Unmanned yard intelligent management and control system based on artificial intelligence

Abstract

The invention relates to the technical field of unmanned storage yard intelligent management and control, and discloses an unmanned storage yard intelligent management and control system based on artificial intelligence, which comprises a data acquisition module, a data preprocessing module, a stockpile matching module, a stacking scheme generation module, a stacking scheme acquisition module and an intelligent management and control module, through the arrangement of the data acquisition module and the stockpiling matching module, the data quality and accuracy are improved by acquiring the environmental data and acquiring the fused environmental data, and a foundation is laid for subsequent region division based on the fused environmental data, so that the environment similarity of the divided regions is high, and the efficiency is improved. The divided areas can store the stacked goods meeting the corresponding conditions based on different environment conditions, the safety of the goods in the stacking yard can be guaranteed, the stacking area with the optimal stacking condition is obtained through the stacking matching module for stacking, and the safety and storage of the stacked goods can be guaranteed to the maximum extent.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned storage yard intelligent management and control, and more specifically to an unmanned storage yard intelligent management and control system based on artificial intelligence. Background Art

[0002] Unmanned yards are an important application of automation and intelligent technologies in the field of logistics and warehousing. Unmanned yards use automated equipment to load, unload and stack containers, reducing manual operations and significantly improving operational efficiency and accuracy. Automated equipment can operate 24 hours a day without interruption and is not restricted by working hours, ensuring the efficient operation of the yard around the clock, reducing the demand for manual labor and lowering labor costs.

[0003] However, the existing unmanned yard intelligent management and control system has limited adaptability, and additional adaptation measures are required when facing containers of different sizes, weights and types. The public document with publication number CN118004767A discloses a U-shaped automated container terminal dynamic stacking optimization system. When stacking containers, it first identifies the location, attributes and status information of the stacked containers, then plans the container stacking path, and stacks the identified containers according to the stacking path. Then, stacking optimization is performed according to different optimization goals, and the stacked containers are moved within the minimum range to optimize the stacking of the stacked containers. The data management module is used to process the data generated in the container identification module, the path planning module, the stacking optimization algorithm module and the intelligent control module. The data management module is used to automatically transport and stack the containers, so that the containers are automatically stacked according to the stacking path, which effectively shortens the stacking time of the containers, facilitates the automated stacking management of the containers, and improves the operation efficiency of the yard.

[0004] However, the above-mentioned public documents did not analyze the optimal stacking area when stacking goods of different sizes and types. The stacking of goods requires consideration of the impact of the stacking environment on the goods to ensure the safety of the goods during stacking. At the same time, after path selection, the conventional path selection method is used to obtain the optimal stacking path. It does not consider that there may be other work tasks such as cargo loading while the goods are stacked. Therefore, it is impossible to coordinate and comprehensively and intelligently manage the work tasks in the yard.

[0005] In view of this, the present application proposes an unmanned storage yard intelligent management and control system based on artificial intelligence to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides an unmanned storage yard intelligent management and control system based on artificial intelligence to solve the problems existing in the above-mentioned background technology.

[0007] The present invention provides the following technical solutions: an unmanned yard intelligent management and control system based on artificial intelligence, comprising a data acquisition module, a data preprocessing module, a stockpile matching module, a stacking plan generation module, a stacking plan acquisition module and an intelligent management and control module;

[0008] The data acquisition module is used to collect the data to be matched and the matching data, and transmit them to the data preprocessing module;

[0009] The data preprocessing module is used to receive data from the data acquisition module, perform preprocessing operations on the data, obtain data that can be directly used, and then transmit it to the stockpile matching module;

[0010] The stacking matching module is used to receive the data preprocessed by the data preprocessing module, match the objects to be stacked, obtain the attributes of the objects to be stacked, analyze the matching degree between the objects to be stacked and the stacking area based on the attributes of the objects to be stacked, and obtain the calibrated stacking point;

[0011] The stacking scheme generating module is used to receive data from the stacking material matching module and generate a stacking scheme for a plurality of objects to be stacked;

[0012] The stacking scheme acquisition module is used to acquire the best stacking scheme from a plurality of stacking schemes of objects to be stacked;

[0013] The intelligent management and control module is used to execute the best stacking plan and perform intelligent management and control; if the best stacking plan cannot be obtained, an intelligent warning is performed.

[0014] Preferably, the data to be matched is data of items to be stacked, including the name, packaging type, packaging size, quantity, weight and stacking level of the items to be stacked; the packaging types include container packaging, cage packaging, carton packaging, glass bottle packaging, plastic bottle packaging and mesh bag packaging; if the packaged items are regular in shape, the packaging size is expressed as the length, width, height and floor space of the items; if the packaged items are irregular in shape, the packaging size is expressed as floor space and volume; the stacking levels include ordinary stacking and emergency stacking;

[0015] The matching data is the data of the stacking area, including the stacking area number, environmental data and spatial data; the stacking areas are distinguished according to the environmental data of each area, and digital labels are set for numbering. The environmental data includes the ambient temperature, ambient humidity and sunshine conditions at each time point, and the spatial data includes the stackable area, stackable volume, ground flatness and spatial layout of the stacking area.

[0016] Preferably, the specific manner of distinguishing the stacking areas according to the environmental data of each area is:

[0017] Step S01: Collect environmental data of q points in the stacking area, build a continuous fusion model to fuse the environmental data to obtain fused environmental data;

[0018] Step S02: Perform cluster analysis based on the fused environment data to obtain Q stacking areas.

[0019] Preferably, the specific method of step S01 is:

[0020] Collect environmental data from q points in the stacking area, divide the environmental data into several data sets, and divide the data sets into training part and output part. For each data set in the training part, use the regression analysis method to build a single model; for each single model, calculate the model weight Among them, ω m is the weight of the single model, k is the adjustment parameter of the model weight, and its value range is [0, 1], n m is the number of environmental data in the data set corresponding to the single model m; N is the number of data of all environmental data; d m is the error evaluation index of the single model m, c and a are positive real number parameters;

[0021] Based on the calculated model weights, the monomer models are fused to obtain a continuous fusion model, and the output part is input into the continuous fusion model, and the output obtained is the continuous fusion data feature LR;

[0022] LR=∑ m ω m ×SC m SC m It is the output value of the single model for the input sample, and the input sample is the data of the output part;

[0023] Based on the constructed continuous fusion model, the fused environmental data of q points in the stacking area are obtained.

[0024] Preferably, the specific method of step S02 is:

[0025] Collect environmental data of q points in the stacking area, set digital labels for each collection point, and construct the environmental similarity matrix U, U = comm(a, b), where comm(a, b) represents the environmental similarity between the area with digital label a and the area with digital label b; the formula is expressed as:

[0026]

[0027] Among them, k 1is the weight parameter, za is the fusion environment data feature vector of the area with digital label a, zb is the fusion environment data feature vector of the area with digital label b; cov(za,zb) is the covariance of za and zb, σ(za) is the standard deviation of za, σ(zb) is the standard deviation of zb; a=1, 2, 3, …, q, b=1, 2, 3, …, q;

[0028] Based on the environmental similarity matrix, the collection areas are clustered.

[0029] Preferably, the specific method of clustering the collection areas is:

[0030] Each region is regarded as a separate small cluster, with a total of q small clusters. The gravity value between every two small clusters is calculated based on the environmental similarity matrix. A merging threshold d is preset to control cluster merging. The small clusters are iteratively merged. In each iteration, the two small clusters L1 and L2 with the largest gravity value are found.

[0031] If the gravity value between L1 and L2 is greater than the merging threshold d, L1 and L2 are merged into a new cluster L′, and the gravity value between the new cluster L′ and all other small clusters is recalculated; otherwise, no iterative merging is performed; the execution is repeated until the gravity value between no two small clusters is greater than the merging threshold d, at which time Q new clusters are obtained, that is, Q stacking areas.

[0032] Preferably, the specific method of obtaining the attributes of the items to be stacked is:

[0033] Construct a knowledge graph for stacking items, input the name of the item to be stacked, match the name of the item to be stacked with the name of the stacking item in the knowledge graph, and obtain the entity attributes of the successfully matched stacking item as the attributes of the item to be stacked; the matching method is string matching, including exact matching and approximate matching. Exact matching is to compare the string corresponding to the item to be stacked with the string corresponding to the entity in the knowledge graph of stacking items, and if they are exactly the same, then match; approximate matching is to use a string similarity algorithm to calculate the similarity between the string corresponding to the item to be stacked and the string corresponding to the entity in the knowledge graph of stacking items, and match the entity in the knowledge graph of stacking items with the highest similarity with the item to be stacked; a combination of the two is used for matching, and a successfully matched entity is obtained as a successfully matched stacking item, and the entity attributes corresponding to the successfully matched entity are used as the attributes of the item to be stacked;

[0034] The specific method of constructing the knowledge graph of stacked items is:

[0035] Collecting historical stacked item information and stackable item information, wherein the stackable item information is information about items that can be stacked and can be obtained using web crawler technology, and the historical stacked item information is information about items that have been stacked historically, and the item information includes optimal stacking temperature, humidity, and light conditions for the items;

[0036] The collected historical stacked items and stackable items are numbered, and a knowledge graph of stacked items is constructed. The knowledge graph of stacked items includes entities, entity attribute information, entity relationship information, and entity types. The entities are the names of the collected historical stacked items and stackable items; the entity relationship information is the relationship between entities, that is, the relationship between each stacked item; the entity attribute information defines the attributes of the entity, that is, the optimal stacking temperature, humidity, and lighting conditions of the entity; the entity type defines the item category to which the entity belongs.

[0037] Preferably, the specific method of obtaining the calibration stacking point is:

[0038] The acquired attributes of the items to be stacked are combined with the matching data and matched with the matching data. The matching method is semantic matching, and the semantic matching uses a word embedding method. The word embedding method converts the attributes of the items to be stacked and the environmental data of each stacking area into a vector form through a word embedding model, and matches them by calculating the similarity between the vectors. The vectors are arranged from large to small according to the similarity, and the area with the highest similarity is used as the calibration stacking point.

[0039] Preferably, the stacking scheme includes a storage route and a corresponding storage scheme, wherein the storage scheme is the storage quantity of each stacking area, and the storage route is the route from the unloading point of the items to be stacked to the stacking area;

[0040] The method for obtaining the storage route in the stacking scheme of the objects to be stacked is as follows:

[0041] Get all u routes from each loading and unloading point to each stacking area, mark the repeated sections in the u routes, get the repeated sections and the number of repetitions, record the sections with more than half of the total number of routes as hot sections, and the remaining repeated sections as ordinary sections. The routes without marked sections in the u routes are recorded as idle routes;

[0042] Get all the routes from the unloading point of the items to be stacked to the stacking area, and determine the stacking level of the items to be stacked. If the items to be stacked are for emergency stacking, the route marked with hot sections will be used as the storage route. If the items to be stacked are for ordinary stacking, a secondary judgment will be made. If there are other loading tasks in the yard at the same time, the idle route will be selected as the storage route. If there are no other loading tasks in the yard, all u routes will be used as storage routes; if there are multiple stacking areas, the storage routes will be obtained in sequence according to the stacking area order; if the idle route is zero, the route marked with ordinary sections will replace the idle route.

[0043] Preferably, the storage solution is obtained in the following manner:

[0044] Obtain the remaining stackable volume and area of ​​the calibrated stacking point, and obtain the volume and area occupied by the items to be stacked based on the data of the items to be stacked. If the calibrated stacking point cannot stack all the items to be stacked, the unstackable parts will be stacked in the area with the second highest similarity. The storage plan can be expressed as: CF = {w1:S1, w2:S2...wQ:SQ}, where CF is the storage plan, w1 is the area with the highest similarity, that is, the calibrated stacking point, S1 is the storage quantity corresponding to the area, that is, the stacking quantity, w2 is the area with the second highest similarity, S2 is the storage quantity corresponding to the area, wQ is the area with the Qth similarity, that is, the last area in the arrangement, and SQ is the storage quantity corresponding to the area;

[0045] The specific method of obtaining the optimal stacking solution is:

[0046] Obtain a storage route, and use an optimization algorithm to obtain the best storage route. The stacking plan formed by combining the best storage route with the storage plan is the best stacking plan.

[0047] Technical effects and advantages of the present invention:

[0048] (1) The present invention is provided with a data acquisition module and a stacking material matching module, which is conducive to acquiring fused environmental data by acquiring environmental data, thereby improving data quality and accuracy, and laying a foundation for subsequent regional division based on fused environmental data. Therefore, the environmental similarity of the divided areas is high, and the divided areas can store stacked items that meet corresponding conditions based on different environmental conditions, which can ensure the safety of goods when they are stacked in the yard. The stacking area with the best stacking conditions is obtained by the stacking material matching module for stacking, which can maximize the safety and storage of the stacked goods.

[0049] (2) The present invention is provided with a stacking plan generation module, which is conducive to marking the route as hot sections, ordinary sections and idle sections, so as to better coordinate various transportation tasks in the yard and make better use of route resources. When there is a loading task in the yard, the loading task is given priority, and the stacking level of the items to be stacked is determined. If the items are emergency stacked, the requirements are met first. If the items are ordinary stacked, the loading task route is not occupied. The hot section is the section that multiple routes need to pass through, that is, the central point section. Many transportation tasks need to pass through here. Therefore, avoiding the hot section for ordinary stacked items can improve the efficiency of the yard transportation task, make full use of route resources, and respond to various emergencies and transportation tasks at any time, thereby enhancing the intelligent management capability of the unmanned yard. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a structural diagram of the unmanned storage yard intelligent management and control system based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0051] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. In addition, the forms of the various structures recorded in the following embodiments are merely illustrative. The unmanned storage yard intelligent management and control system based on artificial intelligence involved in the present invention is not limited to the various structures recorded in the following embodiments. All other implementations obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0052] like Figure 1 As shown, the present invention provides an unmanned yard intelligent management and control system based on artificial intelligence, including a data acquisition module, a data preprocessing module, a stockpile matching module, a stacking plan generation module, a stacking plan acquisition module and an intelligent management and control module;

[0053] The data acquisition module is used to collect the data to be matched and the matching data, and transmit them to the data preprocessing module; the data to be matched is the data of the items to be stacked, including but not limited to the name, packaging type, packaging size, quantity, weight and stacking level of the items to be stacked; the packaging type includes packaging forms such as container packaging, cargo cage packaging, carton packaging, glass bottle packaging, plastic bottle packaging and mesh bag packaging. If the packaged items are regular in shape, the packaging size is expressed as the length, width, height and floor space of the items; if the packaged items are irregular in shape, the packaging size is expressed as floor space and volume; the stacking level includes ordinary stacking and emergency stacking. "Put" means that the items to be stacked can be arranged according to the normal stacking schedule, and "emergency stacking" means that the items to be stacked need to be stacked as quickly as possible. The stacking level can be calibrated by technicians in this field according to the actual situation of the items to be stacked; the matching data is the data of the stacking area, including but not limited to the stacking area number, environmental data and spatial data; the stacking areas are distinguished according to the environmental data of each area, and digital labels are set for numbering. The environmental data includes but is not limited to the ambient temperature, ambient humidity and sunshine conditions at each time point, and the spatial data includes but is not limited to the stackable area, stackable volume, ground flatness and spatial layout of the stacking area;

[0054] The data preprocessing module is used to receive data from the data acquisition module, perform preprocessing operations on the data, and transmit the obtained data that can be directly used to the stockpile matching module; the preprocessing operation includes cleaning the data and standardizing the data; the cleaning data includes but is not limited to deleting duplicate, erroneous or incomplete data to ensure the data quality in the subsequent analysis process, and the standardization process is to unify the data of different source formats into a standard format to facilitate the use of the data in the subsequent analysis process;

[0055] The stacking matching module is used to receive the data preprocessed by the data preprocessing module, match the objects to be stacked, obtain the attributes of the objects to be stacked, analyze the matching degree between the objects to be stacked and the stacking area based on the attributes of the objects to be stacked, and obtain the calibrated stacking point;

[0056] The stacking scheme generating module is used to receive data from the stacking matching module and generate stacking schemes for multiple items to be stacked, wherein the stacking schemes include storage routes and corresponding storage schemes, wherein the storage routes are the storage quantities of each stacking area, and the storage routes are the routes from the unloading point of the items to be stacked to the stacking area;

[0057] The stacking scheme acquisition module is used to acquire the best stacking scheme from a plurality of stacking schemes of the objects to be stacked; the best stacking scheme is acquired by using any one of the optimization algorithms;

[0058] The intelligent management and control module is used to execute the best stacking plan and perform intelligent management and control; if the best stacking plan cannot be obtained, an intelligent warning is performed.

[0059] In this embodiment, it should be specifically explained that the specific manner in which the stacking areas are distinguished according to the environmental data of each area is:

[0060] Step S01: Collect environmental data of q points in the stacking area, build a continuous fusion model to fuse the environmental data to obtain fused environmental data;

[0061] Step S02: Perform cluster analysis based on the fused environment data to obtain Q stacking areas.

[0062] In this embodiment, it should be specifically explained that the specific method of step S01 is:

[0063] Collect environmental data from q points in the stacking area, divide the environmental data into several data sets, and divide the data sets into training part and output part. For each data set in the training part, use regression analysis method to build a single model to describe the pattern of data in the data set; for each single model, calculate the model weight Among them, ω m is the weight of the single model, k is the adjustment parameter of the model weight, with a value range of [0, 1], which controls the relative importance of the two weight items; n m is the number of environmental data in the data set corresponding to the single model m; N is the number of data of all environmental data; d m is the error evaluation index of the single model m, such as root mean square error, R square, etc.; c and a are positive real number parameters used to control the shape of the function;

[0064] Based on the calculated model weights, the monomer models are fused to obtain a continuous fusion model, and the output part is input into the continuous fusion model, and the output obtained is the continuous fusion data feature LR;

[0065] LR=∑ m ω m ×SC m SC m It is the output value of the single model for the input sample, and the input sample is the data of the output part;

[0066] Based on the constructed continuous fusion model, the fused environmental data of q points in the stacking area are obtained, so that the environmental data can be better analyzed. Fusion of data can improve data quality and accuracy, eliminate duplicate, contradictory or inconsistent information, and the fused data can provide more comprehensive and multi-dimensional information.

[0067] In this embodiment, it should be specifically explained that the specific method of step S02 is:

[0068] Collect environmental data of q points in the stacking area, set digital labels for each collection point, and construct the environmental similarity matrix U, U = comm(a, b), where comm(a, b) represents the environmental similarity between the area with digital label a and the area with digital label b; the formula is expressed as:

[0069]

[0070] Among them, k 1 is the weight parameter, za is the fusion environment data feature vector of the area with digital label a, zb is the fusion environment data feature vector of the area with digital label b; cov(za,zb) is the covariance of za and zb, σ(za) is the standard deviation of za, σ(zb) is the standard deviation of zb; a=1, 2, 3, …, q, b=1, 2, 3, …, q;

[0071] Based on the environmental similarity matrix, the collection areas are clustered; the specific method is:

[0072] Each region is regarded as a separate small cluster, with a total of q small clusters. The gravity value between every two small clusters is calculated based on the environmental similarity matrix. A merging threshold d is preset to control cluster merging. The small clusters are iteratively merged. In each iteration, the two small clusters L1 and L2 with the largest gravity value are found.

[0073] If the gravity value between L1 and L2 is greater than the merging threshold d, L1 and L2 are merged into a new cluster L′, and the gravity value between the new cluster L′ and all other small clusters is recalculated; otherwise, no iterative merging is performed; the execution is repeated until the gravity value between no two small clusters is greater than the merging threshold d, at which time Q new clusters are obtained, that is, Q stacking areas; the fused environmental data of the Q stacking areas have high similarity, that is, the environmental conditions of the Q stacking areas are highly similar, and the environmental temperature, humidity and sunshine conditions at each time point are highly similar, so they can be regarded as the same area category, and the stacking materials that meet the same environmental material requirements are stacked.

[0074] In this embodiment, it should be specifically explained that the calculation formula of the gravity value is expressed as: Where Y(L1,L2) is the gravitational force between L1 and L2, G is the gravitational constant, c1 is the number of fused environment data in small cluster L1, c2 is the number of fused environment data in small cluster L2, Dis(L1,L2) is the distance between small clusters L1 and L2; that is, the average value of the distance between all data pairs in the two small clusters, and the distance measurement method between data pairs can be calculated using the Euclidean distance formula; ω is a hyperparameter used to control the overall size; σ(L1) is the standard deviation of the fused environment data feature vector in small cluster L1, σ(L2) is the standard deviation of the fused environment data feature vector in small cluster L2, reflecting the degree of dispersion within the small cluster; μ(L1) is the mean value of the fused environment data feature vector in small cluster L1, μ(L2) is the mean value of the fused environment data feature vector in small cluster L2, reflecting the overall level of the small cluster; comm * (L1, L2) is the mean similarity between the regions of the small clusters L1 and L2; the data pairs in the small clusters are region pairs formed by the regions in the small clusters; for example, if there is a small cluster with region 1 and region 2, and another small cluster with region 3 and region 4, then region 1 and region 3 are a region pair, and region 1 and region 4 are a region pair;

[0075] By adjusting the size of the merge threshold d, the strictness of the merge can be controlled, thereby affecting the final number of clusters. The size of the merge threshold d needs to be determined based on the geographical location and area of ​​the actual stacking area to determine a reasonable merge threshold so that the ideal number of clusters Q is finally obtained.

[0076] In this embodiment, it should be specifically explained that the specific method of obtaining the attributes of the items to be stacked is:

[0077] Construct a knowledge graph for stacking items, input the name of the item to be stacked, match the name of the item to be stacked with the name of the stacking item in the knowledge graph, and obtain the entity attributes of the successfully matched stacking item as the attributes of the item to be stacked; the matching method is string matching, including exact matching and approximate matching. Exact matching is to compare the string corresponding to the item to be stacked with the string corresponding to the entity in the knowledge graph of stacking items, and if they are exactly the same, then match; approximate matching is to use a string similarity algorithm to calculate the similarity between the string corresponding to the item to be stacked and the string corresponding to the entity in the knowledge graph of stacking items, and match the entity in the knowledge graph of stacking items with the highest similarity with the item to be stacked; a combination of the two is used for matching, and a successfully matched entity is obtained as a successfully matched stacking item, and the entity attributes corresponding to the successfully matched entity are used as the attributes of the item to be stacked;

[0078] The specific method of constructing the knowledge graph of stacked items is:

[0079] Collecting historical stacked item information and stackable item information, wherein the stackable item information is information about items that can be stacked and can be obtained using web crawler technology, and the historical stacked item information is information about items that have been stacked in the past, and the item information includes optimal stacking temperature, humidity, and light conditions for the items, wherein the light conditions specifically include information describing light storage conditions such as avoiding direct sunlight and storing in a dark place;

[0080] The collected historical stacked items and stackable items are numbered, and a knowledge graph of stacked items is constructed. The knowledge graph of stacked items includes entities, entity attribute information, entity relationship information, and entity types. Entities are the names of the collected historical stacked items and stackable items; entity relationship information is the relationship between entities, that is, the relationship between each stacked item; entity attribute information defines the attributes of the entity, that is, the optimal stacking temperature, humidity, and lighting conditions of the entity; entity type defines the category of item to which the entity belongs, such as food, clothing, daily necessities, etc.

[0081] In this embodiment, it should be specifically explained that the specific method of obtaining the calibrated stacking point is:

[0082] The acquired attributes of the items to be stacked are combined with the matching data and matched with the matching data. The matching method is semantic matching, and the semantic matching uses a word embedding method. The word embedding method converts the attributes of the items to be stacked and the environmental data of each stacking area into a vector form through a word embedding model, and matches them by calculating the similarity between the vectors. The vectors are arranged from large to small according to the similarity, and the area with the highest similarity is used as the calibrated stacking point. The word embedding method is a prior art and will not be described in detail in this embodiment.

[0083] In this embodiment, it should be specifically explained that the method for obtaining the storage route in the stacking scheme of the objects to be stacked is:

[0084] Get all u routes from each loading and unloading point to each stacking area, mark the repeated sections in the u routes, get the repeated sections and the number of repetitions, record the sections with more than half of the total number of routes as hot sections, and the remaining repeated sections as ordinary sections. The routes without marked sections in the u routes are recorded as idle routes;

[0085] Get all routes from the unloading point of the items to be stacked to the stacking area, and determine the stacking level of the items to be stacked. If the items to be stacked are for emergency stacking, the route marked with hotspot sections will be used as the storage route. If the items to be stacked are for ordinary stacking, a secondary determination will be made. If there are other loading tasks in the yard at the same time, the idle route will be selected as the storage route. If there are no other loading tasks in the yard, all u routes will be used as storage routes. If there are multiple stacking areas, the storage routes will be obtained in sequence according to the stacking area order. If the idle route is zero, the route marked with ordinary sections will replace the idle route.

[0086] Marking the routes as hot sections, ordinary sections and idle sections can better coordinate various transport tasks in the yard and make better use of route resources. When there are loading tasks in the yard, the loading tasks will be given priority, and the stacking levels of the stacked items will be determined. If they are emergency stacked items, the requirements will be met first. If they are ordinary stacked items, the loading task route will not be occupied. Hot sections are sections that multiple routes need to pass through, that is, the central point sections. Many transport tasks need to pass through here. Therefore, avoiding hot sections for ordinary stacked items can improve the efficiency of yard transport tasks, make full use of route resources, and be able to respond to various emergencies and transport tasks at any time.

[0087] In this embodiment, it should be specifically explained that the storage solution is obtained in the following manner:

[0088] Obtain the remaining stackable volume and area of ​​the calibrated stacking point, and obtain the volume and area occupied by the items to be stacked based on the data of the items to be stacked. If the calibrated stacking point cannot stack all the items to be stacked, the unstackable part will be stacked in the area with the second highest similarity. The storage plan can be expressed as: CF = {w1: S1, w2: S2...wQ: SQ}, where CF is the storage plan, w1 is the area with the highest similarity, that is, the calibrated stacking point, and S1 is the storage quantity corresponding to the area, that is, the stacking quantity. w2 is the second area in similarity ranking, S2 is the storage quantity corresponding to this area, wQ is the Qth area in similarity ranking, that is, the last area in ranking, SQ is the storage quantity corresponding to this area, and so on. If all storage is completed by w2, the storage quantities in the remaining areas are all zero. The stacking quantity unit can be set according to the packaging type and quantity of the items to be stacked. If the number of items to be stacked is 1, the calibrated stacking point cannot meet the stacking requirements, and w2 meets the stacking requirements, then S1 is zero and S2 is 1.

[0089] In this embodiment, it should be specifically explained that the specific method of obtaining the optimal stacking solution is:

[0090] Obtain a storage route, and use an optimization algorithm to obtain the best storage route. The stacking scheme formed by combining the best storage route with the storage scheme is the best stacking scheme. The optimization algorithm can be any one of the algorithms such as genetic algorithm, particle swarm algorithm, ant colony algorithm, etc. The fitness function or objective function used can be specifically set according to the actual situation, and the shortest time, the least energy consumption or the most straight sections can be selected. The route does not conflict with other routes, that is, it does not collide with other task processes during the storage process. If there are multiple stacking areas, the best storage route is obtained for the corresponding storage routes in the order of the stacking areas;

[0091] If the items to be stacked are for emergency stacking, and there are other tasks on the storage route, and there is no storage route that meets the conditions, the best storage route will be selected from the routes marked with ordinary sections;

[0092] If the best storage route cannot be obtained, an intelligent warning will be issued, and the fitness function or objective function needs to be adjusted accordingly.

[0093] In this embodiment, it should be specifically explained that the difference between this embodiment and the prior art lies mainly in that this embodiment has a data acquisition module and a stacking material matching module, which is conducive to acquiring fused environmental data by acquiring environmental data, thereby improving data quality and accuracy, and laying a foundation for subsequent regional division based on fused environmental data. Therefore, the environmental similarity of the divided areas is high, and the divided areas can store stacked items that meet the corresponding conditions based on different environmental conditions, which can ensure the safety of goods when they are stacked in the yard. The stacking area with the best stacking conditions is obtained through the stacking material matching module for stacking, which can maximize the safety and storage of the stacked goods. It is equipped with a stacking plan generation module, which is conducive to The marking of hot sections, ordinary sections and idle sections on the route can better coordinate various transportation tasks in the yard and make better use of route resources. When there is a loading task in the yard, the loading task will be given priority, and the stacking level of the stacked items will be determined. If it is an emergency stacking item, the requirements will be met first. If it is an ordinary stacking item, the loading task route will not be occupied; the hot section is the section that multiple routes need to pass through, that is, the central point section, and many transportation tasks need to pass through here. Therefore, avoiding the hot section for ordinary stacked items can improve the efficiency of the yard transportation task, make full use of route resources, and respond to various emergencies and transportation tasks at any time, enhancing the intelligent management capabilities of the unmanned yard.

[0094] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0095] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An unmanned storage yard intelligent management and control system based on artificial intelligence, characterized by: It includes data acquisition module, data preprocessing module, stacking material matching module, stacking plan generation module, stacking plan acquisition module and intelligent management and control module; The data acquisition module is used to collect the data to be matched and the matching data, and transmit them to the data preprocessing module; The data preprocessing module is used to receive data from the data acquisition module, perform preprocessing operations on the data, obtain data that can be directly used, and then transmit it to the stockpile matching module; The stacking matching module is used to receive the data preprocessed by the data preprocessing module, match the objects to be stacked, obtain the attributes of the objects to be stacked, analyze the matching degree between the objects to be stacked and the stacking area based on the attributes of the objects to be stacked, and obtain the calibrated stacking point; The stacking scheme generating module is used to receive data from the stacking material matching module and generate a stacking scheme for a plurality of objects to be stacked; The stacking scheme acquisition module is used to acquire the best stacking scheme from a plurality of stacking schemes of objects to be stacked; The intelligent management and control module is used to execute the best stacking plan and perform intelligent management and control; if the best stacking plan cannot be obtained, an intelligent warning is performed.

2. According to claim 1, an unmanned storage yard intelligent management and control system based on artificial intelligence is characterized by: The data to be matched is data of the items to be stacked, including the name, packaging type, packaging size, quantity, weight and stacking level of the items to be stacked; the packaging types include container packaging, cage packaging, carton packaging, glass bottle packaging, plastic bottle packaging and mesh bag packaging; if the packaged items are regular in shape, the packaging size is expressed as the length, width, height and floor space of the items; if the packaged items are irregular in shape, the packaging size is expressed as floor space and volume; the stacking levels include ordinary stacking and emergency stacking; The matching data is the data of the stacking area, including the stacking area number, environmental data and spatial data; the stacking areas are distinguished according to the environmental data of each area, and digital labels are set for numbering. The environmental data includes the ambient temperature, ambient humidity and sunshine conditions at each time point, and the spatial data includes the stackable area, stackable volume, ground flatness and spatial layout of the stacking area.

3. According to claim 2, an unmanned storage yard intelligent management and control system based on artificial intelligence is characterized by: The specific method of distinguishing the stacking areas according to the environmental data of each area is as follows: Step S01: Collect environmental data of q points in the stacking area, build a continuous fusion model to fuse the environmental data to obtain fused environmental data; Step S02: Perform cluster analysis based on the fused environment data to obtain Q stacking areas.

4. The unmanned storage yard intelligent management and control system based on artificial intelligence according to claim 3 is characterized by: The specific method of step S01 is: Collect environmental data of q points in the stacking area, divide the environmental data into several data sets, and divide the data sets into training part and output part. For each data set in the training part, use regression analysis method to build a single model; For each individual model, calculate the model weight Among them, ω m is the weight of the single model, k is the adjustment parameter of the model weight, and its value range is [0, 1], n m is the number of environmental data in the data set corresponding to the single model m; N is the number of data of all environmental data; d m is the error evaluation index of the single model m, c and a are positive real number parameters; Based on the calculated model weights, the monomer models are fused to obtain a continuous fusion model, and the output part is input into the continuous fusion model, and the output obtained is the continuous fusion data feature LR; LR=∑ m ω m ×SC m SC m It is the output value of the single model for the input sample, and the input sample is the data of the output part; Based on the constructed continuous fusion model, the fused environmental data of q points in the stacking area are obtained.

5. The unmanned storage yard intelligent management and control system based on artificial intelligence according to claim 4 is characterized by: The specific method of step S02 is: Collect environmental data of q points in the stacking area, set digital labels for each collection point, and construct the environmental similarity matrix U, U = comm(a, b), where comm(a, b) represents the environmental similarity between the area with digital label a and the area with digital label b; the formula is expressed as: Wherein, k1 is the weight parameter, za is the fusion environment data feature vector of the area with digital label a, zb is the fusion environment data feature vector of the area with digital label b; cov(za,zb) is the covariance of za and zb, σ(za) is the standard deviation of za, σ(zb) is the standard deviation of zb; a=1, 2, 3, …, q, b=1, 2, 3, …, q; Based on the environmental similarity matrix, the collection areas are clustered.

6. The unmanned storage yard intelligent management and control system based on artificial intelligence according to claim 5 is characterized by: The specific method of clustering the acquisition areas is as follows: Each area is regarded as a separate small cluster, with a total of q small clusters, and the attraction value between every two small clusters is calculated based on the environmental similarity matrix; The preset merging threshold d is used to control the merging of clusters. The small clusters are merged iteratively. In each iteration, two small clusters L1 and L2 with the largest gravity values ​​are found. If the gravity value between L1 and L2 is greater than the merging threshold d, L1 and L2 are merged into a new cluster L′, and the gravity value between the new cluster L′ and all other small clusters is recalculated; otherwise, no iterative merging is performed; the execution is repeated until the gravity value between no two small clusters is greater than the merging threshold d, at which time Q new clusters are obtained, that is, Q stacking areas.

7. The unmanned storage yard intelligent management and control system based on artificial intelligence according to claim 6 is characterized by: The specific method of obtaining the attributes of the items to be stacked is: Construct a knowledge graph for stacking items, input the name of the item to be stacked, match the name of the item to be stacked with the name of the stacking item in the knowledge graph, and obtain the entity attributes of the successfully matched stacking item as the attributes of the item to be stacked; the matching method is string matching, including exact matching and approximate matching. Exact matching is to compare the string corresponding to the item to be stacked with the string corresponding to the entity in the knowledge graph of stacking items, and if they are exactly the same, then match; approximate matching is to use a string similarity algorithm to calculate the similarity between the string corresponding to the item to be stacked and the string corresponding to the entity in the knowledge graph of stacking items, and match the entity in the knowledge graph of stacking items with the highest similarity with the item to be stacked; a combination of the two is used for matching, and a successfully matched entity is obtained as a successfully matched stacking item, and the entity attributes corresponding to the successfully matched entity are used as the attributes of the item to be stacked; The specific method of constructing the knowledge graph of stacked items is: Collecting historical stacked item information and stackable item information, wherein the stackable item information is information about items that can be stacked and can be obtained using web crawler technology, and the historical stacked item information is information about items that have been stacked historically, and the item information includes optimal stacking temperature, humidity, and light conditions for the items; The collected historical stacked items and stackable items are numbered, and a knowledge graph of stacked items is constructed. The knowledge graph of stacked items includes entities, entity attribute information, entity relationship information, and entity types. The entities are the names of the collected historical stacked items and stackable items; the entity relationship information is the relationship between entities, that is, the relationship between each stacked item; the entity attribute information defines the attributes of the entity, that is, the optimal stacking temperature, humidity, and lighting conditions of the entity; the entity type defines the item category to which the entity belongs.

8. The unmanned storage yard intelligent management and control system based on artificial intelligence according to claim 7 is characterized by: The specific method of obtaining the calibration stacking point is: The acquired attributes of the items to be stacked are combined with the matching data and matched with the matching data. The matching method is semantic matching, and the semantic matching uses a word embedding method. The word embedding method converts the attributes of the items to be stacked and the environmental data of each stacking area into a vector form through a word embedding model, and matches them by calculating the similarity between the vectors. The vectors are arranged from large to small according to the similarity, and the area with the highest similarity is used as the calibration stacking point.

9. The unmanned storage yard intelligent management and control system based on artificial intelligence according to claim 8 is characterized by: The stacking scheme includes a storage route and a corresponding storage scheme, wherein the storage scheme is the storage quantity of each stacking area, and the storage route is the route from the unloading point of the items to be stacked to the stacking area; The method for obtaining the storage route in the stacking scheme of the objects to be stacked is as follows: Get all u routes from each loading and unloading point to each stacking area, mark the repeated sections in the u routes, get the repeated sections and the number of repetitions, record the sections with more than half of the total number of routes as hot sections, and the remaining repeated sections as ordinary sections. The routes without marked sections in the u routes are recorded as idle routes; Get all routes from the unloading point of the items to be stacked to the stacking area, and determine the stacking level of the items to be stacked. If the items to be stacked are for emergency stacking, the route marked with hotspot sections will be used as the storage route. If the items to be stacked are for ordinary stacking, a secondary determination will be made. If there are other loading tasks in the yard at the same time, the idle route will be selected as the storage route. If there are no other loading tasks in the yard, all u routes will be used as storage routes. If there are multiple stacking areas, the storage routes are obtained in sequence according to the stacking area order; if the idle route is zero, the route marked with ordinary sections will replace the idle route.

10. The unmanned storage yard intelligent management and control system based on artificial intelligence according to claim 9 is characterized in that: The storage solution is obtained as follows: Obtain the remaining stackable volume and area of ​​the calibrated stacking point, and obtain the volume and area occupied by the items to be stacked based on the data of the items to be stacked. If the calibrated stacking point cannot stack all the items to be stacked, the unstackable parts will be stacked in the area with the second highest similarity. The storage plan can be expressed as: CF = {w1:S1, w2:S2...wQ:SQ}, where CF is the storage plan, w1 is the area with the highest similarity, that is, the calibrated stacking point, S1 is the storage quantity corresponding to the area, that is, the stacking quantity, w2 is the area with the second highest similarity, S2 is the storage quantity corresponding to the area, wQ is the area with the Qth similarity, that is, the last area in the arrangement, and SQ is the storage quantity corresponding to the area; The specific method of obtaining the optimal stacking solution is: Obtain a storage route, and use an optimization algorithm to obtain the best storage route. The stacking plan formed by combining the best storage route with the storage plan is the best stacking plan.

Images