Storage AGV operation path intelligent management and control system based on Internet of Things

Abstract

The invention discloses a storage AGV operation path intelligent management and control system based on the Internet of Things, and relates to the technical field of intelligent storage. The method comprises the following steps: dividing a warehouse into a plurality of grid areas, obtaining priority coefficients of AGVs, carrying out path planning on the AGVs to obtain an initial operation path, obtaining a grid state of each grid area, judging whether path conflicts exist or not, carrying out path optimization and equipment optimization on each AGV with the path conflicts, and obtaining an optimal optimization scheme. An intelligent management and control model is constructed, a passing area of each AGV is obtained, and subsequent path planning is carried out on each AGV; the method can effectively detect a potential path conflict problem, can improve the operation efficiency of AGVs in a warehouse on the whole, and can directly obtain an optimal optimization scheme corresponding to subsequent different path conflicts.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent warehousing, and specifically to an intelligent control system for the operation path of an AGV for warehousing based on the Internet of Things. Background Art

[0002] Intelligently controlling the operation path of an AGV for warehousing is a technology that combines Internet of Things technology with AGV operation management. Through various sensors, communication networks, and intelligent algorithms, it realizes real-time monitoring, path planning, and dynamic adjustment of the AGV in the warehouse, thereby improving the automation level, operation efficiency, and safety of warehousing logistics;

[0003] In the prior art, attention is often only paid to the current operation path of the AGV, while ignoring its subsequent operation position, resulting in the failure to predict in advance the path conflicts that will occur, and thus unable to avoid them in advance. Moreover, in the prior art, there is a lack of an effective detection method for the passing frequency of different areas in the warehouse, making the passing frequency of some areas too high, more likely to have path conflicts, and thus affecting the overall operation efficiency. In view of the deficiencies of the prior art, the present invention provides an intelligent control system for the operation path of an AGV for warehousing based on the Internet of Things. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent control system for the operation path of an AGV for warehousing based on the Internet of Things.

[0005] The purpose of the present invention can be achieved through the following technical solutions: An intelligent control system for the operation path of an AGV for warehousing based on the Internet of Things, including the following modules:

[0006] A map construction module, used to obtain the basic information of the warehouse, construct a corresponding GIS distribution map, and perform grid division on the GIS distribution map to obtain a number of grid areas;

[0007] A device evaluation module, used to respectively obtain the operation status and operation information of the AGV, and obtain the priority coefficient of the AGV according to the operation information;

[0008] A path planning module, used to perform path planning on the AGV according to the operation status to obtain a corresponding initial operation path, combine the operation information and the initial operation path to obtain the grid status of each grid area, and determine whether there are path conflicts between different AGVs;

[0009] A first optimization module, used to respectively perform path optimization and device optimization on each AGV with path conflicts, and obtain corresponding optimal optimization solutions, and construct a corresponding intelligent control model according to different path conflicts and their optimal optimization solutions;

[0010] The second optimization module is used to obtain the conflict coefficient of each grid area, and combine the priority coefficient to obtain the passing area of each AGV, and use the passing area and the intelligent control model to perform subsequent path planning for each AGV.

[0011] Further, the process of obtaining the basic information of the warehouse and constructing the corresponding GIS distribution map, and dividing the GIS distribution map into grids to obtain several grid areas includes:

[0012] The basic information includes the warehouse boundary, shelf location, operation area, ground condition, and driving area. Use GIS technology to construct the GIS distribution map of the warehouse according to the obtained basic information;

[0013] In the GIS distribution map, the area where all AGVs can drive is used as the control area, and the control area in the GIS distribution map is divided into several grid areas according to the equipment size of the AGV.

[0014] Further, the process of respectively obtaining the running state and running information of the AGV, and obtaining the priority coefficient of the AGV according to the running information includes:

[0015] The running state of the AGV in the warehouse is monitored in real time. The running state includes the working state and the non-working state. The running information of the AGV in the working state in the warehouse is collected in real time. The running information includes the current position, delivery position, driving speed, battery power, and load weight;

[0016] Let the driving speed be Y a 、the battery power be Y b 、the load weight be Y c Correspondingly set the weight parameters Q a 、Q b 、Q c to obtain the priority coefficient R of the AGV;

[0017]

[0018] Further, the process of performing path planning for the AGV according to the running state to obtain the corresponding initial running path includes:

[0019] Perform path planning for all AGVs in the working state respectively, obtain the current position and delivery position of each AGV, and use the Dijkstra algorithm to construct an initial running path between the current position and the delivery position of each AGV.

[0020] Further, the process of combining the running information and the initial running path to obtain the grid state of each grid area and determining whether there is a path conflict between different AGVs includes:

[0021] Mark the grid area where the current position of the corresponding AGV is located as occupied according to the operation information, set the prediction period, and obtain the predicted positions of the corresponding AGV in the subsequent several prediction periods on its current running path according to the current position and driving speed;

[0022] Mark the grid areas where the respective predicted positions of the same AGV are located as to-be-occupied states, and use the corresponding moments when the AGV arrives at each grid area as its to-be-occupied time nodes. Respectively obtain the predicted positions of different AGVs in the subsequent several prediction periods on their current running paths, and mark the grid areas corresponding to all the predicted positions;

[0023] Judge whether different AGVs exist simultaneously in a single grid area in the to-be-occupied state at the same to-be-occupied time node. If so, judge that there is a path conflict, mark the grid area as a conflict area, and include the corresponding different AGVs in the same conflict device combination.

[0024] Further, the process of respectively performing path optimization and device optimization on each AGV with a path conflict and obtaining the corresponding best optimization scheme includes:

[0025] Obtain the priority coefficients of each AGV in a single conflict device combination, let the AGV with the largest priority coefficient continue to drive along its original running path, and perform path optimization and device optimization on the other AGVs in the conflict device combination;

[0026] The path optimization refers to taking the current position of the corresponding AGV as the starting point and using the Dijkstra algorithm to construct an optimized running path again between the current position and the delivery position of the AGV;

[0027] The device optimization refers to reducing the driving speed of the AGV so that it arrives at the conflict area after the to-be-occupied time node and then resumes its original driving speed;

[0028] According to the driving speed and running path of the same AGV after path optimization and device optimization, predict the corresponding moment when the AGV arrives at the delivery position under path optimization and device optimization, and take the optimization scheme with the shortest time consumption as the best optimization scheme.

[0029] Further, the process of constructing a corresponding intelligent control model according to different path conflicts and their best optimization schemes includes:

[0030] Obtain the data sets of each AGV under different path conflicts and their best optimization schemes. The data set includes the conflict areas, priority coefficients, driving speeds, running paths corresponding to each AGV, and the data sets of other AGVs in their conflict device combinations;

[0031] Generate an intelligent control set based on the data sets of each AGV and their optimal optimization solutions under different path conflicts, and divide the obtained intelligent control set into a training set and a test set;

[0032] Construct a convolutional neural network. Use the data sets of different AGVs in the training set as the input data of the convolutional neural network, and use the corresponding optimal optimization solutions in the training set as the output data of the convolutional neural network. Train the convolutional neural network to obtain an initial convolutional neural network;

[0033] Use the test set to verify the model of the initial convolutional neural network, and output the initial convolutional neural network whose test error threshold is less than or equal to the preset value as the corresponding intelligent control model.

[0034] Furthermore, obtain the conflict coefficient of each grid area, and combine the priority coefficient to obtain the passing area of each AGV. The process of subsequent path planning for each AGV using the passing area and the intelligent control model includes:

[0035] Set an analysis period. When an analysis period is reached, obtain the conflict coefficient of each grid area in the most recent analysis period. Multiply the priority coefficient of a single AGV at the same moment by the conflict coefficient of a single grid area as their passing coefficient;

[0036] Set a passing threshold. If the passing coefficient is less than or equal to the passing threshold, mark the grid area as the passing area of the AGV. Respectively obtain the passing coefficients between each AGV and each grid area at the same moment, and respectively obtain the passing areas of each AGV. Different AGVs are only allowed to perform path planning and path optimization within their passing areas during their corresponding analysis periods;

[0037] When it is determined that an AGV has a path conflict, obtain its data set and input it into the intelligent control model. Use the intelligent control model to output the corresponding optimal optimization solution, and control the AGV to continue driving according to its optimal optimization solution.

[0038] Compared with the prior art, the beneficial effects of the present invention are:

[0039] The present invention divides the areas where AGVs can pass in the warehouse into several grid areas, and ensures that each grid area can accommodate a single AGV. It can convert complex path planning problems into simple allocation problems of grid areas. Based on each grid area, it judges whether different AGVs will exist in a single grid area at the same time, and can effectively detect potential path conflict problems. By obtaining the priority coefficient of the AGV, the one with the larger priority coefficient continues to drive along the original running path, and different optimization methods are respectively adopted for those with smaller priority coefficients;

[0040] According to the parameters of AGVs under different path conflicts and their final optimal optimization solutions, constructing corresponding intelligent control models is beneficial to directly obtain the optimal optimization solutions corresponding to subsequent different path conflicts. And according to the conflict coefficients of different grid regions in the same time period, combined with the priority coefficients of AGVs, the passing areas for different AGVs are specified, which can alleviate the problem of excessive passing frequency in some grid regions, and then improve the operation efficiency of AGVs in the warehouse as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is the schematic diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0042] As Figure 1 shown, an intelligent control system for the operation path of an AGV for warehousing based on the Internet of Things includes the following modules:

[0043] A map construction module for obtaining the basic information of the warehouse, constructing a corresponding GIS distribution map, and dividing the GIS distribution map into grids to obtain several grid regions;

[0044] An equipment evaluation module for respectively obtaining the operation status and operation information of the AGV, and obtaining the priority coefficient of the AGV according to the operation information;

[0045] A path planning module for performing path planning on the AGV according to the operation status to obtain a corresponding initial operation path, combining the operation information and the initial operation path to obtain the grid status of each grid region, and determining whether there is a path conflict between different AGVs;

[0046] A first optimization module for respectively performing path optimization and equipment optimization on each AGV with a path conflict, and obtaining a corresponding optimal optimization solution, and constructing a corresponding intelligent control model according to different path conflicts and their optimal optimization solutions;

[0047] A second optimization module for obtaining the conflict coefficient of each grid region, and obtaining the passing area of each AGV in combination with the priority coefficient, and performing subsequent path planning on each AGV by using the passing area and the intelligent control model.

[0048] It should be further noted that in the specific implementation process, the process of obtaining the basic information of the warehouse, constructing a corresponding GIS distribution map, and dividing the GIS distribution map into grids to obtain several grid regions includes:

[0049] The basic information refers to the various data necessary for constructing the GIS distribution map of the warehouse, including the warehouse boundary, shelf location, operation area, ground condition, driving area, etc. The GIS distribution map of the warehouse is constructed by using GIS technology according to the obtained basic information;

[0050] In the GIS distribution map, the area where all AGVs can travel is regarded as its control area, and the equipment size of a single AGV is obtained. The equipment size is used to reflect the occupied area of the AGV. The control area in the GIS distribution map is divided into grids according to the equipment size of the AGV, and the entire control area is divided into several grid areas, and the area of each grid area is larger than the equipment size of each AGV.

[0051] It should be further noted that in the specific implementation process, the process of respectively obtaining the operating status and operating information of the AGV and obtaining the priority coefficient of the AGV according to the operating information includes:

[0052] The operating status of each AGV in the warehouse is monitored in real time. All AGVs that are in motion are marked as the working state, and all other AGVs in the idle, charging, loading and unloading, etc. situations are marked as the non-working state;

[0053] The operating information of each AGV in the warehouse that is in the working state is collected in real time. The operating information includes the current position, delivery position, traveling speed, battery power, and carrying weight;

[0054] Mark the traveling speed, battery power, and carrying weight among them as Y a 、Y b 、Y c respectively, and set corresponding weight parameters for them, denoted as Q a 、Q b 、Q c respectively, and obtain the priority coefficient of each AGV, denoted as R;

[0055]

[0056] It should be further noted that in the specific implementation process, the process of performing path planning on the AGV according to the operating status to obtain the corresponding initial operating path includes:

[0057] Path planning is performed on all AGVs in the working state respectively. Taking any one AGV as an example, obtain the current position and delivery position of the AGV, and use the Dijkstra algorithm to construct an initial operating path between the current position and the delivery position of the AGV. The specific steps are as follows:

[0058] Respectively take the centers of each grid area as the corresponding planning nodes, initialize the distances from the current position to each planning node to infinity, initialize the distance from the current position to the AGV itself to 0, and create an empty priority queue to store the unprocessed planning nodes;

[0059] Select the planning node closest to the current position from the priority queue, use the selected planning node as the preferred node, and remove it from the priority queue;

[0060] For each adjacent planning node of the preferred node, obtain the distance from the current position to this adjacent planning node passing through the preferred node. If the distance obtained by passing through the preferred node is shorter than the distance recorded for this adjacent planning node, update the distance of this adjacent planning node;

[0061] Repeat the above steps until all planning nodes have been visited or the delivery location has been included. Starting from the delivery location, according to the shortest distance precursor planning node recorded for each planning node, backtrack in reverse order to obtain the corresponding initial operation path;

[0062] The initial operation path must satisfy two conditions. First, the grid areas corresponding to the respective planning nodes that make up the initial operation path must be connected end to end. Second, the grid areas corresponding to the respective planning nodes that make up the initial operation path must allow the AGV to pass.

[0063] It should be further noted that in the specific implementation process, the process of obtaining the grid status of each grid area in combination with the operation information and the initial operation path and judging whether there is a path conflict between different AGVs includes:

[0064] Mark the grid area where the current position of a single AGV is located as the occupied state according to the operation information, set the prediction period, and obtain the predicted positions of the single AGV in the subsequent several prediction periods on its current operation path based on the current position and the driving speed. The predicted positions are used to reflect the positions that the single AGV will reach at subsequent time nodes according to its driving speed;

[0065] Mark the grid areas where the respective predicted positions of the same AGV are located as the to-be-occupied state, and obtain the to-be-occupied time nodes. The to-be-occupied time nodes are used to reflect the corresponding moments when the corresponding AGV reaches the grid area. Respectively obtain the predicted positions of different AGVs in the subsequent several prediction periods on their current operation paths, and mark the grid areas corresponding to all the predicted positions;

[0066] Judge whether there are different AGVs in the same to-be-occupied time node for a single grid area in the to-be-occupied state. If so, judge that there is a path conflict, mark the grid area as a conflict area, and include the corresponding different AGVs in the same conflict device combination. If not, judge that there is no path conflict and do not perform any other operations on it.

[0067] It should be further noted that in the specific implementation process, the process of separately optimizing the paths and devices of each AGV with path conflicts and obtaining the corresponding optimal optimization solutions includes:

[0068] Taking any conflict device combination as an example, obtain the priority coefficients of each AGV within the conflict device combination, and let the AGV with the largest priority coefficient continue to travel along its original running path, and perform path optimization and device optimization on the other AGVs within the conflict device combination;

[0069] The path optimization refers to taking the current position of the corresponding AGV as the starting point and using the Dijkstra algorithm to reconstruct an optimized running path between the current position and the delivery position of the AGV;

[0070] The device optimization refers to reducing the running speed of the AGV so that it arrives at the conflict area after the time node to be occupied, and then restoring its original running speed;

[0071] According to the running speed and running path (including the initial running path or the optimized running path) of the same AGV after path optimization and device optimization, predict the corresponding moment when the AGV arrives at the delivery position under path optimization and device optimization, and take the optimization solution with the shortest time consumption as the optimal optimization solution, and control the corresponding AGV to continue to travel according to the optimal optimization solution.

[0072] It should be further noted that in the specific implementation process, the process of constructing the corresponding intelligent control model according to different path conflicts and their optimal optimization solutions includes:

[0073] Obtain the data sets of each AGV under different path conflicts and their corresponding optimal optimization solutions. The data sets include the conflict areas, priority coefficients, running speeds, running paths corresponding to each AGV, and the data sets of other AGVs within their conflict device combinations;

[0074] According to the data sets of each AGV under different path conflicts and their corresponding optimal optimization solutions, generate an intelligent control set, and divide the obtained intelligent control set into a training set and a test set;

[0075] Construct a convolutional neural network, take the data sets of different AGVs in the training set as the input data of the convolutional neural network, take the corresponding optimal optimization solutions in the training set as the output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network;

[0076] Use the test set to verify the model of the initial convolutional neural network, and output the initial convolutional neural network with a test error threshold less than or equal to the preset value as the corresponding intelligent control model.

[0077] It should be further noted that in the specific implementation process, the conflict coefficient of each grid area is obtained, and the passing area of each AGV is obtained in combination with the priority coefficient. The process of subsequent path planning for each AGV using the passing area and the intelligent control model includes:

[0078] Set an analysis period. When an analysis period is reached, obtain the number of times each grid area is marked as a conflict area in the most recent analysis period, which is recorded as the conflict coefficient. The conflict coefficient can reflect the frequency of different grid areas being passed by different AGVs in the most recent analysis period;

[0079] Take the product of the priority coefficient of a single AGV and the conflict coefficient of a single grid area at the same moment as their passing coefficient. Set a passing threshold. If the passing coefficient is less than or equal to the passing threshold, mark this grid area as the passing area of this AGV. If the passing coefficient is greater than the passing threshold, mark this grid area as the non-passing area of this AGV;

[0080] Adopt the same method to obtain the passing coefficients between each AGV and each grid area at the same moment respectively, and obtain the passing areas of each AGV respectively. Different AGVs are only allowed to perform path planning and path optimization within their corresponding passing areas during their respective analysis periods;

[0081] Taking a single AGV as an example, when it is determined that there is a path conflict for this AGV, obtain its data set and input it into the intelligent control model, and use the intelligent control model to output the corresponding best optimization plan, and control the AGV to continue driving according to its best optimization plan.

[0082] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent management and control system for AGV operation paths for warehousing based on the Internet of Things, characterized in that: Includes the following modules: The map construction module is used to obtain the basic information of the warehouse and construct the corresponding GIS distribution map, and to grid the GIS distribution map to obtain several grid areas; The equipment evaluation module is used to obtain the operation status and operation information of the AGV respectively, and obtain the priority coefficient of the AGV according to the operation information; The path planning module is used to plan the path of the AGV according to the running status to obtain the corresponding initial running path, obtain the grid status of each grid area by combining the running information and the initial running path, and determine whether there is a path conflict between different AGVs; The first optimization module is used to perform path optimization and equipment optimization on each AGV with path conflicts, obtain the corresponding optimal optimization solution, and build a corresponding intelligent management and control model according to different path conflicts and their optimal optimization solutions; The second optimization module is used to obtain the conflict coefficient of each grid area, and obtain the passage area of ​​each AGV in combination with the priority coefficient, and use the passage area and intelligent management and control model to perform subsequent path planning for each AGV.

2. According to the IoT-based intelligent management and control system for AGV operation paths for warehousing in claim 1, it is characterized in that: The process of constructing a GIS distribution map and obtaining several grid areas includes: The basic information includes warehouse boundaries, shelf locations, operating areas, ground conditions, and driving areas. GIS technology is used to construct a GIS distribution map of the warehouse based on the basic information obtained; In the GIS distribution map, all areas where AGVs can travel are regarded as control areas, and the control areas in the GIS distribution map are divided into several grid areas according to the equipment size of the AGV.

3. According to the IoT-based intelligent management and control system for AGV operation paths for warehousing in claim 2, it is characterized in that: The process of obtaining the priority coefficient of AGV based on operation information includes: Real-time monitoring of the operating status of AGVs in the warehouse, including working status and non-working status, and real-time collection of operating information of AGVs in working status in the warehouse, including current location, delivery location, driving speed, battery power, and load weight; is the driving speed Y a 、Battery power Y b , load capacity Y c Set the corresponding weight parameters Q respectively a , Q b , Q c , obtain the priority coefficient R of AGV; 4. According to the IoT-based intelligent management and control system for AGV operation paths for warehousing according to claim 3, it is characterized in that: The process of path planning for AGV to obtain the initial operation path includes: Path planning is performed for all AGVs in working state, the current position and delivery position of each AGV are obtained, and the Dijkstra algorithm is used to build an initial operation path between the current position and delivery position of each AGV.

5. According to the IoT-based intelligent management and control system for AGV operation paths for warehousing according to claim 4, it is characterized in that: The process of obtaining the grid status of each grid area and determining whether there is a path conflict includes: According to the operation information, the grid area where the current position of the corresponding AGV is located is marked as occupied, a prediction cycle is set, and the predicted position of the corresponding AGV in the next several prediction cycles on its current operation path is obtained according to the current position and driving speed; The grid areas where the predicted positions of the same AGV are located are marked as being in a waiting-to-be-occupied state, and the corresponding time when the AGV arrives at each grid area is used as its waiting-to-be-occupied time node. The predicted positions of different AGVs in the subsequent prediction cycles on their current running paths are obtained respectively, and the grid areas corresponding to all the predicted positions are marked; It is judged whether there are different AGVs in a single grid area in the waiting-to-be-occupied state at the same waiting-to-be-occupied time node. If so, it is judged that there is a path conflict, the grid area is marked as a conflict area, and the corresponding different AGVs are included in the same conflicting equipment combination.

6. The intelligent management and control system for AGV operation paths for warehousing based on the Internet of Things according to claim 5 is characterized in that: The process of performing path optimization and equipment optimization on each AGV with path conflicts and obtaining the best optimization solution includes: Obtain the priority coefficients of each AGV in a single conflicting equipment combination, and continue the AGV with the largest priority coefficient along its original running path, and perform path optimization and equipment optimization on other AGVs in the conflicting equipment combination; The path optimization refers to taking the current position of the corresponding AGV as the starting point and using the Dijkstra algorithm to re-construct an optimized operation path between the current position of the AGV and the delivery position; The equipment optimization refers to reducing the driving speed of the AGV so that it reaches the conflict area after the occupied time node and then resumes its original driving speed; According to the driving speed and running path of the same AGV after path optimization and equipment optimization, the corresponding time when the AGV arrives at the delivery location under path optimization and equipment optimization is predicted, and the optimization plan with the shortest time is taken as the best optimization plan.

7. The intelligent management and control system for AGV operation paths for warehousing based on the Internet of Things according to claim 6 is characterized in that: The process of building an intelligent management and control model includes: Obtain the data sets of each AGV under different path conflicts and their best optimization solutions, wherein the data sets include the conflict area, priority coefficient, driving speed, running path corresponding to each AGV and the data sets of other AGVs in the conflicting equipment combination; Generate an intelligent control set based on the data sets of each AGV under different path conflicts and their best optimization solutions, and divide the obtained intelligent control set into a training set and a test set; Construct a convolutional neural network, use different AGV data sets in the training set as input data of the convolutional neural network, use the corresponding best optimization solution in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network; The test set is used to verify the model of the initial convolutional neural network, and the output of the initial convolutional neural network that is less than or equal to the preset test error threshold is used as the corresponding intelligent management and control model.

8. The intelligent management and control system for AGV operation paths for warehousing based on the Internet of Things according to claim 7 is characterized in that: The process of obtaining the passage area of ​​each AGV and performing subsequent path planning includes: Set the analysis cycle. When an analysis cycle is reached, obtain the conflict coefficient of each grid area in the most recent analysis cycle, and take the product of the priority coefficient of a single AGV and the conflict coefficient of a single grid area at the same time as the pass coefficient of the two; Set a pass threshold. If the pass coefficient is less than or equal to the pass threshold, mark the grid area as the pass area of ​​the AGV. Obtain the pass coefficients between each AGV and each grid area at the same time, and obtain the pass area of ​​each AGV. Different AGVs are only allowed to perform path planning and path optimization in the pass area within their corresponding analysis cycle. When it is determined that there is a path conflict in the AGV, its data set is obtained and input into the intelligent management and control model, and the intelligent management and control model is used to output the corresponding optimal optimization plan to control the AGV to continue driving according to its optimal optimization plan.

Images