CN113587931A - Path planning method, device, equipment and storage medium - Google Patents
Path planning method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN113587931A CN113587931A CN202110776397.3A CN202110776397A CN113587931A CN 113587931 A CN113587931 A CN 113587931A CN 202110776397 A CN202110776397 A CN 202110776397A CN 113587931 A CN113587931 A CN 113587931A
- Authority
- CN
- China
- Prior art keywords
- target
- target point
- initial
- current
- target object
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 77
- 210000002569 neuron Anatomy 0.000 claims abstract description 185
- 230000000694 effects Effects 0.000 claims abstract description 161
- 238000013528 artificial neural network Methods 0.000 claims abstract description 110
- 230000008569 process Effects 0.000 claims description 22
- 238000013507 mapping Methods 0.000 claims description 10
- 238000010276 construction Methods 0.000 claims description 5
- 238000012804 iterative process Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 description 14
- 230000006870 function Effects 0.000 description 10
- 230000004913 activation Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 210000004205 output neuron Anatomy 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
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
- G01C21/20—Instruments for performing navigational calculations
-
- 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
- G01C21/20—Instruments for performing navigational calculations
- G01C21/206—Instruments for performing navigational calculations specially adapted for indoor navigation
-
- 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
- G01C21/3415—Dynamic re-routing, e.g. recalculating the route when the user deviates from calculated route or after detecting real-time traffic data or accidents
-
- 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
- G01C21/3446—Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags, using precalculated routes
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0219—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory ensuring the processing of the whole working surface
-
- 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
-
- 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
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Aviation & Aerospace Engineering (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a path planning method, a device, equipment and a storage medium, wherein the method comprises the following steps: acquiring a target grid map; determining coordinate positions of a plurality of target points and initial coordinate positions of a plurality of target objects in the target grid map; respectively constructing a biological enlightening neural network corresponding to each target point based on preset neuron activity information and a coordinate position of each target point, wherein the preset neuron activity information represents preset activity state information of a neuron corresponding to each target point in the corresponding biological enlightening neural network; and performing task allocation on the plurality of target objects based on the initial coordinate positions of the plurality of target objects and the biological heuristic neural network corresponding to each target object, and determining an initial task path of each target object. The scheme provided by the application can realize dynamic task allocation and path planning of multiple AGVs, and improves the transportation efficiency of warehouse logistics on the basis of reducing the total length of the task path.
Description
Technical Field
The present application relates to a neighborhood of path planning technologies, and in particular, to a method, an apparatus, a device, and a storage medium for path planning.
Background
An agv (automated Guided vehicle), i.e. an automatic navigation vehicle, is an important tool for developing and utilizing ground resources and saving labor cost. At present, the AGV is often used for the storage logistics neighborhood, and the AGV can realize the full automation of article transport and handling process through self-owned auto-control handling mechanism and navigation head. In order to realize efficient operation of warehouse logistics, the application of multiple AGV path planning technology is particularly important. The multi-AGV path planning technology is characterized in that a surrounding environment is firstly sensed, then an optimal or approximately optimal feasible path from a starting point to a final target point is planned in an environment space according to an evaluation performance standard (minimum cost, shortest path, shortest time or minimum energy consumption), and finally a task is efficiently completed at the target point.
The basis of path planning of multiple target points aiming at multiple AGVs is task allocation of the multiple AGVs. Initial task allocation is based on the distance of the AGVs to the target points, and each target point is initially allocated to a certain AGV reasonably and accurately. If the initial task allocation result is simply used as the final task allocation result, the result is often not optimal, because after the initial task allocation, the position of each AGV is constantly changed, some AGVs are allocated with some initial target points, but as each AGV moves, the AGV corresponding to the initial target points may be farther away from the initial target points, and other AGVs may be closer to the initial target points, so that a more scientific and effective path planning method needs to be provided.
Disclosure of Invention
The application provides a path planning method, a path planning device, equipment and a storage medium, which can achieve reasonable distribution of target points to obtain an optimal path through dynamic adjustment of tasks, and improve the transportation efficiency of warehouse logistics, and the technical scheme of the application is as follows:
in one aspect, a method for path planning is provided, where the method includes:
acquiring a target grid map;
determining coordinate positions of a plurality of target points and initial coordinate positions of a plurality of target objects in the target grid map;
respectively constructing a biological enlightening neural network corresponding to each target point based on preset neuron activity information and a coordinate position of each target point, wherein the preset neuron activity information represents preset activity state information of a neuron corresponding to each target point in the corresponding biological enlightening neural network;
performing task allocation on the plurality of target objects based on the initial coordinate positions of the plurality of target objects and the biological heuristic neural network corresponding to each target object, and determining an initial task path of each target object, wherein the initial task path is path information formed by grids of the initial target points corresponding to each target object;
controlling each target object to advance along the initial task path, taking the position information of the grid which passes through at present as the current coordinate position of each target object in the process that each target object advances along the initial task path, and determining the current target-passing point of each target object;
determining a target point to be adjusted in the current non-target points of each target object based on a path length ratio algorithm and path information formed by a grid where the current non-target points are located;
task adjustment is carried out on the multiple target objects based on the current coordinate positions of the multiple target objects and a biological heuristic neural network corresponding to each target point to be adjusted in the target points to be adjusted of the multiple target objects, and a current task path of each target object is determined;
and taking the current task path of each target object as the target task path of each target object under the condition that the number of the current non-passing target points is zero.
In another aspect, a path planning apparatus is provided, the apparatus including:
the target grid map acquisition module is used for acquiring a target grid map;
a coordinate position determination module for determining coordinate positions of a plurality of target points and initial coordinate positions of a plurality of target objects in the target grid map;
the biological elicitation neural network construction module is used for respectively constructing a biological elicitation neural network corresponding to each target point based on preset neuron activity information and coordinate positions of each target point, wherein the preset neuron activity information represents preset activity state information of neurons corresponding to each target point in the corresponding biological elicitation neural network;
the task allocation module is used for allocating tasks to the target objects based on the initial coordinate positions of the target objects and the biological heuristic neural network corresponding to each target point, and determining an initial task path of each target object, wherein the initial task path is path information formed by grids where the initial target points corresponding to each target object are located;
the target object control module is used for controlling each target object to advance along the initial task path, taking the position information of the grid which passes through at present as the current coordinate position of each target object in the process that each target object advances along the initial task path, and determining the current target point which does not pass through of each target object;
a target point to be adjusted determining module, configured to determine, based on a path length ratio algorithm and path information formed by a grid where the current target point that has not passed is located, a target point to be adjusted in the current target point that has not passed of each target object;
the task adjusting module is used for performing task adjustment on the plurality of target objects based on the current coordinate positions of the plurality of target objects and the biological heuristic neural network corresponding to each target point to be adjusted in the target points to be adjusted of the plurality of target objects, and determining the current task path of each target object;
and the target task path module is used for taking the current task path of each target object as the target task path of each target object under the condition that the number of the current non-passing target points is zero.
In another aspect, a path planning apparatus is provided, which includes a processor and a memory, where at least one instruction or at least one program is stored in the memory, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the path planning method as described above.
In another aspect, a computer-readable storage medium is provided, in which at least one instruction or at least one program is stored, and the at least one instruction or the at least one program is loaded and executed by a processor to implement the path planning method as described above.
The path planning method, the device, the equipment and the storage medium have the following technical effects:
according to the technical scheme provided by the application, dynamic task allocation and path planning of multiple AGVs are achieved by adopting a biological heuristic Neural network and integrating the ratios of the lengths of the raster paths between the target points (DRGBNN), for a certain AGV allocated with a plurality of initial target points, in the moving process of the AGV, some target points which possibly need to be dynamically adjusted are found out according to the ratios of the lengths of the raster paths between adjacent target points in the initial target points, and then task allocation is carried out again according to the current coordinate positions of the AGV, so that the total length of the running paths of the AGV is reduced, and meanwhile, the transportation efficiency of storage logistics is improved on the basis of reducing the resource consumption of the AGV.
Drawings
In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a schematic flowchart of a path planning method according to an embodiment of the present application;
FIG. 2 is a schematic diagram of a target grid map provided by embodiments of the present application;
FIG. 3 is a flow chart of a method for constructing a bio-heuristic neural network according to an embodiment of the present disclosure;
FIG. 4 is a schematic diagram of a two-dimensional structure of an initial neural network provided by an embodiment of the present application;
fig. 5 is a schematic flowchart of a task allocation method according to an embodiment of the present application;
fig. 6 is a flowchart illustrating a process of determining a traversal order of the initial target point corresponding to each target object based on the first neuron activity information corresponding to each target object and the initial target point according to an embodiment of the present application;
fig. 7 is a schematic flowchart of a process of determining a target point to be adjusted in the current non-target points of each target object based on a path length ratio algorithm and path information formed by a grid where the current non-target points are located according to an embodiment of the present application;
FIG. 8 is a schematic diagram of a raster forming path in which a currently-unviewed target point is located according to an embodiment of the present disclosure;
fig. 9 is a schematic flowchart of a task adjustment method according to an embodiment of the present application;
fig. 10 is a schematic flowchart of another path planning method provided in the embodiment of the present application;
fig. 11 is a schematic diagram of a path planning apparatus according to an embodiment of the present application;
fig. 12 is a hardware structure block diagram of a server of a path planning method according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort belong to the protection scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or server that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
A path planning method provided in the embodiment of the present application is described below, and fig. 1 is a schematic flow chart of the path planning method provided in the embodiment of the present application. It is noted that the present specification provides the method steps as described in the examples or flowcharts, but may include more or less steps based on routine or non-inventive labor. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. In actual system or product execution, sequential execution or parallel execution (e.g., parallel processor or multi-threaded environment) may be possible according to the embodiments or methods shown in the figures. Specifically, as shown in fig. 1, the method may include:
and S101, acquiring a target grid map.
In the embodiments of the present specification, the target grid map is generally obtained by rasterizing the target planar map. The target plane map may include, but is not limited to, a warehouse plane map, and specifically, the warehouse plane map may include a location of the target cargo, a parking location of a warehouse AGV (Automated Guided Vehicle), and a location of the obstacle.
In the embodiment of the present specification, the length and the width of the warehouse plan map are L and W, respectively, the length and the width of a single grid are m and n, respectively, since the number of grids is an integer, an integer function is required in the rasterization process, and the total number Z of grids of the target grid map obtained by rasterizing the warehouse plan map is as shown in formula (1):
in practical application, the length and width of a single grid need to be preset based on the complexity of the warehouse plan map and the precision requirement of path planning, and in an optional embodiment, for the consideration of convenient calculation, each grid in the target grid map is scaled to be a square with the side length being a unit length.
And S103, determining the coordinate positions of a plurality of target points and the initial coordinate positions of a plurality of target objects in the target grid map.
Specifically, the position coordinates of the target points may be positions of the target goods, the target objects may be the storage AGVs, grids where the target points are located and grids where the target objects are located in the target grid map are determined based on the positions of the target goods in the storage plane map and the parking positions of the storage AGVs, the position information of the grids where the target points are located is used as coordinate positions of the target points, and the position information of the initial grids where the target objects are located is used as initial coordinate positions of the target objects.
In practical application, the coordinate positions of the multiple obstacles in the target grid map can be determined based on the positions of the multiple obstacles in the warehouse plane map.
In one specific embodiment, as shown in FIG. 2, FIG. 2 is a schematic diagram of an object grid map. Marking a plurality of target points, a plurality of target objects and a plurality of obstacles in the target grid map, wherein the marking mark of the target object is A, and obtaining a target object A1~A5(ii) a Marking mark of target pointThe number is T, and the target point T is obtained1~T18(ii) a The obstacles are marked with black boxes.
And S105, respectively constructing a biological enlightening neural network corresponding to each target point based on preset neuron activity information and a coordinate position of each target point, wherein the preset neuron activity information represents preset activity state information of a neuron corresponding to each target point in the corresponding biological enlightening neural network.
In practical applications, a biological enlightened Neural network (GBNN) is an adaptive artificial Neural network and is composed of many interconnected neurons. Each neuron transmits the output neuron activity information to other neurons through the connection weight, and simultaneously receives the activity information transmitted by other neurons.
In an embodiment of the present specification, as shown in fig. 3, fig. 3 is a schematic flow chart of a method for constructing a bio-heuristic neural network provided in an embodiment of the present specification, and specifically, the method may include:
s301, traversing the target points.
And S303, constructing an initial neural network corresponding to the currently traversed target point by taking the neuron corresponding to the currently traversed target point as an initial neuron.
Specifically, taking a neuron corresponding to a currently traversed target point as an initial neuron, and constructing an initial neural network corresponding to the currently traversed target point based on a preset mathematical model, where fig. 4 is a schematic diagram of a two-dimensional structure of the initial neural network, where the preset mathematical model is shown in formula (2):
in the formula (2), C(i,j)Representing the activity information of the neurons at (I, j) coordinate positions, t representing the time, g (x) being the activation function, I(i,j)Bias terms representing the (i, j) coordinate position, and (p, q) another neuron sitsMark position, W((i,j),(p,q))And (b) representing the weight between the neurons of the two coordinate positions (i, j) and (p, q), and m and n represent the scale of the neural network. The numerical value corresponding to the activity information of each neuron at the next moment is obtained by calculating through an activation function according to the numerical value corresponding to the activity information of other neurons in the neural network at the previous moment and the accumulation and bias term of the weight, wherein the preset weight W((i,j),(p,q))As shown in equation (3), the preset activation function g (x) is shown in equation (4):
in formula (3), (i, j) and (p, q) are the coordinate positions of two neurons,representing the euclidean distance of two neurons. Based on the preset weight value, in the preset neural network, when the neurons of the two coordinate positions (i, j) and (p, q) are adjacent neurons, the weight value between the neurons is e-2d(ii) a When the neurons in the two coordinate positions are not adjacent neurons or both represent the same neuron, the weight value between them is 0.
The formula (4) is a preset activation function, and after each neuron receives activity information transmitted from other neurons at the previous time, the received activity information is calculated by the activation function, so that the current activity information of each neuron can be obtained.
S305, performing initialization setting on the corresponding initial neural network according to the preset mapping relation between each grid in the target grid map and each neuron in the corresponding initial neural network.
Specifically, the preset mapping relationship may be a one-to-one correspondence relationship, and the activity information of each neuron may be used as the neuron activity information of the corresponding grid.
In the target grid map, introducing a bias term + E into a neuron corresponding to a grid where a target point is located, introducing a bias term-E into a neuron corresponding to a grid where an obstacle is located, setting bias terms of neurons corresponding to grids where other positions are located to be 0, and obtaining bias terms as shown in formula (5):
after calculation is performed through the activation function, the value corresponding to the activity information of each neuron is between 0 and the value corresponding to the preset activity information of the initial neuron, and the value of the value corresponding to the E is generally larger than the value corresponding to the preset activity information.
In practical applications, the value corresponding to the preset activity information is an upper limit value of a value corresponding to activity information of each neuron in the preset neural network, generally, the value corresponding to the preset activity information may be set to 1, and the value of E may be a number greater than 1, for example, E may be 2 or 3.
And S307, inputting the preset neuron activity information and the coordinate position of the currently traversed target point into the corresponding initial neural network to perform iterative computation on the neuron activity information of each grid until the neuron activity information of each grid output by the corresponding initial neural network remains unchanged, wherein the neuron activity information is used for mapping the shortest grid path length between each grid and the grid where the currently traversed target point is located, and a value corresponding to the neuron activity information is in an inverse relationship with the corresponding shortest grid path length.
The target point traversed at present is the target point T1For example, the corresponding initial neural network is T1The corresponding initial neural network takes the preset activity information of the initial neuron as T at the initial moment1The predetermined neuron activity information is, in practice, to initiate the predetermined activity of the neuronThe value corresponding to the information is initialized to 1, i.e. T1Initializing the value corresponding to the preset neuron activity information to 1, and dividing T in the target grid map1And initializing the numerical value corresponding to the preset neuron activity information of other grids except the grid to be 0. Because the current activity information of each neuron is obtained by calculating the activation function of the activity information received by each neuron and transmitted by other neurons at the last moment, the change of the corresponding numerical value of each neuron activity information in the preset neural network is from T1The grid begins and gradually expands around.
At the first moment, due to the reception of T1Corresponds to an influence of a value of 1, and the neuron activity information of eight neighborhood grids around the grid where T1 is located changes from 0. According to the preset weight W as shown in formula (3)((i,j),(p,q))(p, q) is T1(i, j) is T1Any coordinate position in the coordinate positions of the eight neighborhood grids around the grid, the weight W of the grid where T1 is located and the eight neighborhood grids around the grid((i,j),(p,q))Is not 0 and C(p,q)When the neuron activity information corresponding to the eight surrounding neighborhood grids is calculated according to the formula (2) as 1, W((i,j),(p,q))The bias term of the eight neighboring positions is not 0, so that the values corresponding to the neuron activity information of the eight neighboring positions change from 0 to a value after the activation function g (x) shown in formula (4). From the initial moment to the first moment, except for T1The number corresponding to the neuron activity information is not 0, the number corresponding to the neuron activity information of the surrounding eight neighborhood grids is changed from 0 to a corresponding value according to the formula (2), and in the current preset neural network, the number of grid positions of the neuron activity information corresponding to the number not 0 is changed from one to nine.
At the second moment, due to T1The influence that the corresponding numerical value of the neuron activity information of the peripheral eight neighborhood grids is not 0, and the corresponding numerical value of the neuron activity information of the twenty-four neighborhood grids around the target point is not 0. By analogy, T is constructed in the manner1Corresponding preset nerveIn the network, because the weights of the neurons and the weights of the neurons in the preset neural network are 0, after a limited number of cycles, the preset neural network tends to be in a stable state, and the construction of the preset neural network is finished. Specifically, the steady state may include when the corresponding value of the activity information of each neuron at the last time and the next time is kept constant, that is, the neuron activity information of each grid at the last time and the neuron activity information of each grid at the next time are kept constant.
S309, taking the initial neural network corresponding to the current as the biological heuristic neural network corresponding to the target point traversed currently.
The target point to which the current traversal is continued is the target point T1For example, T, which is currently in steady state1Corresponding initial neural network as T1A corresponding bio-heuristic neural network.
In the examples of this specification, T1In a corresponding bio-heuristic neural network, T1The neuron activity information of (a) corresponds to the maximum value, generally, T1The neuron activity information of (A) is kept unchanged in the neural network construction process, namely T1The neuron activity information is the same as the preset neuron activity information; the corresponding numerical values of the neuron activity information of other grids are all smaller than T1And a value corresponding to the neuron activity information of (a), and T1The longer the shortest grid distance between the grids, the smaller its value corresponding to the neuron activity information.
According to and build a target point T1And respectively determining the biological enlightening neural networks corresponding to the target points by using a similar method of the corresponding biological enlightening neural networks.
And S107, distributing tasks to the target objects based on the initial coordinate positions of the target objects and the biological heuristic neural network corresponding to each target object, and determining an initial task path of each target object, wherein the initial task path is path information formed by grids of the initial target points corresponding to each target object.
In a specific embodiment, as shown in fig. 5, fig. 5 is a flowchart illustrating a task allocation method provided in an embodiment of the present specification, specifically, the method may include:
and S501, inputting the initial coordinate positions of the plurality of target objects into the bio-inspired neural network corresponding to each target point, respectively, to perform neuron activity information calculation, to obtain first neuron activity information of the plurality of target objects corresponding to each target point, where the first neuron activity information is used to map a shortest raster path length between each target object and the corresponding target point when the first neuron activity information is mapped to the initial coordinate positions, and a numerical value corresponding to the first neuron activity information is in an inverse relationship with the corresponding shortest raster path length.
S503, determining an initial target object corresponding to each target point based on the first neuron activity information of the plurality of target objects corresponding to each target point.
And S505, determining an initial target point corresponding to each target object according to the initial target object corresponding to each target point.
Specifically, as shown in formula (6), MjIs a set of first neuron activity information in the bio-heuristic neural network corresponding to the jth target point for the plurality of target objects, where Cj(xi,yi)Representing the first neuron activity information of the ith target object in the biological heuristic neural network corresponding to the jth target point. As shown in the formula (7), (x)max,ymax) Indicating that the first neuron activity information corresponding to the jth target point corresponds to the initial coordinate position of the target object having the largest numerical value after the task assignment is completed, max indicating identification information of the target object, U being a set of the initial coordinate positions of the plurality of target objects, and setting the max-th target object as the initial target object corresponding to the jth target point.
Continue with the target point T1For example, the initial coordinate positions of a plurality of target objects are input to T, respectively1Calculating neuron activity information by corresponding biological heuristic neural network to obtain T1First neuron activity information of a plurality of corresponding target objects, and in particular, first neuron activity information for mapping each target object with T at an initial coordinate position1And the corresponding numerical value of the first neuron activity information is in inverse proportion to the corresponding shortest raster path length. A plurality of target objects are taken as A1~A5For example, assume T1The first neuron activity information of the corresponding plurality of target objects is shown in table (1):
watch (1)
In particular, due to T1Corresponding biological heuristic neural network A1The first neuron activity information of (a) corresponds to the maximum value, namely A1And T1The distance between the shortest raster paths is less than other target objects and T1The shortest raster path distance between them, thus A1As target point T1The corresponding initial target object.
According to and at the target point T1Determining the initial target objects similarly, sequentially determining the initial target objects corresponding to a plurality of target points, as shown in table (2):
watch (2)
Integrating the table (2) to obtain the target object A1To A5The initial target points corresponding to the respective targets are shown in table (3):
watch (3)
S507, determining a traversal order of the initial target point corresponding to each target object based on the first neuron activity information corresponding to each target object and the initial target point.
In an alternative embodiment, as shown in fig. 6, the determining the traversal order of the initial target point corresponding to each target object based on the first neuron activity information corresponding to each target object and the initial target point may include:
s601, traversing the plurality of target objects.
And S603, taking the initial target point with the maximum numerical value corresponding to the first neuron activity information of the currently traversed target object as the preset target point of the currently traversed target object.
S605, using other target points except the preset target point in the initial target point corresponding to the currently traversed target object as remaining target points.
And S607, respectively inputting the coordinate positions of the residual target points into the biological heuristic neural network corresponding to the preset target point to calculate the neuron activity information, so as to obtain the first neuron activity information of the residual target points.
S609, updating the preset target point based on the largest initial target point in the first neuron activity information of the remaining target points.
S611, updating the remaining target points based on the updated preset target point.
S613, repeating the above-mentioned iterative process of inputting the coordinate positions of the remaining target points into the bio-inspired neural networks corresponding to the preset target points respectively to calculate the neuron activity information based on the updated remaining target points, and obtaining the first neuron activity information of the remaining target points to update the remaining target points based on the updated preset target points.
And S615, when the number of the current remaining target points is zero, taking the sequence of the preset target points determined in the iteration process as the traversal sequence of the initial target points.
The target object traversed at present is A4For example, the determination A is explained4The specific process of the corresponding traversal sequence of the initial target point is as follows:
1) as shown in Table (3), A4The corresponding initial target point position is T10、T11And T12Obtaining A4At T10、T11And T12Corresponding first neuronal activity information in the biological heuristic neural network, as shown in table (4):
watch (4)
2) A is to be4The initial target point T with the maximum value corresponding to the first neuron activity information10As A4The preset target point of (1);
in particular, due to T10Corresponding to A4Is the largest, so A is among the three initial target points4Distance T10Has the shortest raster path, therefore, T is10As A4The first preset target point to be traversed.
3) A is to be4Corresponding initial target point by T10The other target points are used as residual target points, namely the residual target points are T11And T12;
4) Will T11And T12Respectively inputting T10Calculating neuron activity information by corresponding biological heuristic neural network to obtain T11And T12The first neuronal activity information of (a);
specifically, T is determined10After the traversal order of T is first, T is added11And T12Middle distance T10The closer target point is designated as A4The second preset target point to be traversed, therefore, T needs to be calculated10Corresponding T in biological heuristic neural network11And T12To determine T11And T12And T10Distance.
5) Will T11And T12The target point with larger value corresponding to the activity information of the first neuron is taken as A4The second preset target point to be traversed is taken as A4A third preset target point to be traversed;
6) the current number of the remaining target points is zero, and T is determined in the iterative process10、T11And T12As the traversal order of the initial target point.
According to the above determination A4And respectively determining the traversal orders of the initial target points corresponding to the target objects by using a method similar to the traversal orders of the corresponding initial target points.
S509, generating the initial task path based on the traversal order of the initial target point.
Specifically, the shortest raster paths of any two adjacent initial target points in the initial target points may be determined based on the traversal order of the initial target points, and finally, the determined multiple shortest raster paths are synthesized to obtain the initial task path of each target object.
In a specific embodiment, TjAnd Tj+1Two adjacent initial target points of a target object, from TjTo Tj+1The shortest raster path may be represented by GjA path point which can be T-th-order for the target objectjTo Tj+1The target object selects T from eight neighborhood grids around the target object to be passed throughjThe grid with the maximum value corresponding to the corresponding first neuron activity information is used as the grid to be moved next, and the rest is done to obtain TjTo Tj+1G on the shortest raster path in betweenjA path point, the first path point on the shortest raster path, namely TjThe coordinate position of the last path point on the path, namely Tj+1The position of the coordinate is based on the GjA path point, determining TjTo Tj+1The shortest raster path. And in the same way, finally generating the initial task path of the target object.
As can be seen from the above embodiments, with the technical solution provided in this embodiment, a more reasonable initial task path can be generated by using the bio-heuristic neural network for task allocation of multiple target objects and multiple target points, and the resource utilization rate of each target object is improved.
And S109, controlling each target object to advance along the initial task path, taking the position information of the grid which passes through currently as the current coordinate position of each target object in the process that each target object advances along the initial task path, and determining the current non-passing target point of each target object.
In some embodiments, during the process that each target object advances along the initial task path, in response to a task adjustment instruction triggered by the management front end, position information of a currently passed grid may be used as a current coordinate position of each target object, and a current non-passed target point of each target object may be determined. Specifically, the management front-end may include a smart phone, a desktop computer, a tablet computer, a notebook computer, a digital assistant, a smart wearable device, and other types of physical devices, and may also include software running in the physical devices, such as an application program (APP), a web page, a wechat applet, and the like. The management front end can be used for managing task scheduling in warehouse logistics.
And S111, determining a target point to be adjusted in the current non-target points of each target object based on a path length ratio algorithm and path information formed by the grids where the current non-target points are located.
In practical application, target points to be adjusted in current target points which do not pass through target points of the target objects can be determined based on a path length ratio algorithm, and the target objects corresponding to the target points which are not reasonably allocated individually are adjusted through redistribution of the target points to be adjusted, so that the total length of the running paths of the target objects is reduced, and the transportation efficiency of warehouse logistics is improved on the basis of reducing resource consumption.
In an alternative embodiment, as shown in fig. 7, the determining the target point to be adjusted in the current non-target points of each target object based on the path length ratio algorithm and the path information formed by the raster where the current non-target point is located may include:
s701, traversing the target objects.
And S703, calculating the shortest raster path length between grids of any two adjacent current non-target points based on the traversal sequence of the current non-target points of the currently traversed target object.
S705, determining a first raster path length between the currently traversed target object and a corresponding first currently-unviewed target point based on the current coordinate position of the currently traversed target object.
And S707, when a ratio between a second raster path length and the first raster path length satisfies a preset ratio condition, adding a second currently-unviewed target point corresponding to the currently-traversed target object to the target point to be adjusted of the currently-traversed target object, where the second raster path length is a shortest raster path length between the corresponding first currently-unviewed target point and the corresponding second currently-unviewed target point.
S709, traversing a second current non-passed target point corresponding to the currently traversed target object to a penultimate current non-passed target point corresponding to the currently traversed target object based on the traversal order of the current non-passed target point of the currently traversed target object.
S711, when a ratio between a fourth raster path length and a third raster path length satisfies a preset ratio condition, adding a next current target point, which is located after a current target point, to a target point to be adjusted of the currently traversed target object, where the fourth raster path length is a shortest raster path length between the current target point and the next current target point, and the third raster path length is a shortest raster path length between the current target point and a previous current target point, which is located before the current target point.
Specifically, dist (T) may be usedj,Tj+1) Any two adjacent current non-passed target points T corresponding to the currently traversed target objectjAnd Tj+1Shortest raster path length between the grids, from TjTo Tj+1May be represented by GjA path point is formed, and the path point can be formed from T for the currently traversed target objectjTo Tj+1Selecting T from eight neighborhood grids around the position of the currently traversed target object by the grid needing to pass throughjThe grid with the maximum value corresponding to the corresponding first neuron activity information is used as the grid to be moved next, and the rest is done to obtain TjTo Tj+1G on the shortest raster path in betweenjAnd (4) path points.
LkRepresents TjAnd Tj+1The k-th path point, k being 1, 2, …, G on the shortest raster pathj。L1Representing the first path point on the shortest raster path, i.e. TjThe position of the coordinate; l isGjRepresenting the last path point on the shortest raster path, i.e. Tj+1The coordinate position of the location. (u)k,vk) Represents a path point Lk(u) coordinate position of (c)k+1,vk+1) Represents a path point Lk+1As shown in equation (8), the path point LkTo Lk+1European distance L between themk+1-LkAnd | l is:
as shown in formula (9), A5Corresponding any two adjacent current non-target points TjAnd Tj+1The shortest raster path length in between is:
αjshowing that the current non-target point is located on the path formed by the grid from TjTo Tj+1And the shortest grid path segment and the slave Tj-1To TjOne having SiThere are a total of Si-1 such length ratios on the paths that currently do not pass the target point. The first current target point which is not passed by the target object and is required to go is obtained according to task allocation, adjustment is not required, the length ratio is irrelevant, the corresponding length ratio is calculated from the second current target point which is not passed by the target object, whether the task allocation adjustment is required for the j +1 th current target point which is not passed by the target object on the path formed by the grid where the current target point is not passed by is judged according to the length ratio shown in the formula (10), and T0Is the current coordinate position of the currently traversed target object.
When a preset ratio condition is met, adding the j +1 th current target point which is not passed to the target point to be adjusted of the currently traversed target object, specifically, the preset ratio may be preset based on the precision requirement of path planning in the actual application, for example, the preset ratio condition may set αjIs > 2.
The target object traversed at present is AiFor example, assume AiIs Ta、TbAnd TcThe traversal order is Ta—Tb—Tc,Ta、TbAnd TcThe paths formed by the grids are shown in fig. 8. Based on the initial task path, AiFirst go to TaThen proceed to TbFinally go to Tc. First length ratio alpha1Express | | Tb-TaI and Ta-AiBetween the two raster path lengthsBased on a1Can judge TbWhether task allocation needs to be adjusted, when alpha1When the preset ratio condition is met, the T is setbAdding AiThe target point to be adjusted; second distance ratio alpha2Express | | Tc-TbI and Tb-AaThe ratio between the two raster path lengths, | | based on α2Can judge TcWhether task allocation needs to be adjusted, when alpha2When the preset ratio condition is met, the T is setcAdding AiThe target point to be adjusted.
As can be seen from the above embodiments, with the technical solution of this embodiment, a target point to be adjusted in target points, which is not currently passed through by a target object, can be determined based on a path length ratio algorithm, and target objects corresponding to respective unreasonably allocated target points are adjusted by reallocation of the target point to be adjusted, so as to generate a more reasonable task path.
And S113, performing task adjustment on the plurality of target objects based on the current coordinate positions of the plurality of target objects and the biological heuristic neural network corresponding to each target point to be adjusted in the target points to be adjusted of the plurality of target objects, and determining the current task path of each target object.
In an alternative embodiment, as shown in fig. 9, fig. 9 is a flowchart illustrating a task adjusting method, and specifically, the method may include:
s901, respectively inputting the current coordinate positions of the target objects into the bio-heuristic neural network corresponding to each target point to be adjusted to perform neuron activity information calculation, so as to obtain second neuron activity information of each target point to be adjusted corresponding to the target objects, where the second neuron activity information is used to map the shortest raster path length between each target object and the corresponding target point to be adjusted when the second neuron activity information is mapped at the current coordinate position, and a value corresponding to the second neuron activity information is in an inverse relationship with the corresponding shortest raster path length.
And S903, determining a current target object corresponding to each target point to be adjusted based on the second neuron activity information of the target objects corresponding to each target point to be adjusted.
S905, determining a current target point corresponding to each target object according to the current target object corresponding to each target point to be adjusted.
Specifically, the specific steps of performing task adjustment and determining the current target point corresponding to each target object are similar to the steps from "inputting the initial coordinate positions of the target objects into the bio-inspired neural network corresponding to each target point respectively to perform neuron activity information calculation in S501 to S505, and obtaining the first neuron activity information of the target objects corresponding to each target point" to "determining the initial target point corresponding to each target object according to the initial target object corresponding to each target point", and the specific steps may refer to the relevant descriptions in S501 to S505, which is not described herein again.
And S907, determining the traversal sequence of the current target point corresponding to each target object based on the second neuron activity information corresponding to each target object and the current target point.
Specifically, the specific steps for determining the traversal order of the current target point corresponding to each target object are similar to the steps for determining the traversal order of the initial target point corresponding to each target object in S601 to S615, and the specific steps may refer to the related descriptions in S601 to S615, and are not described herein again.
S909, a target point other than the target point to be adjusted of each target object among the current non-passing target points of each target object is taken as the target point to be passed of each target object.
And S911, generating the current task path based on the traversal order of the target point to be passed and the traversal order of the current target point.
Specifically, the specific step of generating the current task path of each target object here is similar to the step of generating the initial task path of each target object in S509, and the specific step may refer to the related description in S509, and is not described herein again.
And S115, taking the current task path of each target object as the target task path of each target object when the number of the current non-passing target points is zero.
Specifically, when the number of the current non-passed target points is zero, that is, after all the target points are traversed by a certain target object, the task adjustment is ended.
In an alternative embodiment, as shown in fig. 10, after the task adjustment is performed on the plurality of target objects based on the current coordinate positions of the plurality of target objects and the bio-heuristic neural network corresponding to each target point to be adjusted in the target points to be adjusted of the plurality of target objects, and the current task path of each target object is determined, the method may further include:
and S117, based on the current task path, repeatedly executing the step of controlling each target object to advance to the biological enlightening neural network corresponding to each target point to be adjusted based on the current coordinate positions of the plurality of target objects and the target points to be adjusted of the plurality of target objects to perform task adjustment on the plurality of target objects, and determining the current task path of each target object until the number of current target points which are not passed of each target object is zero.
In practical application, in the process that each target object advances along the current path, the target point corresponding to each target object can be repeatedly adjusted, dynamic allocation of the target points in path planning is realized, and the transportation efficiency of warehouse logistics is further improved while the resource loss of the AGV is reduced.
As can be seen from the above embodiments, according to the technical solutions provided by the embodiments of the present specification, dynamic task allocation and path planning of multiple AGVs are implemented by using a Bio-inspired Neural network and blending the ratios of lengths of raster paths between target points (DRGBNN).
An embodiment of the present application provides a path planning apparatus, as shown in fig. 11, the apparatus includes:
a target grid map obtaining module 1110, configured to obtain a target grid map;
a coordinate position determination module 1120 for determining coordinate positions of a plurality of target points and initial coordinate positions of a plurality of target objects in the target grid map;
a biological heuristic neural network constructing module 1130, configured to respectively construct a biological heuristic neural network corresponding to each target point based on preset neuron activity information and a coordinate position of each target point, where the preset neuron activity information represents preset activity state information of a neuron corresponding to each target point in the corresponding biological heuristic neural network;
a task allocation module 1140, configured to perform task allocation on the plurality of target objects based on the initial coordinate positions of the plurality of target objects and the bio-heuristic neural network corresponding to each target point, and determine an initial task path of each target object, where the initial task path is path information formed by a grid where the initial target point corresponding to each target object is located;
a target object control module 1150, configured to control each target object to advance along the initial task path, and in a process that each target object advances along the initial task path, take position information of a currently passing grid as a current coordinate position of each target object, and determine a current non-passing target point of each target object;
a target point to be adjusted determining module 1160, configured to determine a target point to be adjusted in the current non-target points of each target object based on a path length ratio algorithm and path information formed by a grid where the current non-target points are located;
a task adjustment module 1170, configured to perform task adjustment on the multiple target objects based on the current coordinate positions of the multiple target objects and a bio-heuristic neural network corresponding to each target point to be adjusted in the target points to be adjusted of the multiple target objects, and determine a current task path of each target object;
a target task path module 1180, configured to take the current task path of each target object as the target task path of each target object when the number of the current target points that have not passed is zero.
In an embodiment of the present disclosure, the biological heuristic neural network constructing module 1130 may include:
a target point traversing unit for traversing the plurality of target points;
the initial neural network construction unit is used for constructing an initial neural network corresponding to the currently traversed target point by taking a neuron corresponding to the currently traversed target point as an initial neuron;
an initialization setting unit, configured to perform initialization setting on the corresponding initial neural network according to a preset mapping relationship between each grid in the target grid map and each neuron in the corresponding initial neural network;
an iterative calculation unit, configured to input preset neuron activity information and a coordinate position of the currently traversed target point into the corresponding initial neural network to perform iterative calculation on the neuron activity information of each grid until the neuron activity information of each grid output by the corresponding initial neural network remains unchanged, where the neuron activity information is used to map a shortest grid path length between each grid and a grid where the currently traversed target point is located, and a value corresponding to the neuron activity information is in an inverse relationship with the corresponding shortest grid path length;
and the biological enlightening neural network determining unit is used for taking the currently corresponding initial neural network as the biological enlightening neural network corresponding to the currently traversed target point.
In a specific embodiment, the task assigning module 1140 may include:
a first neuron activity information calculation unit configured to input initial coordinate positions of the plurality of target objects into the biological heuristic neural network corresponding to each of the target points, respectively, to calculate neuron activity information, to obtain first neuron activity information of the plurality of target objects corresponding to each of the target points, wherein the first neuron activity information is used for mapping a shortest raster path length between each of the target objects and the corresponding target point when the first neuron activity information is mapped to the initial coordinate positions, and a numerical value corresponding to the first neuron activity information is in an inverse relationship with the corresponding shortest raster path length;
an initial target object specifying unit configured to specify an initial target object corresponding to each of the target points based on first neuron activity information of the plurality of target objects corresponding to each of the target points;
an initial target point determining unit, configured to determine an initial target point corresponding to each target object according to the initial target object corresponding to each target point;
an initial target point traversal order determination unit configured to determine a traversal order of an initial target point corresponding to each target object based on first neuron activity information of each target object corresponding to the initial target point;
and the initial task path generating unit is used for generating the initial task path based on the traversal sequence of the initial target point.
In an optional embodiment, the initial target point traversal order determining unit may include:
a first target object traversing unit for traversing the plurality of target objects;
a preset target point determining unit, configured to use an initial target point with a largest numerical value corresponding to first neuron activity information of a currently traversed target object as a preset target point of the currently traversed target object;
a remaining target point determining unit, configured to use, as a remaining target point, another target point, except the preset target point, in the initial target point corresponding to the currently traversed target object;
a neuron activity information calculation unit, configured to input the coordinate positions of the remaining target points to the bio-inspired neural network corresponding to the preset target point, respectively, to perform neuron activity information calculation, so as to obtain first neuron activity information of the remaining target points;
a preset target point updating unit for updating the preset target point based on a largest initial target point among the first neuron activity information of the remaining target points;
a remaining target point updating unit configured to update the remaining target point based on the updated preset target point;
an iteration execution unit, configured to repeat, based on the updated remaining target points, the above-mentioned biological heuristic neural network that respectively inputs the coordinate positions of the remaining target points to the preset target points to perform neuron activity information calculation, so as to obtain first neuron activity information of the remaining target points, and to perform an iteration process that updates the remaining target points based on the updated preset target points;
and a traversal order determining unit, configured to use, when the number of the current remaining target points is zero, the order of the preset target points determined in the iteration process as the traversal order of the initial target points.
In an alternative embodiment, the target point to be adjusted determining module 1160 may include:
a second target object traversing unit for traversing the plurality of target objects;
the path length calculation unit is used for calculating the shortest raster path length between grids of any two adjacent current non-target points based on the traversal sequence of the current non-target points of the currently traversed target object;
a first raster path length unit, configured to determine, based on a current coordinate position of the currently traversed target object, a first raster path length between the currently traversed target object and a corresponding first currently-unviewed target point;
a first target point to be adjusted determining unit, configured to add a second current non-passing target point corresponding to the currently traversed target object to a target point to be adjusted of the currently traversed target object when a ratio of a second raster path length to the first raster path length satisfies a preset ratio condition, where the second raster path length is a shortest raster path length between the corresponding first current non-passing target point and the corresponding second current non-passing target point;
a current non-target-point traversing unit, configured to traverse a second current non-target point corresponding to the currently traversed target object to a penultimate current non-target point corresponding to the currently traversed target object based on a traversal order of the current non-target point of the currently traversed target object;
a second target point to be adjusted determining unit, configured to add, when a ratio of a fourth raster path length to a third raster path length satisfies a preset ratio condition, a next current target point that is not passed after a current target point that is currently traversed to the target object to be adjusted of the currently traversed target object, where the fourth raster path length is a shortest raster path length between the current target point that is currently traversed to the target object that is not currently passed and the next current target point that is not currently passed, and the third raster path length is a shortest raster path length between the current target point that is currently traversed to the target object that is currently not passed before the current target point that is currently traversed to the target object that is currently passed.
In an alternative embodiment, the task adjustment module 1170 may comprise:
a second neuron activity information calculation unit, configured to input current coordinate positions of the target objects into a bio-heuristic neural network corresponding to each target point to be adjusted respectively for neuron activity information calculation, so as to obtain second neuron activity information of the target objects corresponding to each target point to be adjusted, where the second neuron activity information is used to map, when the current coordinate position is located, a shortest raster path length between each target object and the corresponding target point to be adjusted, and a value corresponding to the second neuron activity information is in an inverse relationship with the corresponding shortest raster path length;
a current target object determining unit, configured to determine a current target object corresponding to each target point to be adjusted based on second neuron activity information of the plurality of target objects corresponding to each target point to be adjusted;
a current target point determining unit, configured to determine, according to a current target object corresponding to each target point to be adjusted, a current target point corresponding to each target object;
a current target point traversal order determining unit, configured to determine, based on second neuron activity information corresponding to each target object and the current target point, a traversal order of the current target point corresponding to each target object;
a target point to be passed determination unit configured to take a target point of each of the target objects that is not currently passed, except for a target point to be adjusted of each of the target objects, as a target point to be passed of each of the target objects;
and the current task path generating unit is used for generating the current task path based on the traversal order of the target point to be passed and the traversal order of the current target point.
In an optional embodiment, the path planning apparatus may further include:
and a repeated execution module, configured to repeatedly execute, based on the current task path, the step of controlling each target object to advance to the biological heuristic neural network corresponding to each target point to be adjusted based on the current coordinate positions of the plurality of target objects and the target points to be adjusted of the plurality of target objects along the initial task path to perform task adjustment on the plurality of target objects, and determine the current task path of each target object until the number of current target points of each target object that have not passed is zero.
The embodiment of the present application provides a path planning apparatus, which includes a processor and a memory, where the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the path planning method provided by the above method embodiment.
The memory may be used to store software programs and modules, and the processor may execute various functional applications and data processing by operating the software programs and modules stored in the memory. The memory can mainly comprise a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs needed by functions and the like; the storage data area may store data created according to the use of the above-described apparatus, and the like. Further, the memory 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 volatile solid state storage device. Accordingly, the memory may also include a memory controller to provide the processor access to the memory.
The method embodiments provided in the embodiments of the present application may be executed in a mobile terminal, a computer terminal, a server, or a similar computing device, that is, the computer device may include a mobile terminal, a computer terminal, a server, or a similar computing device. Taking the example of running on a server, fig. 12 is a hardware structure block diagram of the server of the path planning method provided in the embodiment of the present application. As shown in fig. 12, the server 1200 may have a relatively large difference due to different configurations or performances, and may include one or more Central Processing Units (CPUs) 1210 (the processors 1210 may include but are not limited to Processing devices such as a microprocessor MCU or a programmable logic device FPGA), a memory 1230 for storing data, and one or more storage media 1220 (e.g., one or more mass storage devices) for storing applications 1223 or data 1222. Memory 1230 and storage media 1220, among other things, may be transient storage or persistent storage. The program stored in the storage medium 1220 may include one or more modules, each of which may include a series of instruction operations for a server. Further, the central processor 1210 may be configured to communicate with the storage medium 1220, and execute a series of instruction operations in the storage medium 1220 on the server 1200. The server 1200 may also include one or more power supplies 1260, one or more wired or wireless network interfaces 1250, one or more input-output interfaces 1240, and/or one or more operating systems 1221, such as Windows Server, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, and so forth.
The input/output interface 1240 may be used to receive or transmit data via a network. The specific example of the network described above may include a wireless network provided by a communication provider of the server 1200. In one example, the input/output Interface 1240 includes a Network Interface Controller (NIC) that may be coupled to other Network devices via a base station to communicate with the internet. In one example, the input/output interface 1240 may be a Radio Frequency (RF) module, which is used to communicate with the internet in a wireless manner.
It will be understood by those skilled in the art that the structure shown in fig. 12 is merely illustrative and is not intended to limit the structure of the electronic device. For example, server 1200 may also include more or fewer components than shown in FIG. 12, or have a different configuration than shown in FIG. 12.
The present application further provides a storage medium, where the storage medium may be disposed in a server to store at least one instruction or at least one program for implementing a path planning method in one of the method embodiments, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the path planning method provided in the method embodiment.
Alternatively, in this embodiment, the storage medium may be located in at least one network server of a plurality of network servers of a computer network. Optionally, in this embodiment, the storage medium may include, but is not limited to: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.
As can be seen from the above embodiments of the path planning method, apparatus, device, or storage medium provided by the present application, the technical solution provided by the present application implements dynamic task allocation and path planning for multiple AGVs by using a Bio-inspired Neural network and incorporating a Ratio of lengths of raster paths between target points (Distance Ratio Glasius Bio-interpolated Neural Networks, DRGBNN).
It should be noted that: the sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the apparatus, device and storage medium embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference may be made to some descriptions of the method embodiments for relevant points.
It will be understood by those skilled in the art that all or part of the steps of implementing the above embodiments may be performed by hardware, or may be performed by a program to instruct relevant hardware to perform the steps, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk, an optical disk, or the like.
The above description is only exemplary of the present application and should not be taken as limiting the present application, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.
Claims (10)
1. A method of path planning, the method comprising:
acquiring a target grid map;
determining coordinate positions of a plurality of target points and initial coordinate positions of a plurality of target objects in the target grid map;
respectively constructing a biological enlightening neural network corresponding to each target point based on preset neuron activity information and a coordinate position of each target point, wherein the preset neuron activity information represents preset activity state information of a neuron corresponding to each target point in the corresponding biological enlightening neural network;
performing task allocation on the plurality of target objects based on the initial coordinate positions of the plurality of target objects and the biological heuristic neural network corresponding to each target object, and determining an initial task path of each target object, wherein the initial task path is path information formed by grids of the initial target points corresponding to each target object;
controlling each target object to advance along the initial task path, taking the position information of the grid which passes through at present as the current coordinate position of each target object in the process that each target object advances along the initial task path, and determining the current target-passing point of each target object;
determining a target point to be adjusted in the current non-target points of each target object based on a path length ratio algorithm and path information formed by a grid where the current non-target points are located;
task adjustment is carried out on the multiple target objects based on the current coordinate positions of the multiple target objects and a biological heuristic neural network corresponding to each target point to be adjusted in the target points to be adjusted of the multiple target objects, and a current task path of each target object is determined;
and taking the current task path of each target object as the target task path of each target object under the condition that the number of the current non-passing target points is zero.
2. The method of claim 1, wherein the separately constructing the bio-heuristic neural network corresponding to each target point based on the preset neuron activity information and the coordinate position of each target point comprises:
traversing the plurality of target points;
taking a neuron corresponding to a currently traversed target point as an initial neuron, and constructing an initial neural network corresponding to the currently traversed target point;
initializing the corresponding initial neural network according to a preset mapping relation between each grid in the target grid map and each neuron in the corresponding initial neural network;
inputting preset neuron activity information and coordinate positions of the currently traversed target point into the corresponding initial neural network to perform iterative computation on the neuron activity information of each grid until the neuron activity information of each grid output by the corresponding initial neural network is kept unchanged, wherein the neuron activity information is used for mapping the shortest grid path length between each grid and the grid where the currently traversed target point is located, and a numerical value corresponding to the neuron activity information is in an inverse proportion relation with the corresponding shortest grid path length;
and taking the currently corresponding initial neural network as the biological heuristic neural network corresponding to the currently traversed target point.
3. The method of claim 1 or 2, wherein the task assigning the plurality of target objects based on the initial coordinate locations of the plurality of target objects and the bio-heuristic neural network corresponding to each target point, wherein determining the initial task path for each target object comprises:
respectively inputting the initial coordinate positions of the multiple target objects into the biological enlightening neural network corresponding to each target point to calculate neuron activity information, so as to obtain first neuron activity information of the multiple target objects corresponding to each target point, wherein the first neuron activity information is used for mapping the shortest raster path length between each target object and the corresponding target point when the first neuron activity information is mapped at the initial coordinate positions, and the numerical value corresponding to the first neuron activity information is in an inverse proportion relation with the corresponding shortest raster path length;
determining an initial target object corresponding to each target point based on first neuron activity information of the plurality of target objects corresponding to the each target point;
determining an initial target point corresponding to each target object according to the initial target object corresponding to each target point;
determining a traversal order of an initial target point corresponding to each target object based on first neuron activity information of each target object corresponding to the initial target point;
and generating the initial task path based on the traversal sequence of the initial target point.
4. The method of claim 3, wherein the determining the traversal order of the initial target point corresponding to each target object based on the first neuron activity information corresponding to each target object to the initial target point comprises:
traversing the plurality of target objects;
taking the initial target point with the maximum numerical value corresponding to the first neuron activity information of the currently traversed target object as a preset target point of the currently traversed target object;
taking other target points except the preset target point in the initial target point corresponding to the currently traversed target object as residual target points;
respectively inputting the coordinate positions of the residual target points into a biological heuristic neural network corresponding to the preset target point to calculate neuron activity information, so as to obtain first neuron activity information of the residual target points;
updating the preset target point based on the largest initial target point in the first neuron activity information of the remaining target points;
updating the remaining target points based on the updated preset target points;
repeating the iterative process of inputting the coordinate positions of the residual target points into the biological heuristic neural network corresponding to the preset target point respectively to calculate the neuron activity information based on the updated residual target points, and obtaining the first neuron activity information of the residual target points to update the residual target points based on the updated preset target points;
and when the number of the current residual target points is zero, taking the sequence of the preset target points determined in the iteration process as the traversal sequence of the initial target points.
5. The method according to claim 1, wherein the determining the target point to be adjusted in the current non-target points of each target object based on a path length ratio algorithm and path information formed by a grid where the current non-target points are located comprises:
traversing the plurality of target objects;
calculating the shortest raster path length between grids of any two adjacent current non-passed target points based on the traversal sequence of the current non-passed target points of the currently traversed target object;
determining a first raster path length between the currently traversed target object and a corresponding first current non-traversed target point based on the current coordinate position of the currently traversed target object;
when the ratio of a second raster path length to the first raster path length meets a preset ratio condition, adding a second current non-passing target point corresponding to the currently traversed target object to a target point to be adjusted of the currently traversed target object, wherein the second raster path length is the shortest raster path length between the corresponding first current non-passing target point and the corresponding second current non-passing target point;
traversing a second current non-target point corresponding to the currently traversed target object to a penultimate current non-target point corresponding to the currently traversed target object based on a traversal order of current non-target points of the currently traversed target object;
when the ratio of a fourth raster path length and a third raster path length meets a preset ratio condition, adding a next current target point which is not passed after a current target point which is traversed to be currently to a target point to be adjusted of the currently traversed target object, wherein the fourth raster path length is the shortest raster path length between the current target point which is traversed to be currently not passed and the next current target point which is not passed, and the third raster path length is the shortest raster path length between the current target point which is traversed to be currently not passed and the last current target point which is traversed to be currently to be not passed before the current target point which is traversed to be currently.
6. The method of claim 1, wherein the task adjustment of the plurality of target objects based on the current coordinate locations of the plurality of target objects and the biological heuristic neural network corresponding to each of the target points to be adjusted of the plurality of target objects, the determining the current task path of each target object comprising:
respectively inputting the current coordinate positions of the target objects into a biological heuristic neural network corresponding to each target point to be adjusted to calculate neuron activity information, so as to obtain second neuron activity information of each target point to be adjusted corresponding to the target objects, wherein the second neuron activity information is used for mapping the shortest raster path length between each target object and the corresponding target point to be adjusted when the second neuron activity information is mapped at the current coordinate position, and the numerical value corresponding to the second neuron activity information and the corresponding shortest raster path length are in an inverse proportion relation;
determining a current target object corresponding to each target point to be adjusted based on second neuron activity information of the target objects corresponding to each target point to be adjusted;
determining a current target point corresponding to each target object according to the current target object corresponding to each target point to be adjusted;
determining a traversal order of a current target point corresponding to each target object based on second neuron activity information corresponding to each target object and the current target point;
taking a target point of the current non-passing target point of each target object except the target point to be adjusted of each target object as the target point to be passed of each target object;
and generating the current task path based on the traversal sequence of the target point to be passed and the traversal sequence of the current target point.
7. The method of claim 6, further comprising:
and repeatedly executing the step of controlling each target object to advance to the biological enlightening neural network corresponding to each target point to be adjusted in the target points to be adjusted based on the current coordinate positions of the plurality of target objects and the plurality of target objects based on the initial task path based on the current task path to perform task adjustment on the plurality of target objects, and determining the current task path of each target object until the number of current target points which are not passed by each target object is zero.
8. A path planning apparatus, the apparatus comprising:
the target grid map acquisition module is used for acquiring a target grid map;
a coordinate position determination module for determining coordinate positions of a plurality of target points and initial coordinate positions of a plurality of target objects in the target grid map;
the biological elicitation neural network construction module is used for respectively constructing a biological elicitation neural network corresponding to each target point based on preset neuron activity information and coordinate positions of each target point, wherein the preset neuron activity information represents preset activity state information of neurons corresponding to each target point in the corresponding biological elicitation neural network;
the task allocation module is used for allocating tasks to the target objects based on the initial coordinate positions of the target objects and the biological heuristic neural network corresponding to each target point, and determining an initial task path of each target object, wherein the initial task path is path information formed by grids where the initial target points corresponding to each target object are located;
the target object control module is used for controlling each target object to advance along the initial task path, taking the position information of the grid which passes through at present as the current coordinate position of each target object in the process that each target object advances along the initial task path, and determining the current target point which does not pass through of each target object;
a target point to be adjusted determining module, configured to determine, based on a path length ratio algorithm and path information formed by a grid where the current target point that has not passed is located, a target point to be adjusted in the current target point that has not passed of each target object;
the task adjusting module is used for performing task adjustment on the plurality of target objects based on the current coordinate positions of the plurality of target objects and the biological heuristic neural network corresponding to each target point to be adjusted in the target points to be adjusted of the plurality of target objects, and determining the current task path of each target object;
and the target task path module is used for taking the current task path of each target object as the target task path of each target object under the condition that the number of the current non-passing target points is zero.
9. A path planning apparatus, characterized in that the apparatus comprises a processor and a memory, in which at least one instruction or at least one program is stored, which is loaded and executed by the processor to implement the path planning method according to any one of claims 1 to 7.
10. A computer-readable storage medium, in which at least one instruction or at least one program is stored, which is loaded and executed by a processor to implement the path planning method according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110776397.3A CN113587931B (en) | 2021-07-08 | 2021-07-08 | Path planning method, device, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110776397.3A CN113587931B (en) | 2021-07-08 | 2021-07-08 | Path planning method, device, equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113587931A true CN113587931A (en) | 2021-11-02 |
CN113587931B CN113587931B (en) | 2024-07-09 |
Family
ID=78246675
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110776397.3A Active CN113587931B (en) | 2021-07-08 | 2021-07-08 | Path planning method, device, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113587931B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118003342A (en) * | 2024-04-09 | 2024-05-10 | 北京科技大学 | Path planning method for space crawling mechanical arm |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007257094A (en) * | 2006-03-22 | 2007-10-04 | Toyota Motor Corp | Method for creating path, moving element and moving element control system |
KR20110026776A (en) * | 2009-09-08 | 2011-03-16 | 부산대학교 산학협력단 | Method for path-planning for actual robots |
US20120254153A1 (en) * | 2011-03-31 | 2012-10-04 | Microsoft Corporation | Shortest path determination in databases |
JP2014174048A (en) * | 2013-03-11 | 2014-09-22 | Aisin Aw Co Ltd | Guide route search system |
CN111307153A (en) * | 2020-02-26 | 2020-06-19 | 河海大学 | Multi-AUV task allocation and path planning method based on hexagonal grid map |
CN112215465A (en) * | 2020-09-04 | 2021-01-12 | 燕山大学 | Auction model-based distributed robust heterogeneous multi-AUV task allocation method |
CN112465127A (en) * | 2020-11-29 | 2021-03-09 | 西北工业大学 | Multi-agent cooperative target searching method based on improved biological heuristic neural network |
CN112529235A (en) * | 2019-09-18 | 2021-03-19 | 北京三快在线科技有限公司 | Path determining method and device, electronic equipment and computer readable storage medium |
CN112673234A (en) * | 2020-01-17 | 2021-04-16 | 华为技术有限公司 | Path planning method and path planning device |
CN112947457A (en) * | 2021-02-25 | 2021-06-11 | 薛明 | Dynamic wireless power supply path planning method for sorting robot groups |
CN113033970A (en) * | 2021-03-05 | 2021-06-25 | 西北工业大学 | AUV cluster layered distribution method for large-scale tasks |
US20220081001A1 (en) * | 2020-09-15 | 2022-03-17 | Beijing Baidu Netcom Science Technology Co., Ltd. | Method and apparatus of determining guide path, method and apparatus of controlling driving of vehicle, and electronic device |
-
2021
- 2021-07-08 CN CN202110776397.3A patent/CN113587931B/en active Active
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007257094A (en) * | 2006-03-22 | 2007-10-04 | Toyota Motor Corp | Method for creating path, moving element and moving element control system |
KR20110026776A (en) * | 2009-09-08 | 2011-03-16 | 부산대학교 산학협력단 | Method for path-planning for actual robots |
US20120254153A1 (en) * | 2011-03-31 | 2012-10-04 | Microsoft Corporation | Shortest path determination in databases |
JP2014174048A (en) * | 2013-03-11 | 2014-09-22 | Aisin Aw Co Ltd | Guide route search system |
CN112529235A (en) * | 2019-09-18 | 2021-03-19 | 北京三快在线科技有限公司 | Path determining method and device, electronic equipment and computer readable storage medium |
CN112673234A (en) * | 2020-01-17 | 2021-04-16 | 华为技术有限公司 | Path planning method and path planning device |
CN111307153A (en) * | 2020-02-26 | 2020-06-19 | 河海大学 | Multi-AUV task allocation and path planning method based on hexagonal grid map |
CN112215465A (en) * | 2020-09-04 | 2021-01-12 | 燕山大学 | Auction model-based distributed robust heterogeneous multi-AUV task allocation method |
US20220081001A1 (en) * | 2020-09-15 | 2022-03-17 | Beijing Baidu Netcom Science Technology Co., Ltd. | Method and apparatus of determining guide path, method and apparatus of controlling driving of vehicle, and electronic device |
CN112465127A (en) * | 2020-11-29 | 2021-03-09 | 西北工业大学 | Multi-agent cooperative target searching method based on improved biological heuristic neural network |
CN112947457A (en) * | 2021-02-25 | 2021-06-11 | 薛明 | Dynamic wireless power supply path planning method for sorting robot groups |
CN113033970A (en) * | 2021-03-05 | 2021-06-25 | 西北工业大学 | AUV cluster layered distribution method for large-scale tasks |
Non-Patent Citations (4)
Title |
---|
YUE XIU; ZHANG WEI: "UAV path planning based on intelligent algorithm", ELECTRONIC SCIENCE AND TECHNOLOGY, vol. 32, no. 2, pages 9 - 13 * |
倪斌;陈雄;鲁公羽;: "基于神经网络的未知环境路径规划算法研究", 计算机工程与应用, no. 11, 11 April 2006 (2006-04-11), pages 73 - 76 * |
刘雨;朱大奇: "基于生物启发自组织神经网络的任务分配与路径规划", ***仿真技术, vol. 13, no. 3, pages 230 - 234 * |
陈少斌;陈冲;: "动态环境下移动机器人路径规划的一种新方法", 福州大学学报(自然科学版), vol. 37, no. 05, 28 October 2009 (2009-10-28), pages 701 - 704 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118003342A (en) * | 2024-04-09 | 2024-05-10 | 北京科技大学 | Path planning method for space crawling mechanical arm |
Also Published As
Publication number | Publication date |
---|---|
CN113587931B (en) | 2024-07-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Giordani et al. | A distributed multi-agent production planning and scheduling framework for mobile robots | |
Li et al. | Tasks assigning and sequencing of multiple AGVs based on an improved harmony search algorithm | |
Zhu et al. | A neural network approach to dynamic task assignment of multirobots | |
EP3605334A1 (en) | Incentive control for multi-agent systems | |
Alitappeh et al. | Multi-robot exploration in task allocation problem | |
Zhou et al. | A novel path planning algorithm based on plant growth mechanism | |
CN110031007B (en) | Flight path planning method and device and computer readable storage medium | |
CN110688920A (en) | Unmanned control method and device and server | |
CN110766145A (en) | Learning task compiling method of artificial intelligence processor and related product | |
CN111798097B (en) | Autonomous mobile robot task allocation processing method based on market mechanism | |
Couceiro et al. | Introducing the fractional order robotic Darwinian PSO | |
Luo et al. | A multi-scale map method based on bioinspired neural network algorithm for robot path planning | |
Pascarella et al. | Parallel and distributed computing for UAVs trajectory planning | |
CN113587931A (en) | Path planning method, device, equipment and storage medium | |
Wang et al. | Cooperative task allocation for multi-robot systems based on multi-objective ant colony system | |
CN110766146B (en) | Learning task compiling method of artificial intelligence processor and related product | |
Kravets et al. | Markovian learning methods in decision-making systems | |
Yang et al. | Risk-aware motion planning for very-large-scale robotics systems using conditional value-at-risk | |
CN113627646B (en) | Path planning method, device, equipment and medium based on neural network | |
Ho et al. | Preference-based multi-objective multi-agent path finding | |
Skobtsov et al. | Synthesis production schedules based on ant colony optimization method | |
Pavel et al. | Control of open mobile robotic platform using deep reinforcement learning | |
CN113627648A (en) | Task allocation method, device, equipment and storage medium | |
CN113687657B (en) | Method and storage medium for multi-agent formation dynamic path planning | |
Kim et al. | Motion optimization using state-dispersion-based phase space partitions |
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 |