the traveling time of the unmanned express vehicle at the side of the task to be serviced, road condition parameters at the side of the task to be serviced, oil consumption of the unmanned express vehicle at the side of the task to be serviced, time of loading and unloading goods at the side of the task to be serviced, task quantity at the side of the task to be serviced, and current capacity of the unmanned express vehicle at the side of the task to be serviced may be sequentially and correspondingly stored in the feature matrix, for example, for the side e1 of the task to be serviced, a parameter of the side e1 of the task to be serviced in the feature matrix may be represented as
Wherein u is1(e1)、u2(e1) Respectively representing the two end points of the edge e1, if endPoint is warehouse, then f (u) is 1, otherwise f (u) is 0, v0 is warehouse, dc (u, v0) may be the distance from the task edge to warehouse, d (e1) may be the task amount of the task edge to be served,
the ratio of the task amount of the task edge to be served e1 to the task amount of all the tasks to be served can be determined, and Q can be the total capacity of the unmanned express delivery vehicle.
And step 205, inputting the structural matrix and the feature matrix into a preset atlas neural network to generate a task matrix.
The graph convolution neural network may be a neural network that generates structure matrices and feature matrices into embedded vectors corresponding to each task edge to be serviced, and further, the graph convolution neural network may include three convolution layers, and the K-th convolution layer is calculated by HK＝σK(CHK-1WK) Wherein H isKIs the output of the Kth convolutional layer, WKIs the parameter of the Kth convolution layer, when K is 1, H1May be a feature matrix.
Specifically, a graph convolution neural network having three convolution layers may be generated in advance, a structure matrix and a feature matrix may be used as input of the graph convolution neural network, and embedded vectors corresponding to output edges of each task to be served may be merged to serve as a task matrix, where the embedded vectors may be vectors generated by compressing information corresponding to edges of the task to be served and relationship information with edges of other tasks to be served.
And step 206, inputting each task vector in the task matrix into a sequence determination unit of a path planning model, and sequentially arranging output results of the sequence determination unit to form a task sequence, wherein each task vector corresponds to a task edge to be served.
The sequence determining unit may be a neural network, specifically, a recurrent neural network, configured to generate a to-be-serviced task edge service sequence, where the sequence determining unit may have two hidden layers, may determine, based on an input task matrix, a service sequence corresponding to each to-be-serviced task edge, and an output result may be a service sequence of the to-be-serviced task edges, and the task sequence may be the to-be-serviced task edges sorted according to the service sequence, for example, the to-be-serviced task edges may be A, B, C and D, and the task sequence may be B, C, A, D.
Specifically, each task vector of the task matrix may be input into a sequence determining unit of the path planning model, each task vector corresponds to each task edge to be served, output results of the sequence determining unit may be arranged according to an output sequence to generate a task sequence, and the unmanned express delivery vehicle may sequentially serve according to the sequence of the task edges to be served in the task sequence.
And step 207, inputting the task sequence into a direction determining unit of a path planning model, and acquiring the direction sequence output by the direction determining unit.
The direction determining unit may specifically include two hidden layers, a two-dimensional probability vector is output at each step to represent corresponding probabilities in two directions of the task edge to be served, and a direction with the highest probability is selected as a direction corresponding to the task edge to be served; the direction sequence may be a set representing the access directions of the tasks to be served, and in the direction sequence, 0 may be used to represent the access sequence from the first endpoint to the second endpoint of the task to be served, and 1 may be used to represent the access sequence from the second endpoint to the first endpoint of the task to be served.
Specifically, the task sequence may be input into a direction determining unit of the path planning model, the direction determining unit determines the access direction of each to-be-served task edge in the task sequence, the probabilities of the to-be-served edges in each access direction are respectively calculated, the access direction with the highest probability may be selected as the access direction corresponding to the to-be-served task edge, if the task sequence is (B, C, D, a), the direction sequence determined by the direction determining unit may be (1, 0, 1, 0), and the access directions in the direction sequence respectively correspond to the to-be-served task edges in the task sequence.
And step 208, combining the task sequence and the direction sequence to generate a planning path.
The planned path can be generated according to the task sequence and the direction sequence when the unmanned express delivery vehicle serves the route of each task side to be served.
Specifically, the sequence and the access direction of the task sides of the unmanned express delivery vehicle serving the service to be served can be determined according to the task sequence and the direction sequence, the task sides to be served are connected according to the sequence and the direction, a route generated by connection can be used as a planned route, and further, a method for determining the route between the task sides to be served can be obtained by using a traditional Dijkstra algorithm.
The technical scheme of the embodiment of the invention comprises the steps of extracting a request position in a service request, mapping the service request to a corresponding street to be served according to the request position, taking the street to be served as a task side to be served, generating a structure matrix according to distance information of the task side to be served, generating a feature matrix according to attribute information of the task side to be served, inputting the structure matrix and the feature matrix into a preset graph convolution neural network to generate a task matrix, inputting a task vector in the task matrix into a sequence determination unit of a path planning model, sequentially arranging output results to generate a task sequence, inputting the task sequence into a direction determination unit to obtain the direction sequence, generating a planning path based on the task sequence and the direction sequence, enabling the task side to be a minimum unit for providing service, improving task planning efficiency, reducing calculation cost of path planning, and effectively improving the conveying efficiency of the unmanned express delivery vehicle, the cost of the delivery service can be reduced.
On the basis of the above embodiment, the sequence determination unit determines the output result by:
determining a probability distribution value of each task vector in the task matrix based on a set probability calculation formula; determining a target task edge corresponding to the maximum probability distribution value; if the task quantity of the target task edge is smaller than a load threshold, taking the target task edge as an output result, obtaining a task matrix after shielding the task vector corresponding to the target task edge, and returning to execute the calculation operation of probability distribution until all task vectors are shielded; otherwise, the task matrix after shielding the task vector corresponding to the target task edge is obtained and returns to execute the calculation operation of probability distribution until all the task vectors are shielded.
The set probability calculation formula may be a formula of probability distribution values of task vectors corresponding to the task sides to be served, the probability distribution values may be probabilities of the unmanned express delivery vehicle accessing the corresponding task sides to be served, the probability distribution values may be calculated according to the probability calculation formula, the target task sides may be task sides determined according to the probability distribution values from the task sides to be served, the task quantities may be the number of services to be provided on the target task sides, for example, the weight or capacity of the express delivery, the compliance threshold may be the maximum load weight or capacity of the unmanned express delivery vehicle, the shielding may be operations for representing that the sequence determination unit does not consider when determining the target task sides, and the operations may include removing the task vectors corresponding to the target task sides from a task matrix.
Specifically, an encoding matrix generated by performing an encoding operation on task vectors may be generated first, a context state may be generated based on the encoding, a corresponding probability distribution value may be determined by the encoding matrix, the context state, and a set probability calculation formula, a side to be serviced with the largest probability distribution value may be used as a target task side, a task amount stored in association with the target task side may be obtained, the task amount is compared with a load threshold, if the task amount is smaller than the load threshold, the target task side is output as a result, a shielding operation is performed on the target task side, a calculation operation of probability distribution is returned to be performed until the side to be serviced corresponding to each task vector is shielded, if the task amount is not smaller than the load threshold, the target task side may be shielded, a calculation operation of probability distribution is returned to be performed, and the calculation operation of probability distribution is performed as follows:
firstly, encoding task vectors to generate an encoding matrix XtAnd ehtFor example, task matrices
Is a dynamic feature of the task edge to be serviced, DtThe 1 st column in the list indicates the remaining capacity of the current unmanned express delivery vehicle (each row is the same, the initial value is 1, namely 100%), and the 2 nd column indicates the task amount of the current task side (all are less than 1, 0 indicates that no service is needed, and actually the proportion of the capacity of the unmanned express delivery vehicle needs to be occupied). D1Is initialized according to the side information of the task to be served. Dt＞1The value of (a) is obtained by adjusting according to the output of the sequence determination unit, if the output of the sequence determination unit is 1, the value indicates the task edge e to be served1Not needing to be served any more, column 1 all values minus the task amount d (e) of the task edge to be served1) Column 2, row 2 becomes 0; if the sequence determination unit output is 0, this indicates that the garage is returned and all values in the first column are reset to 1. When all of the columns 2 are 0, the encoding is completed,
Xtis generated by S and DtThe components are spliced to form the composite material,
at＝softmax(at)∈(0-1)T+1
probt＝softmax(probt)∈(0-1)T+1
Obtaining the distribution probability corresponding to each task edge to be served
On the basis of the above-described embodiment, the direction determining unit outputs the direction sequence by performing:
determining a probability value of each set trend of each target task edge in the input task sequence; setting the direction corresponding to the maximum probability value as a direction result of the target task edge; and arranging corresponding direction results according to the sequence of the direction sequence, forming the direction sequence of the task sequence and outputting the direction sequence.
The set trend may be a direction from the first end point to the second end point and a direction from the second end point to the first end point while the task to be served is running, the probability value may be a possible probability of each set trend while each task to be served is running, and the probability value may be calculated by the direction determining unit based on the task sequence, the task amount of each task to be served and the load threshold.
Specifically, the task vector corresponding to the target task edge in the task sequence may be based on the following formula
The probability values of the corresponding target task vectors in the set trends are obtained through calculation, the set trend corresponding to the maximum probability value can be used as the direction result of the target task edge and can be represented by 1 or 0, for example, the set trend from the first endpoint to the second endpoint of the task edge to be served can be represented by 0, the set trend from the second endpoint to the first endpoint of the task edge to be served can be represented by 1, and the direction results can be sorted according to the virtual sequence of the tasks to form a direction sequence and output.
On the basis of the above embodiment, merging the task sequence and the direction sequence to generate a planned path includes: sequentially extracting target task edges and corresponding direction results from the task sequence and the direction sequence respectively; marking the direction of the corresponding target service edge according to the result of each direction; and connecting the target service sides with the direction marks in sequence to generate a planning path.
The direction mark can be marked in the map information corresponding to the target service side, and the planned path can be a route of the unmanned express delivery vehicle for providing service for the task to be served.
Specifically, the target task sides and the direction results can be sequentially obtained from the sequence according to the sequence of the task sequence and the direction sequence, the map information corresponding to the target service sides can be marked according to the direction results, the trend of the service provided by the unmanned express delivery vehicle is described, then the marked target task sides can be connected according to the sequence, a planned path can be generated based on the connected target task sides, and it can be understood that the mode of connecting each target task side can include a dijkstra algorithm, an interpolation method and the like.
For example, fig. 4 is an example diagram of a path planning method provided by the second embodiment of the present invention, referring to fig. 4, step 1 in the second embodiment of the present invention may convert task edges to be served into vertices to generate a graph to be served, step 2 may generate a structure matrix C and a feature matrix F based on the graph to be served, and step 3 may generate a task matrix from the structure matrix C and the feature matrix F through a three-layer graph neural network 21
Step 4 can be used for making task matrix
The sequence determination unit 22 based on the recurrent neural network generates the task sequence pi, wherein the recurrent neural network shown in fig. 4 is a time sequence structure, only one output result is output at each moment, elements in the task sequence pi are not output simultaneously, step 5 can input the task sequence pi into the direction determination unit 23 of the bidirectional recurrent neural network to obtain the direction sequence di, and step 6 can generate the task sequence pi and the direction sequence diAnd forming a corresponding path planning solution.
It is understood that, in the above example, 3, 4 and 5 of the steps can be implemented in a manner that each step is executed by a separate neural network model, or executed by a neural network model, and the model can be provided with the function of implementing steps 3, 4 and 5.
EXAMPLE III
Fig. 5 is a schematic structural diagram of a path planning apparatus provided in a third embodiment of the present invention, and referring to fig. 5, the path planning apparatus provided in the third embodiment of the present invention can execute the path planning method provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method. The device can be implemented by software and/or hardware, and specifically comprises: a task edge determination module 301, a matrix determination module 302, and a path generation module 303.
The task edge determining module 301 is configured to determine a task edge to be served according to the service request.
A matrix determining module 302, configured to determine a task matrix based on the task edge to be served.
And a path generating module 303, configured to determine a planned path according to the task matrix and a preset path planning model.
According to the technical scheme of the embodiment of the invention, the task edge determining module determines the task edge to be served according to the service request, the matrix determining module determines the task matrix based on the task edge to be served, and the path generating module determines the planned path based on the task matrix and the preset path planning model, so that the task edge becomes the minimum unit for carrying out service, the unmanned express delivery vehicle can carry out service in a task edge mode, the path planning efficiency is improved, the calculation cost is reduced, and the conveying efficiency of the unmanned express delivery vehicle can be improved.
On the basis of the above embodiment of the present invention, the task edge determining module includes:
and the extracting unit is used for extracting the request position in the service request.
And the generating unit is used for mapping the service request to a corresponding street to be served according to the request position and taking the street to be served as a task side to be served.
On the basis of the above embodiment of the present invention, the matrix determination module includes:
and the structure matrix generating unit is used for generating a structure matrix according to the distance information of the task edge to be served.
And the characteristic matrix generating unit is used for generating a characteristic matrix according to the attribute information of the task edge to be served.
And the task matrix generating unit is used for inputting the structure matrix and the characteristic matrix into a preset atlas neural network to generate a task matrix.
On the basis of the above embodiment of the present invention, the path generating module includes:
and the task generating unit is used for inputting each task vector in the task matrix into the sequence determining unit of the path planning model, sequentially arranging the output results of the sequence determining unit and forming a task sequence, wherein each task vector corresponds to one task edge to be served.
And the direction generating unit is used for inputting the task sequence into the direction determining unit of the path planning model and acquiring the direction sequence output by the direction determining unit.
And the planning path generating unit is used for combining the task sequence and the direction sequence to generate a planning path.
On the basis of the above embodiment of the present invention, the sequence determining unit is specifically configured to:
determining a probability distribution value of each task vector in the task matrix based on a set probability calculation formula; determining a target task edge corresponding to the maximum probability distribution value; if the task quantity of the target task edge is smaller than a load threshold, taking the target task edge as an output result, obtaining a task matrix after shielding the task vector corresponding to the target task edge, and returning to execute the calculation operation of probability distribution until all task vectors are shielded; otherwise, the task matrix after shielding the task vector corresponding to the target task edge is obtained and returns to execute the calculation operation of probability distribution until all the task vectors are shielded.
On the basis of the above embodiment of the present invention, the direction determining unit is specifically configured to:
determining a probability value of each set trend of each target task edge in the input task sequence; setting the direction corresponding to the maximum probability value as a direction result of the target task edge; and arranging corresponding direction results according to the sequence of the direction sequence, forming the direction sequence of the task sequence and outputting the direction sequence.
On the basis of the above embodiment, the planned path generating unit includes:
and the parameter acquisition subunit is used for respectively extracting the target task edge and the corresponding direction result from the task sequence and the direction sequence in sequence.
And the marking subunit is used for marking the direction of the corresponding target service edge according to the result of each direction.
And the connecting subunit is used for sequentially connecting the target service sides marked by the direction marks to generate a planning path.
On the basis of the above embodiment, the task edge determining module further includes:
a second extraction unit, configured to extract a request location of the service request.
And the second generating unit is used for connecting the service requests within the threshold distance according to the request position so as to generate a task edge to be served.
Example four
Fig. 6 is a schematic structural diagram of an apparatus according to a fourth embodiment of the present invention, as shown in fig. 6, the apparatus includes a processor 40, a memory 41, an input device 42, and an output device 43; the number of processors 40 in the device may be one or more, and one processor 40 is taken as an example in fig. 6; the processor 40, the memory 41, the input device 42 and the output device 43 in the apparatus may be connected by a bus or other means, as exemplified by the bus connection in fig. 6.
The memory 41 is a computer-readable storage medium, and can be used for storing software programs, computer-executable programs, and modules, such as program modules corresponding to the path planning method in the embodiment of the present invention (for example, the task edge determining module 301, the matrix determining module 302, and the path generating module 303 in the path planning apparatus). The processor 40 executes various functional applications of the device and data processing by executing software programs, instructions and modules stored in the memory 41, namely, implements the path planning method described above.
The memory 41 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal, and the like. Further, the memory 41 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, memory 41 may further include memory located remotely from processor 40, which may be connected to the device over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 72 may be used to receive entered numeric or character information and to generate key signal inputs relating to user settings and function controls of the apparatus. The output device 73 may include a display device such as a display screen.
EXAMPLE five
An embodiment of the present invention further provides a storage medium containing computer-executable instructions, which when executed by a computer processor, perform a path planning method, including:
determining a task edge to be served according to the service request;
determining a task matrix based on the task edge to be served;
and determining a planning path according to the task matrix and a preset path planning model.
Of course, the storage medium provided by the embodiment of the present invention contains computer-executable instructions, and the computer-executable instructions are not limited to the operations of the method described above, and may also perform related operations in the path planning method provided by any embodiment of the present invention.
From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.
It should be noted that, in the embodiment of the path planning apparatus, each included unit and each included module are only divided according to functional logic, but are not limited to the above division, as long as the corresponding function can be implemented; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (9)
1. A method of path planning, comprising:
determining a task edge to be served according to the service request;
determining a task matrix based on the task edge to be served;
determining a planning path according to the task matrix and a preset path planning model, wherein the method comprises the following steps: inputting each task vector in the task matrix into a sequence determination unit of a path planning model, sequentially arranging output results of the sequence determination unit to form a task sequence, wherein each task vector corresponds to a task edge to be served; inputting the task sequence into a direction determining unit of a path planning model, and acquiring the direction sequence output by the direction determining unit; merging the task sequence and the direction sequence to generate a planning path;
wherein the sequence determination unit determines an output result by: determining a probability distribution value of each task vector in the task matrix based on a set probability calculation formula; determining a target task edge corresponding to the maximum probability distribution value; if the task quantity of the target task edge is smaller than a load threshold, taking the target task edge as an output result, obtaining a task matrix after shielding the task vector corresponding to the target task edge, and returning to execute the calculation operation of probability distribution until all task vectors are shielded; otherwise, the task matrix after shielding the task vector corresponding to the target task edge is obtained and returns to execute the calculation operation of probability distribution until all the task vectors are shielded.
2. The method of claim 1, wherein determining task edges to be serviced according to service requests comprises:
extracting a request position in the service request;
and mapping the service request to a corresponding street to be served according to the request position, and taking the street to be served as a task edge to be served.
3. The method of claim 1, wherein determining a task matrix based on the task edges to be serviced comprises:
generating a structural matrix according to the distance information of the task edge to be served;
generating a characteristic matrix according to the attribute information of the task edge to be served;
and inputting the structural matrix and the characteristic matrix into a preset graph convolution neural network to generate a task matrix.
4. The method of claim 1, wherein the direction determination unit outputs the sequence of directions by performing:
determining a probability value of each set trend of each target task edge in the input task sequence;
setting the direction corresponding to the maximum probability value as a direction result of the target task edge;
and arranging corresponding direction results according to the sequence of the direction sequence, forming the direction sequence of the task sequence and outputting the direction sequence.
5. The method of claim 4, wherein the merging the sequence of tasks and the sequence of directions to generate a planned path comprises:
sequentially extracting target task edges and corresponding direction results from the task sequence and the direction sequence respectively;
marking the direction of the corresponding target service edge according to the result of each direction;
and connecting the target service sides with the direction marks in sequence to generate a planning path.
6. The method of claim 1, wherein determining task edges to be serviced based on service requests further comprises:
extracting a request position of the service request;
and connecting the service requests within the threshold distance according to the request position to generate a task edge to be served.
7. A path planning apparatus, comprising:
the task edge determining module is used for determining the edge of the task to be served according to the service request;
the matrix determining module is used for determining a task matrix based on the task edge to be served;
a path generating module, configured to determine a planned path according to the task matrix and a preset path planning model, where the path generating module is specifically configured to: inputting each task vector in the task matrix into a sequence determination unit of a path planning model, sequentially arranging output results of the sequence determination unit to form a task sequence, wherein each task vector corresponds to a task edge to be served; inputting the task sequence into a direction determining unit of a path planning model, and acquiring the direction sequence output by the direction determining unit; merging the task sequence and the direction sequence to generate a planning path; wherein the sequence determination unit determines an output result by: determining a probability distribution value of each task vector in the task matrix based on a set probability calculation formula; determining a target task edge corresponding to the maximum probability distribution value; if the task quantity of the target task edge is smaller than a load threshold, taking the target task edge as an output result, obtaining a task matrix after shielding the task vector corresponding to the target task edge, and returning to execute the calculation operation of probability distribution until all task vectors are shielded; otherwise, the task matrix after shielding the task vector corresponding to the target task edge is obtained and returns to execute the calculation operation of probability distribution until all the task vectors are shielded.
8. A computer device, characterized in that the computer device comprises:
one or more processors;
a memory for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the path planning method of any of claims 1-6.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out a path planning method according to any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910482413.0A CN110146102B (en) | 2019-06-04 | 2019-06-04 | Path planning method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910482413.0A CN110146102B (en) | 2019-06-04 | 2019-06-04 | Path planning method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110146102A CN110146102A (en) | 2019-08-20 |
| CN110146102B true CN110146102B (en) | 2021-06-15 |
Family
ID=67590536
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910482413.0A Active CN110146102B (en) | 2019-06-04 | 2019-06-04 | Path planning method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110146102B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111325505B (en) * | 2020-02-14 | 2023-07-18 | 北京百度网讯科技有限公司 | Method and device for processing network points and electronic equipment |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104331746B (en) * | 2014-10-20 | 2018-12-14 | 江苏大学 | A kind of dynamic path optimization system and method for separate type |
| CN107101645B (en) * | 2017-05-05 | 2019-10-01 | 中国联合网络通信集团有限公司 | A kind of paths planning method and path planning apparatus |
| CN109145160A (en) * | 2017-06-28 | 2019-01-04 | 华东师范大学 | Key side is chosen in probability graph and optimizes the method and storage medium of key side |
| CN107560631B (en) * | 2017-08-30 | 2020-02-14 | 国网智能科技股份有限公司 | Path planning method and device and inspection robot |
| US10551202B2 (en) * | 2017-09-11 | 2020-02-04 | Fuji Xerox Co., Ltd. | Generating itinerary destinations and recommendations from scanned content |
| CN109816246A (en) * | 2019-01-28 | 2019-05-28 | 江苏大地物流有限责任公司 | Logistic dispatching system based on GIS and intelligent algorithm |
-
2019
- 2019-06-04 CN CN201910482413.0A patent/CN110146102B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110146102A (en) | 2019-08-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3479306B1 (en) | Method and system for estimating time of arrival | |
| CN111476588B (en) | Order demand prediction method and device, electronic equipment and readable storage medium | |
| AU2016102423A4 (en) | System and method for determining vehicle load status | |
| CN104683405B (en) | The method and apparatus of cluster server distribution map matching task in car networking | |
| WO2020176381A1 (en) | Joint order dispatching and fleet management for online ride-sharing platforms | |
| JP2019507316A (en) | System and method for advising estimated arrival times | |
| CN104599002B (en) | Method and equipment for predicting order value | |
| CN111784018A (en) | Resource scheduling method and device, electronic equipment and storage medium | |
| CN113763700A (en) | Information processing method, information processing device, computer equipment and storage medium | |
| CN109716383A (en) | The system and method for carrying out price estimation using machine learning techniques | |
| CN108337685B (en) | Wireless sensor network data fusion method based on sub-clustering DGM | |
| CN111126799A (en) | Shared network driver and crew matching method based on bipartite graph | |
| CN110146102B (en) | Path planning method, device, equipment and storage medium | |
| Tang et al. | The data and science behind grabshare carpooling | |
| WO2023000261A1 (en) | Regional traffic prediction method and device | |
| CN111105076A (en) | Weather forecast data generation method and device, computer equipment and storage medium | |
| CN112287503B (en) | Dynamic space network construction method for traffic demand prediction | |
| CN113932796A (en) | High-precision map lane line generation method and device and electronic equipment | |
| CN111489214A (en) | Order allocation method, condition setting method and device and electronic equipment | |
| CN111047107B (en) | Road traffic time prediction method, device, electronic equipment and storage medium | |
| CN116151600B (en) | Maintenance method, device, computer equipment and storage medium for shared vehicle | |
| CN111861178A (en) | Service matching model training method, service matching method, device and medium | |
| CN115456238A (en) | Urban trip demand prediction method based on dynamic multi-view coupling graph convolution | |
| CN114548572A (en) | Method, device, equipment and medium for predicting urban road network traffic state | |
| CN114872718A (en) | Vehicle trajectory prediction method, vehicle trajectory prediction device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CB03 | Change of inventor or designer information |
Inventor after: Yang Peng Inventor after: Li Guiying Inventor after: Tang Ke Inventor before: Li Guiying Inventor before: Yang Peng Inventor before: Tang Ke |
|
| CB03 | Change of inventor or designer information |