CN112183831A - Route generation method and device, storage medium and electronic equipment - Google Patents

Route generation method and device, storage medium and electronic equipment Download PDF

Info

Publication number
CN112183831A
CN112183831A CN202010981300.8A CN202010981300A CN112183831A CN 112183831 A CN112183831 A CN 112183831A CN 202010981300 A CN202010981300 A CN 202010981300A CN 112183831 A CN112183831 A CN 112183831A
Authority
CN
China
Prior art keywords
grid
target
route
track
grids
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.)
Pending
Application number
CN202010981300.8A
Other languages
Chinese (zh)
Inventor
曾俊齐
陈水平
叶靖
姜海
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Beijing Sankuai Online Technology Co Ltd
Original Assignee
Tsinghua University
Beijing Sankuai Online Technology Co Ltd
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 Tsinghua University, Beijing Sankuai Online Technology Co Ltd filed Critical Tsinghua University
Priority to CN202010981300.8A priority Critical patent/CN112183831A/en
Publication of CN112183831A publication Critical patent/CN112183831A/en
Pending legal-status Critical Current

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
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Human Resources & Organizations (AREA)
  • Evolutionary Computation (AREA)
  • Economics (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Computing Systems (AREA)
  • Strategic Management (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Databases & Information Systems (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Evolutionary Biology (AREA)
  • Development Economics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Game Theory and Decision Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Remote Sensing (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Navigation (AREA)

Abstract

The present disclosure relates to a route generation method and apparatus, a storage medium, and an electronic device, the method including: acquiring a starting position and an end position of a route to be generated; determining a starting point grid corresponding to the starting point position and an end point grid corresponding to the end point position in a preset grid obtained by coding a road network; inputting the starting point grid and the end point grid into a pre-trained probability model, and determining a target grid route through an output result of the probability model and a preset path search algorithm, wherein the probability model is used for generating a jump probability of any grid jumping to each grid adjacent to the grid based on the end point grid; and fitting the target grid route with the road network, and taking the fitting result as the target route. The method and the device can improve the accuracy of the generated route and reduce the cost of the generated route.

Description

Route generation method and device, storage medium and electronic equipment
Technical Field
The present disclosure relates to the field of path planning, and in particular, to a method and an apparatus for generating a route, a storage medium, and an electronic device.
Background
The path planning is a technology commonly used by people during life and travel, real road network information is mostly used in the current path planning algorithm, however, the current roads are large in quantity and complex in road conditions, the data volume of path planning modeling based on the road conditions is large, the modeling difficulty and the training difficulty are high, and the calculation cost required by the path planning is high; the route generated by the route planning of the road-free network does not consider the real road condition, is often the optimal route under the ideal condition, and has low feasibility in the actual road network.
Disclosure of Invention
An object of the present disclosure is to provide a route generation method and apparatus, a storage medium, and an electronic device, which have solved the above technical problems.
In order to achieve the above object, a first aspect of the present disclosure provides a route generation method, including: acquiring a starting position and an end position of a route to be generated; determining a starting point grid corresponding to the starting point position and an end point grid corresponding to the end point position in a preset grid obtained by coding a road network; inputting the starting point grid and the end point grid into a pre-trained probability model, and determining a target grid route through an output result of the probability model and a preset path search algorithm, wherein the probability model is used for generating a jump probability of any grid jumping to each grid adjacent to the grid based on the end point grid; and fitting the target grid route with the road network, and taking the fitting result as the target route.
Optionally, the determining a target grid route through an output result of the probability model and a preset path search algorithm includes: taking the starting point grid as a target grid, repeatedly executing the step of determining the jump probability of the target grid to the adjacent grid based on the probability model, and taking the grid with the maximum jump probability as a new target grid until the grid with the maximum jump probability is the end point grid; and connecting the target grids in sequence to obtain a target grid route.
Optionally, the probability model includes an embedding layer, a connection layer, and an output layer, where the embedding layer is configured to convert the obtained target mesh into a target vector and convert the obtained end point mesh into an end point vector; the connecting layer is used for connecting the target vector and the destination vector into a connecting vector; the output layer is used for generating the jump probability of the target grid jumping to the adjacent grid based on the connection vector.
Optionally, the probability model further includes a long-short term memory LSTM layer, where the LSTM layer is configured to generate a trajectory vector based on a history sequence corresponding to a current target grid, where the history sequence is obtained by arranging all the determined target grids according to an obtaining order; the output layer is further configured to generate a jump probability distribution based on the trajectory vector and a jump vector corresponding to the target grid, where the jump vector is a parameter adjusted in a training phase of the probability model.
Optionally, the LSTM layer is further configured to: storing a history sequence corresponding to the current target grid; generating a trajectory vector based on a history sequence corresponding to the current target grid, including: and acquiring a history sequence corresponding to the last target grid, adding the current target grid in the history sequence to update the history sequence, and generating a track vector based on the updated history sequence.
Optionally, the probability model is obtained by training through the following steps: acquiring a plurality of historical tracks uploaded by a user, wherein each historical track is a set formed by a plurality of track points; for each historical track, based on the coordinate position of the track point of the historical track, projecting the track point into the preset grid to obtain a plurality of track point grids, and connecting the track point grids according to the arrangement sequence of the track points to obtain a track grid sequence; inputting a plurality of track grid sequences corresponding to a plurality of historical tracks into a probability model to be trained, and adjusting parameters in the probability model based on a preset loss function.
Optionally, the connecting the trace point grids according to the arrangement order of the trace points to obtain a trace grid sequence includes: under the condition that the track point grids corresponding to two track points at adjacent arrangement positions are not adjacent, the centers of the track point grids which are not adjacent are connected, the grids which are connected by wires are used as the track point grids, and the grids are sequentially inserted between the sequence positions of the track point grids which are not adjacent, so that a track grid sequence is obtained.
Optionally, the preset mesh is a regular hexagon mesh, and the obtaining of a plurality of meshes of the locus points by projecting the locus points into the preset mesh includes: judging whether the distance between two track points at adjacent arrangement positions is larger than the side length of a hexagon of the regular hexagon grid or not; under the condition that the distance is smaller than the side length of a hexagon of the regular hexagon grid, acquiring the track point grids corresponding to the two track points; and under the condition that the distance is less than the side length of the hexagon of the regular hexagon grid, connecting the two track points, and taking the grid through which the connecting line passes as a track point grid.
Optionally, the fitting the target grid route to the road network and taking a fitting result as a target route includes: determining a plurality of target road segments in a road network surrounded by grids in the target grid route; and connecting the target road sections between the starting position and the end position to obtain at least one target route.
Optionally, the determining a plurality of target road segments in the road network surrounded by the mesh in the target mesh route includes: determining a target sub-road network surrounded by grids in the target grid route in the road network, wherein the target sub-road network comprises a plurality of target road sections and road condition information corresponding to each road section; connecting the target road sections between the starting position and the ending position to obtain at least one target route, wherein the method comprises the following steps: connecting the target road sections between the starting position and the end position according to route generation conditions and the road condition information, and generating a target route meeting the route generation conditions in the target road sub-network; wherein the route generation condition includes at least one of a distance condition, a turn number condition, a traffic light number condition, and a link type condition.
In a second aspect of the present disclosure, there is provided a route generation device including: the acquisition module is used for acquiring a starting point position and an end point position of a route to be generated; the positioning module is used for determining a starting point grid corresponding to the starting point position and an end point grid corresponding to the end point position in a preset grid obtained by coding a road network; the determining module is used for inputting the starting point grid and the end point grid into a pre-trained probability model, and determining a target grid route according to an output result of the probability model and a preset path searching algorithm, wherein the probability model is used for generating the jump probability of any grid jumping to each grid adjacent to the grid based on the end point grid; and the fitting module is used for fitting the target grid route with the road network and taking the fitting result as the target route.
Optionally, the determining module is configured to use the starting point grid as a target grid, and repeatedly perform the step of determining a jump probability of the target grid jumping to an adjacent grid based on the probability model, and use the grid with the largest jump probability as a new target grid until the grid with the largest jump probability is the ending point grid; and connecting the target grids in sequence to obtain a target grid route.
Optionally, the probability model includes an embedding layer, a connection layer, and an output layer, where the embedding layer is configured to convert the obtained target mesh into a target vector and convert the obtained end point mesh into an end point vector; the connecting layer is used for connecting the target vector and the destination vector into a connecting vector; the output layer is used for generating the jump probability of the target grid jumping to the adjacent grid based on the connection vector.
Optionally, the probability model further includes a long-short term memory LSTM layer, where the LSTM layer is configured to generate a trajectory vector based on a history sequence corresponding to a current target grid, where the history sequence is obtained by arranging all the determined target grids according to an obtaining order; the output layer is further configured to generate a jump probability distribution based on the trajectory vector and a jump vector corresponding to the target grid, where the jump vector is a parameter adjusted in a training phase of the probability model.
Optionally, the LSTM layer is further configured to: the method comprises the steps of storing a history sequence corresponding to a current target grid, obtaining a history sequence corresponding to a previous target grid, adding the current target grid in the history sequence to update the history sequence, and generating a track vector based on the updated history sequence.
Optionally, the device further includes a training module, configured to obtain multiple historical tracks uploaded by a user, where each historical track is a set formed by multiple track points; for each historical track, based on the coordinate position of the track point of the historical track, projecting the track point into the preset grid to obtain a plurality of track point grids, and connecting the track point grids according to the arrangement sequence of the track points to obtain a track grid sequence; inputting a plurality of track grid sequences corresponding to a plurality of historical tracks into a probability model to be trained, and adjusting parameters in the probability model based on a preset loss function.
Optionally, the training module is configured to connect centers of non-adjacent trace point grids under the condition that trace point grids corresponding to two trace points in adjacent arrangement positions are not adjacent, and sequentially insert grids passed by a connection line as the trace point grids between sequence positions of the non-adjacent trace point grids to obtain a trace grid sequence.
Optionally, the preset grid is a regular hexagonal grid, and the training module is configured to determine whether a distance between two trajectory points located at adjacent arrangement positions is greater than a side length of a hexagon of the regular hexagonal grid; under the condition that the distance is smaller than the side length of a hexagon of the regular hexagon grid, acquiring the track point grids corresponding to the two track points; and under the condition that the distance is less than the side length of the hexagon of the regular hexagon grid, connecting the two track points, and taking the grid through which the connecting line passes as a track point grid.
Optionally, the fitting module is configured to determine a plurality of target road segments in a road network surrounded by grids in the target grid route; and connecting the target road sections between the starting position and the end position to obtain at least one target route.
Optionally, the fitting module is configured to determine a target sub-road network surrounded by a mesh in the target mesh route in a road network, where the target sub-road network includes a plurality of target road segments and road condition information corresponding to each road segment; connecting the target road sections between the starting position and the end position according to route generation conditions and the road condition information, and generating a target route meeting the route generation conditions in the target road sub-network; wherein the route generation condition includes at least one of a distance condition, a turn number condition, a traffic light number condition, and a link type condition.
In a third aspect of the disclosure, a computer-readable storage medium is provided, on which a computer program is stored, which when executed by a processor implements the steps of the method of any one of the first aspect of the disclosure
In a fourth aspect of the present disclosure, an electronic device is provided, comprising a memory and a processor; the memory has a computer program stored thereon; a processor is adapted to execute the computer program in the memory to implement the steps of the method of any one of the first aspect of the present disclosure.
Through the technical scheme, the following technical effects can be at least achieved:
the target grid route can be determined based on a probability model and a preset path search algorithm in a preset grid obtained based on road network coding, and the target grid route is fitted with the road network to obtain the target route.
Additional features and advantages of the disclosure will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure without limiting the disclosure. In the drawings:
FIG. 1 is a flow chart illustrating a method of route generation according to an exemplary disclosed embodiment.
FIG. 2 is a schematic diagram illustrating a route fitting process according to an exemplary disclosed embodiment.
FIG. 3 is a schematic diagram illustrating a generation of a trajectory grid sequence according to an exemplary disclosed embodiment.
FIG. 4 is a schematic diagram illustrating another generation of a trajectory grid sequence according to an exemplary disclosed embodiment.
FIG. 5 is a schematic diagram illustrating generation of a target grid route by a probabilistic model and a path search algorithm according to an exemplary disclosed embodiment.
FIG. 6 is a block diagram illustrating a route generation apparatus according to an exemplary disclosed embodiment.
FIG. 7 is a block diagram illustrating an electronic device according to an exemplary disclosed embodiment.
Detailed Description
The following detailed description of specific embodiments of the present disclosure is provided in connection with the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present disclosure, are given by way of illustration and explanation only, not limitation.
The grid used in the present disclosure will be explained first. The specific position in the road network is located through coordinates, the coordinates may be obtained through longitude and latitude codes, or may be obtained through horizontal and vertical coordinate codes only having relative positions, and after a specific coordinate is given, the specific position corresponding to the coordinate and the road condition environment around the position can be searched in the road network. Therefore, under the condition that only the position information in the road network is considered and other road condition information is not considered, the grid covering the road network can be generated by coding the road network, and each coordinate position in the road network can be projected into the grid and has unique grid coordinates.
At present, a rectangular grid and a cellular (i.e. regular hexagonal) grid can be generated by GeoHash coding and H3 coding, respectively, and the GeoHash coding process is as follows: firstly, performing binary calculation on a longitude value and a latitude value respectively to obtain two strings of binary codes with equal length; combining the two strings of binary codes to generate a new string, wherein longitude is set at even positions and latitude is set at odd positions; finally, converting every six groups of the new string into decimal numbers, finding out characters corresponding to the decimal numbers through table lookup, and combining the characters into a GeoHash code. According to different precision requirements, the number of times of binary calculation is different, and finally obtained character coding lengths are also different. In H3 encoding, edges of hexagons are all adjacent, and point neighbors in rectangular division cannot occur, so that isotropy can be guaranteed. It should be noted that the present disclosure does not limit the mesh encoding manner and the mesh shape, and those skilled in the art should know that any shape of mesh generated by other encoding forms should also belong to one of the preset meshes that can be selected in the present disclosure.
FIG. 1 is a flow chart illustrating a method of route generation according to an exemplary disclosed embodiment. As shown in fig. 1, the route generation method includes the steps of:
and S11, acquiring the starting position and the end position of the route to be generated.
The starting and ending positions may be determined based on user selection or may be generated based on delivery orders. The starting position and the ending position corresponding to the delivery order can be determined based on the obtained delivery order. For example, when an order delivered from mall a to residential building B is acquired, the location coordinates corresponding to mall a and residential building B may be determined in the road network, and the location coordinate corresponding to mall a may be used as the starting location, and the location coordinate corresponding to residential building B may be used as the ending location.
S12, determining a starting point grid corresponding to the starting point position and an end point grid corresponding to the end point position in a preset grid obtained by coding the road network.
In the present disclosure, the precision of the preset mesh may be fixed, or may be adjusted according to an application scenario. For example, in a take-out delivery scenario, the distance between the end point position and the start point position is usually not far, and the user needs a fine route, so the accuracy of the preset mesh can be set high, that is, the area of the road network actually covered by each mesh is small, and thus a fine and accurate delivery route can be generated; in the scene of city driving navigation, the distance between the end point position and the start point position may be far (cross-district and cross-city conditions may also occur), and the higher the accuracy of the mesh is, the longer the iteration time required for generating the route is, therefore, the accuracy of the preset mesh can be set to be moderate, that is, the road network area actually covered by each mesh is moderate, so that the driving route can be generated quickly.
And S13, inputting the starting point grid and the end point grid into a pre-trained probability model, and determining a target grid route according to an output result of the probability model and a preset path search algorithm.
The probability model is used for generating the jump probability of any grid jumping to each grid adjacent to the grid based on the terminal grid. That is, by inputting an arbitrary mesh and an end point mesh into the probability model, the probability of jumping from the mesh to each mesh adjacent to the mesh can be obtained under the condition that the mesh starts from the mesh to the end point mesh.
It should be noted that the target grid route may be a sequence of target grids with a rank order, or may be a set of target grids without a rank order, which is not limited by the present disclosure.
In the case that the probabilistic model is an LSTM (Long short-term memory) model or includes an LSTM layer, the probabilistic model may extract a hidden state of a history sequence generated in a probability prediction process of a preamble, and the probabilistic model may output a probability that a current target mesh jumps to an adjacent mesh based on the hidden state of the history sequence. The hidden state may be a track vector generated by the LSTM layer based on the history sequence.
For example, after a starting point grid and an end point grid are input into a probability model, a jump probability of the starting point grid output by the probability model jumping to an adjacent grid is obtained, a grid A jumped to by the starting point grid is determined based on a preset path search algorithm, the probability model is further based on a hidden state of a history sequence of the starting point grid-grid A and a jump probability of the end point grid output grid A jumping to the adjacent grid, a grid B jumped to by the grid A is determined based on the preset path search algorithm, the probability model is further based on a hidden state of the history sequence of the starting point grid-grid A-grid B and a jump probability of the end point grid output grid B jumping to the adjacent grid, and the process is repeated. In this way, historical sequences can be saved based on the LSTM layer and more accurate hop probabilities generated based on the historical sequences.
The probability model is obtained by training a plurality of historical tracks and/or a plurality of appointed sample tracks, and the accuracy of the jump probability obtained by the probability model prediction is higher when the number of training samples is larger.
In the disclosure, the preset path search algorithm may be a Dijkstra algorithm, an a-star algorithm, or other path finding algorithms, and the target grid route is obtained by using the jump probability obtained by the probability model as the traffic revenue in the path finding algorithm, or using the values of the complement, reciprocal, and the like of the jump probability that can perform reverse sequencing on the size of the jump probability as the traffic cost, and combining the traffic revenue and/or the traffic cost required by the algorithm through an iteration or traversal mode preset by the algorithm.
In one possible implementation, the target mesh route is obtained by the following algorithm:
taking the starting point grid as a target grid, repeatedly executing the step of determining the jump probability of the target grid to the adjacent grid based on the probability model, and taking the grid with the maximum jump probability as a new target grid until the grid with the maximum jump probability is the end point grid; and connecting the target grids in sequence to determine a target grid route.
For example, if grid 1 is the starting grid, grid 1 is used as the target grid, and the probability model is input, and the jump probabilities of grid 1 jumping to each adjacent grid are respectively 5% (grid 2), 15% (grid 3), 25% (grid 4), and 55% (grid 5), then the new target grid can be determined to be grid 5; inputting the grid 5 as a target grid into a probability model, and determining that the jump probabilities of the grid 5 jumping to the adjacent grids are respectively 0% (grid 1), 50% (grid 6), 25% (grid 7) and 25% (table 8), so that a new target grid can be determined as grid 8; and circularly executing the steps until the obtained new target grid is the terminal grid, and sequentially connecting the target grids to obtain a target grid route.
And S14, fitting the target grid route with the road network, and taking the fitting result as the target route.
In a possible embodiment, a plurality of target road segments in the road network surrounded by the mesh in the target mesh route are determined, and the target road segments between the starting position and the ending position are connected to obtain at least one target route.
The method comprises the steps of determining target road sections surrounded by grids in a target grid route from a road network through a spatial mapping algorithm, namely constructing an R-tree structure for all road sections in the road network, enabling each road section to correspond to a node in the R-tree, obtaining corresponding sub-indexes in the R-trees based on an outsourcing rectangle of the target grid route, and screening the road sections covered by the target grid route by using shape or other screening tools locally in a node area corresponding to the sub-indexes.
And after the target road section is determined, connecting the target road section between the starting position and the ending position to obtain a target route. It is worth noting that in the case of a low accuracy grid, a grid may cover an area including a plurality of parallel road segments, and therefore, a target route obtained by connecting target road segments may have a plurality of parallel road segments. After a plurality of target routes are acquired, one target route meeting the user condition can be screened out based on the condition selected by the user, and the target route is provided for the user. For example, two routes are generated through the above steps, where route 1 passes through 5 intersections, 3 traffic lights and has a length of 1km, and route 2 passes through 4 intersections, 4 traffic lights and has a length of 0.8km, then route 1 may be provided to the user when the condition selected by the user is "the number of traffic lights is the minimum", and route 2 may be provided to the user when the condition selected by the user is "the shortest length".
Fig. 2 is a schematic diagram of a route fitting process, and as shown in fig. 2, a connecting line between two points is a road segment, and target road segments covered by a grid are shown in a bold manner, and are respectively a, B, c, d, e, f, g, h, i, j, k, l, m, and n, and then the target road segments between a starting point a and a terminal point B are connected to obtain a target route: A-B-c-g-j-m-B.
In a possible implementation manner, determining a target sub-road network surrounded by a mesh in the target mesh route in a road network, where the target sub-road network includes a plurality of target road segments and road condition information corresponding to each road segment, connecting the target road segments between the starting position and the ending position according to a route generation condition and the road condition information, and generating a target route meeting the route generation condition in the target sub-road network; wherein the route generation condition includes at least one of a distance condition, a turn number condition, a traffic light number condition, and a link type condition.
That is, when fitting a route, target links between a start position and an end position are not directly connected, but an optimal route that meets route generation conditions is generated in all the target links based on a certain route generation policy. The route generation condition may be preset by the system or generated in response to a condition setting operation by the user.
In a possible implementation, the probabilistic model in S13 includes an embedding layer, a connection layer, and an output layer, where the embedding layer is configured to convert the obtained target mesh into a target vector and convert the obtained end point mesh into an end point vector; the connecting layer is used for connecting the target vector and the destination vector into a connecting vector; the output layer is used for generating the jump probability of the target grid jumping to the adjacent grid based on the connection vector.
In a possible implementation, the probabilistic model further includes a long-short term memory LSTM layer, where the LSTM layer is configured to generate a trajectory vector based on a history sequence corresponding to a current target mesh, where the history sequence is obtained by arranging all the determined target meshes according to an obtaining order; the output layer is further configured to generate a jump probability distribution based on the trajectory vector and a jump vector corresponding to the target grid, where the jump vector is a parameter adjusted in a training phase of the probability model.
Here, generating the jump probability distribution based on the track vector and the jump vector corresponding to the target mesh may be implemented by performing point multiplication on the track vector and the jump vector to obtain scores of each adjacent mesh, and performing normalization processing on the scores of each mesh.
In a possible implementation, the LSTM layer is further configured to store a history sequence corresponding to the current target mesh; when generating the track vector corresponding to the current target grid, acquiring a history sequence corresponding to the previous target grid, adding the current target grid in the history sequence to update the history sequence, and generating the track vector based on the updated history sequence.
Therefore, repeated calculation of the LSTM layer can be avoided, when the target grid is determined each time, the historical sequence related parameters stored by the previous LSTM layer are directly called, the historical sequence related parameters generated by the current LSTM are stored, and the working efficiency of the LSTM layer is improved.
In one possible embodiment, the probabilistic model in S13 is trained by:
acquiring a plurality of historical tracks uploaded by a user, wherein each historical track is a set formed by a plurality of track points; for each historical track, based on the coordinate position of the track point of the historical track, projecting the track point into the preset grid to obtain a plurality of track point grids, and connecting the track point grids according to the arrangement sequence of the track points to obtain a track grid sequence; inputting a plurality of track grid sequences corresponding to a plurality of historical tracks into a probability model to be trained, and adjusting parameters in the probability model based on a preset loss function.
The user uploading the historical track can be a delivery person who takes out or delivers the goods, or a common owner or a common navigation user, and the user can upload the historical track through a special position uploading device, or can upload the historical track through any device which can acquire a geographical position coordinate, such as a vehicle-mounted terminal, a mobile terminal and a wearable device. The historical track can be uploaded by a user in real time, or can be uploaded to a terminal through a physical copy mode and the like after being stored locally, and is uploaded to a server through the terminal, or can be uploaded through a specific communication mode after being stored offline.
In training, the type of the historical track of the training set may be determined based on an application scenario, for example, when the probabilistic model is a scenario for generating a route of a take-away deliverer, the probabilistic model may be trained with the historical track uploaded by the take-away deliverer, and when the probabilistic model is a scenario for private car route planning, the probabilistic model may be trained with the historical track uploaded by a private car owner; further, different training sets may be determined for different vehicle models to train different probabilistic models, for example, historical tracks uploaded by vehicle owners of all vehicle models of model 1 may be determined to train a probabilistic model, which is used to provide path planning for users of model 1.
In one possible embodiment, the predetermined loss function is set as follows:
given a set of historical tracks
Figure BDA0002687611690000104
And a starting point lsAnd end point ldThen a probability of occurrence of a route p can be defined as
Figure BDA0002687611690000105
Target route p for route planning*The route with the highest probability of occurrence, the target route p*The function of (d) is:
Figure BDA0002687611690000101
through conditional probability decomposition and logarithmic transformation, the function of the target route can be converted into:
Figure BDA0002687611690000102
wherein liIs the ith position in the target route, li-1For the i-1 st position (i.e./in) in the target routeiPrevious position of), Pr (l)i|ls→li-1,ld) Has an end point ofdIn the case of (1), the route is fromsJump toi-1The probability of occurrence of (c).
Therefore, to maximize the probability of all historical tracks occurring, let npFor the length of the trajectory p, the loss function can then be set to:
Figure BDA0002687611690000103
based on the loss function and the historical trajectory, the probability model described above may be trained.
However, it should be understood by those skilled in the art that the above-mentioned loss function is set to maximize the probability of all the historical tracks, and those skilled in the art may simply modify the above-mentioned loss function or design the loss function separately, for example, convert the logarithmized data items into other data processing modes to achieve the similar data item processing effect as the logarithmization, so as to obtain the loss function that can be used for training the probability model, considering the difference of training purposes in different application scenarios. That is, the setting of the loss function may be implemented in many different ways, and the present disclosure does not limit the setting of the loss function.
In order to train a model by using a historical track, the historical track needs to be projected into a preset grid, but because the historical track is generated by reporting a coordinate position in real time by a user, under the condition that a signal is bad or the acquisition frequency is low, track points of the historical track are projected into the preset grid, the situation that grids where adjacent track points are located are discontinuous possibly exists, a complete grid sequence is difficult to generate, and the training of a probability model is not facilitated.
Therefore, in a possible implementation manner, under the condition that the trace point grids corresponding to two trace points at adjacent arrangement positions are not adjacent, the centers of the non-adjacent trace point grids are connected, the grids through which the connecting lines pass are used as the trace point grids, and the grid points are sequentially inserted between the sequence positions of the non-adjacent trace point grids to obtain a trace grid sequence.
It is worth explaining that the problem of discontinuity of grids where the track points are located can be solved by connecting the track points and determining the grids where the track point connecting lines pass through as the track point grids.
As shown in fig. 3, if the grids in which the trace points 1 (indicated by the circles with the reference numbers 1 in fig. 3) and the trace points 2 (indicated by the circles with the reference numbers 2 in fig. 3) are located are not adjacent to each other, the center of the grid in which the trace points 1 are located and the center of the grid in which the trace points 2 are located may be connected, and the grid through which the connection line passes is used as the trace point grid. In fig. 3, the grid of trace points has been shown bold.
Or, in a possible implementation, the preset mesh is a regular hexagon mesh, and it may be determined whether a distance between two trace points located at adjacent arrangement positions is greater than a hexagon side length of the regular hexagon mesh, and under a condition that the distance is smaller than the hexagon side length of the regular hexagon mesh, the trace point mesh corresponding to the two trace points is obtained, and under a condition that the distance is smaller than the hexagon side length of the regular hexagon mesh, the two trace points are connected, and a mesh through which a connection line passes is used as the trace point mesh.
As shown in fig. 4, the trace points 3 (indicated by a circle with the reference number 3 in fig. 4) and the trace points 4 (indicated by a circle with the reference number 4 in fig. 4) are respectively located in two grids, and the connecting line between the trace points 3 and the trace points 4 is greater than the side length of the hexagonal grid, so that the trace points 3 and the trace points 4 can be connected, and the grid through which the connecting line passes is used as a trace point grid. In fig. 4, the grid of trace points has been shown bold.
FIG. 5 is a schematic flow diagram illustrating the generation of a target grid route by a probabilistic model and a path search algorithm according to an exemplary disclosed embodiment. The probability model comprises an LSTM layer, can acquire and store a history sequence and generate a hidden state of state information representing the history sequence, and can generate a jump probability based on the information of the history sequence output by the LSTM layer. As shown in fig. 5, the process includes the following steps:
and S51, inputting the starting point mesh and the end point mesh of the route to be generated into the probability model.
And S52, generating a history sequence based on the starting point grids by the probability model, and adding the starting point grids serving as target grids into the target grid set.
S53, storing the current historical sequence, determining the jump probability of the target grid to the adjacent grid based on the current historical sequence, and determining the grid with the maximum jump probability.
And S54, judging whether the target grid is the end point grid or not. If yes, go to step S56; if not, step S55 is executed.
And S55, adding the grid with the maximum jumping probability as a target grid into the target grid set, and updating the stored history sequence. And returns to step S53.
And S56. And outputting the target grid set as a target grid route.
According to the technical scheme, the target grid route can be determined in the preset grid obtained based on the road network coding based on the probability model and the preset path search algorithm, and the target grid route is fitted with the road network to obtain the target route.
Through the technical scheme, the following technical effects can be at least achieved:
the target grid route can be determined based on a probability model and a preset path search algorithm in a preset grid obtained based on road network coding, and the target grid route is fitted with the road network to obtain the target route.
FIG. 6 is a block diagram illustrating a route generation apparatus according to an exemplary disclosed embodiment. As shown in fig. 6, the apparatus 600 includes:
the obtaining module 610 is configured to obtain a starting point position and an ending point position of a route to be generated.
A positioning module 620, configured to determine a starting point grid corresponding to the starting point position and an ending point grid corresponding to the ending point position in a preset grid obtained by encoding a road network.
A determining module 630, configured to input the starting point mesh and the ending point mesh into a pre-trained probability model, and determine a target mesh route according to an output result of the probability model and a preset path search algorithm, where the probability model is used to generate a jump probability that an arbitrary mesh jumps to each mesh adjacent to the mesh based on the ending point mesh.
And a fitting module 640, configured to fit the target grid route to the road network, and use a fitting result as the target route.
Optionally, the determining module 630 is configured to use the starting point grid as a target grid, and repeatedly perform the step of determining the jump probability of the target grid jumping to an adjacent grid based on the probability model, and using the grid with the largest jump probability as a new target grid until the grid with the largest jump probability is the ending point grid; and connecting the target grids in sequence to obtain a target grid route.
Optionally, the probability model includes an embedding layer, a connection layer, and an output layer, where the embedding layer is configured to convert the obtained target mesh into a target vector and convert the obtained end point mesh into an end point vector; the connecting layer is used for connecting the target vector and the destination vector into a connecting vector; the output layer is used for generating the jump probability of the target grid jumping to the adjacent grid based on the connection vector.
Optionally, the probability model further includes a long-short term memory LSTM layer, where the LSTM layer is configured to generate a trajectory vector based on a history sequence corresponding to a current target grid, where the history sequence is obtained by arranging all the determined target grids according to an obtaining order; the output layer is further configured to generate a jump probability distribution based on the trajectory vector and a jump vector corresponding to the target grid, where the jump vector is a parameter adjusted in a training phase of the probability model.
Optionally, the LSTM layer is further configured to: the method comprises the steps of storing a history sequence corresponding to a current target grid, obtaining a history sequence corresponding to a previous target grid, adding the current target grid in the history sequence to update the history sequence, and generating a track vector based on the updated history sequence.
Optionally, the apparatus 600 further includes a training module, configured to obtain multiple historical tracks uploaded by a user, where each historical track is a set formed by multiple track points; for each historical track, based on the coordinate position of the track point of the historical track, projecting the track point into the preset grid to obtain a plurality of track point grids, and connecting the track point grids according to the arrangement sequence of the track points to obtain a track grid sequence; inputting a plurality of track grid sequences corresponding to a plurality of historical tracks into a probability model to be trained, and adjusting parameters in the probability model based on a preset loss function.
Optionally, the training module is configured to connect centers of non-adjacent trace point grids under the condition that trace point grids corresponding to two trace points in adjacent arrangement positions are not adjacent, and sequentially insert grids passed by a connection line as the trace point grids between sequence positions of the non-adjacent trace point grids to obtain a trace grid sequence.
Optionally, the preset grid is a regular hexagonal grid, and the training module is configured to determine whether a distance between two trajectory points located at adjacent arrangement positions is greater than a side length of a hexagon of the regular hexagonal grid; under the condition that the distance is smaller than the side length of a hexagon of the regular hexagon grid, acquiring the track point grids corresponding to the two track points; and under the condition that the distance is less than the side length of the hexagon of the regular hexagon grid, connecting the two track points, and taking the grid through which the connecting line passes as a track point grid.
Optionally, the fitting module 640 is configured to determine a plurality of target road segments in the road network, which are surrounded by grids in the target grid route; and connecting the target road sections between the starting position and the end position to obtain at least one target route.
Optionally, the fitting module 640 is configured to determine a target sub-road network surrounded by a mesh in the target mesh route in a road network, where the target sub-road network includes a plurality of target road segments and road condition information corresponding to each road segment; connecting the target road sections between the starting position and the end position according to route generation conditions and the road condition information, and generating a target route meeting the route generation conditions in the target road sub-network; wherein the route generation condition includes at least one of a distance condition, a turn number condition, a traffic light number condition, and a link type condition.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Through the technical scheme, the following technical effects can be at least achieved:
the target grid route can be determined based on a probability model and a preset path search algorithm in a preset grid obtained based on road network coding, and the target grid route is fitted with the road network to obtain the target route.
Fig. 7 is a block diagram illustrating an electronic device 700 in accordance with an example embodiment. As shown in fig. 7, the electronic device 700 may include: a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an input/output (I/O) interface 704, and a communication component 705.
The processor 701 is configured to control the overall operation of the electronic device 700, so as to complete all or part of the steps in the route generation method. The memory 702 is used to store various types of data to support operation at the electronic device 700, such as instructions for any application or method operating on the electronic device 700 and application-related data, such as contact data, transmitted and received messages, pictures, audio, video, and the like. The Memory 702 may be implemented by any type of volatile or non-volatile Memory device or combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk, or optical disk. The multimedia components 703 may include screen and audio components. Wherein the screen may be, for example, a touch screen and the audio component is used for outputting and/or inputting audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may further be stored in the memory 702 or transmitted through the communication component 705. The audio assembly also includes at least one speaker for outputting audio signals. The I/O interface 704 provides an interface between the processor 701 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless Communication, such as Wi-Fi, bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IOT, eMTC, or other 5G, etc., or a combination of one or more of them, which is not limited herein. The corresponding communication component 705 may thus include: Wi-Fi module, Bluetooth module, NFC module, etc.
In an exemplary embodiment, the electronic Device 700 may be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for performing the above-described route generation method.
In another exemplary embodiment, a computer readable storage medium is also provided, which comprises program instructions, which when executed by a processor, implement the steps of the route generation method described above. For example, the computer readable storage medium may be the memory 702 described above including program instructions that are executable by the processor 701 of the electronic device 700 to perform the route generation method described above.
The preferred embodiments of the present disclosure are described in detail with reference to the accompanying drawings, however, the present disclosure is not limited to the specific details of the above embodiments, and various simple modifications may be made to the technical solution of the present disclosure within the technical idea of the present disclosure, and these simple modifications all belong to the protection scope of the present disclosure.
It should be noted that, in the foregoing embodiments, various features described in the above embodiments may be combined in any suitable manner, and in order to avoid unnecessary repetition, various combinations that are possible in the present disclosure are not described again.
In addition, any combination of various embodiments of the present disclosure may be made, and the same should be considered as the disclosure of the present disclosure, as long as it does not depart from the spirit of the present disclosure.

Claims (13)

1. A route generation method, characterized in that the method comprises:
acquiring a starting position and an end position of a route to be generated;
determining a starting point grid corresponding to the starting point position and an end point grid corresponding to the end point position in a preset grid obtained by coding a road network;
inputting the starting point grid and the end point grid into a pre-trained probability model, and determining a target grid route through an output result of the probability model and a preset path search algorithm, wherein the probability model is used for generating a jump probability of any grid jumping to each grid adjacent to the grid based on the end point grid;
and fitting the target grid route with the road network, and taking the fitting result as the target route.
2. The method of claim 1, wherein determining a target grid route through the output of the probabilistic model and a preset path search algorithm comprises:
taking the starting point grid as a target grid, repeatedly executing the step of determining the jump probability of the target grid to the adjacent grid based on the probability model, and taking the grid with the maximum jump probability as a new target grid until the grid with the maximum jump probability is the end point grid;
and connecting the target grids in sequence to obtain a target grid route.
3. The method of claim 1, wherein the probabilistic model comprises an embedding layer, a connection layer, an output layer, wherein,
the embedded layer is used for converting the obtained target grid into a target vector and converting the obtained terminal grid into a terminal vector;
the connecting layer is used for connecting the target vector and the destination vector into a connecting vector;
the output layer is used for generating the jump probability of the target grid jumping to the adjacent grid based on the connection vector.
4. The method of claim 3, wherein the probabilistic model further comprises a long-short term memory (LSTM) layer,
the LSTM layer is used for generating a track vector based on a history sequence corresponding to a current target grid, wherein the history sequence is obtained by arranging all the determined target grids according to an obtaining sequence;
the output layer is further configured to generate a jump probability distribution based on the trajectory vector and a jump vector corresponding to the target grid, where the jump vector is a parameter adjusted in a training phase of the probability model.
5. The method of claim 4, wherein the LSTM layer is further configured to:
storing a history sequence corresponding to the current target grid;
generating a trajectory vector based on a history sequence corresponding to the current target grid, including:
and acquiring a history sequence corresponding to the last target grid, adding the current target grid in the history sequence to update the history sequence, and generating a track vector based on the updated history sequence.
6. The method of claim 1, wherein the probabilistic model is trained by:
acquiring a plurality of historical tracks uploaded by a user, wherein each historical track is a set formed by a plurality of track points;
for each historical track, based on the coordinate position of the track point of the historical track, projecting the track point into the preset grid to obtain a plurality of track point grids, and connecting the track point grids according to the arrangement sequence of the track points to obtain a track grid sequence;
inputting a plurality of track grid sequences corresponding to a plurality of historical tracks into a probability model to be trained, and adjusting parameters in the probability model based on a preset loss function.
7. The method of claim 6, wherein connecting the grids of trace points according to the arrangement order of the trace points to obtain a trace grid sequence comprises:
under the condition that the track point grids corresponding to two track points at adjacent arrangement positions are not adjacent, the centers of the track point grids which are not adjacent are connected, the grids which are connected by wires are used as the track point grids, and the grids are sequentially inserted between the sequence positions of the track point grids which are not adjacent, so that a track grid sequence is obtained.
8. The method of claim 6, wherein the predetermined grid is a regular hexagonal grid, and the projecting the trace points into the predetermined grid to obtain a plurality of grids of trace points comprises:
judging whether the distance between two track points at adjacent arrangement positions is larger than the side length of a hexagon of the regular hexagon grid or not;
under the condition that the distance is smaller than the side length of a hexagon of the regular hexagon grid, acquiring the track point grids corresponding to the two track points;
and under the condition that the distance is less than the side length of the hexagon of the regular hexagon grid, connecting the two track points, and taking the grid through which the connecting line passes as a track point grid.
9. The method according to claim 1, wherein fitting the target grid route to a road network and using the fitting result as a target route comprises:
determining a plurality of target road segments in a road network surrounded by grids in the target grid route;
and connecting the target road sections between the starting position and the end position to obtain at least one target route.
10. The method of claim 9, wherein said determining a plurality of target road segments in a road network surrounded by a mesh in said target mesh route comprises:
determining a target sub-road network surrounded by grids in the target grid route in the road network, wherein the target sub-road network comprises a plurality of target road sections and road condition information corresponding to each road section;
connecting the target road sections between the starting position and the ending position to obtain at least one target route, wherein the method comprises the following steps:
connecting the target road sections between the starting position and the end position according to route generation conditions and the road condition information, and generating a target route meeting the route generation conditions in the target road sub-network;
wherein the route generation condition includes at least one of a distance condition, a turn number condition, a traffic light number condition, and a link type condition.
11. A route generation apparatus, characterized in that the apparatus comprises:
the acquisition module is used for acquiring a starting point position and an end point position of a route to be generated;
the positioning module is used for determining a starting point grid corresponding to the starting point position and an end point grid corresponding to the end point position in a preset grid obtained by coding a road network;
the determining module is used for inputting the starting point grid and the end point grid into a pre-trained probability model, and determining a target grid route according to an output result of the probability model and a preset path searching algorithm, wherein the probability model is used for generating the jump probability of any grid jumping to each grid adjacent to the grid based on the end point grid;
and the fitting module is used for fitting the target grid route with the road network and taking the fitting result as the target route.
12. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 10.
13. An electronic device, comprising:
a memory having a computer program stored thereon;
a processor for executing the computer program in the memory to carry out the steps of the method of any one of claims 1 to 10.
CN202010981300.8A 2020-09-17 2020-09-17 Route generation method and device, storage medium and electronic equipment Pending CN112183831A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010981300.8A CN112183831A (en) 2020-09-17 2020-09-17 Route generation method and device, storage medium and electronic equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010981300.8A CN112183831A (en) 2020-09-17 2020-09-17 Route generation method and device, storage medium and electronic equipment

Publications (1)

Publication Number Publication Date
CN112183831A true CN112183831A (en) 2021-01-05

Family

ID=73921639

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010981300.8A Pending CN112183831A (en) 2020-09-17 2020-09-17 Route generation method and device, storage medium and electronic equipment

Country Status (1)

Country Link
CN (1) CN112183831A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115080866A (en) * 2022-08-22 2022-09-20 北京中交兴路信息科技有限公司 Travel path recommendation method and device, storage medium and terminal

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115080866A (en) * 2022-08-22 2022-09-20 北京中交兴路信息科技有限公司 Travel path recommendation method and device, storage medium and terminal

Similar Documents

Publication Publication Date Title
Fajardo et al. A mobile disaster management system using the android technology
US10921136B2 (en) Efficient processing for vector tile generation
CN106767835B (en) Positioning method and device
US10402729B2 (en) Path determination using robust optimization
CN110617831B (en) Method, device and equipment for generating navigation route
CN102762957A (en) Historial traffic data compression
CN105678412A (en) Path planning method and device facing multiple passengers
JP2007241340A (en) N division patrol path search system, route search server, and n division patrol path search method
CN111339230B (en) Vehicle information display method and device, electronic equipment and storage medium
CN111780774B (en) Motion trajectory generation method and device, storage medium and electronic equipment
CN112665601A (en) Path planning method and device, electronic equipment and readable storage medium
KR20220026999A (en) A method, an apparatus, an electronic device and a storage medium for predicting a destination
CN104636457B (en) A kind of method and device of location finding cognition
KR20210117235A (en) Method for managing travel route database based on location and server for the method
KR20200054355A (en) Method, apparatus and computer program for coloring of image, Method, apparatus and computer program for learning of artificial neural network
CN112183831A (en) Route generation method and device, storage medium and electronic equipment
JP2019095569A (en) Map information processing device, map information processing method and map information processing program
CN112417070B (en) Road network topology construction method and device, electronic equipment and storage medium
CN108204817B (en) Navigation route processing method and device
JP2011221888A (en) Area information retrieval server, area information retrieval method and method for creating complement mesh
JP2020067656A (en) Map information processing device, map information processing method, and map information processing program
CN110795519A (en) Markov model and probability statistics-based position prediction method and readable storage medium
KR20080062088A (en) Apparatus and method for showing the way
CN114842207A (en) Road network generation method and device, readable storage medium and electronic equipment
CN114328788A (en) Region coding method, device, equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20210105