Storage cabinet batch delivery and pickup system and batch delivery and pickup method thereof

Abstract

The invention discloses a storage cabinet batch delivery and pickup system and a batch delivery and pickup method thereof, the storage cabinet batch delivery and pickup system comprises a user side device, a central server and an intelligent storage cabinet group, the user side device communicates with the central server through a network, and the central server communicates with the intelligent storage cabinet group through the network. The invention discloses a delivery and pickup method of a locker batch delivery and pickup system. The method comprises the following steps: S1, initiating a task and uploading an order; s2, order processing and path planning; s3, article identification and identity verification; s4, putting articles and updating the progress; and S5, consumable monitoring and supply. According to the invention, order grouping and path planning can be automatically completed, the walking distance and operation time of a courier and a user are effectively reduced, the multi-mode identification technology sub-module and the batch identity verification sub-module can rapidly process article information and identity verification, and the overall efficiency of batch delivery and pickup is significantly improved.

Description

Technical Field

[0001] The present invention relates to the field of logistics delivery and retrieval technology, and in particular to a locker batch delivery and retrieval system and a batch delivery and retrieval method thereof. Background Art

[0002] With the rapid development of the e-commerce industry, logistics business volume has seen explosive growth, and the bulk delivery and collection of packages has become a top priority in the logistics process. In the field of logistics delivery and retrieval, Chinese patent publication number "CN105006075A" discloses a locker bulk delivery and retrieval system and bulk delivery and retrieval method. The bulk delivery and retrieval system includes lockers, smart terminals, and servers. The smart terminals communicate with the servers via a network, and the servers communicate with the lockers via a network. The servers include a server database, a server identity verification module, a server reservation module, and a server compartment information module. The lockers include a human-computer interaction device, a control unit, and multiple compartments.

[0003] Although the above patent solves the problem that traditional delivery and pickup systems cannot deliver and pick up parcels in batches and cannot predict whether there are empty compartments in the lockers, it does not design a special module to plan the shortest time and optimal path for couriers to deliver and users to pick up parcels. In addition, the above patent does not design a specific algorithm in the system to support the overall framework of the system. Therefore, it has certain difficulties in actual implementation. At the same time, it is impossible to obtain big data on batch delivery and pickup to continuously optimize and improve the delivery and pickup system. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a locker batch delivery system and a batch delivery method.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A locker batch delivery and retrieval system includes a user-end device, a central server, and a smart locker group. The user-end device communicates with the central server via a network, and the central server communicates with the smart locker group via a network. The central server includes a batch task intelligent planning module, a virtual locker mapping module, a rapid identification and verification module, an automatic replenishment and recycling module, and a data analysis and optimization module.

[0007] The batch task intelligent planning module consists of an order analysis and grouping submodule, an optimal route planning submodule, and a time estimation and scheduling submodule, which are used to analyze and group orders, plan delivery and pickup routes, estimate delivery and pickup times, and reasonably schedule container resources.

[0008] The virtual locker mapping module consists of a 3D visualization display submodule, a remote operation simulation submodule, and a real-time synchronization update submodule. It is used to create a virtual map of the locker, facilitate users to simulate operations in the virtual map, and allow users to keep abreast of the latest locker status at any time.

[0009] The rapid identification and verification module consists of a multimodal identification technology submodule, a batch identity verification submodule, and an abnormal information warning submodule. It is used to quickly and accurately identify batch item information, support batch user identity verification, and issue warnings in the event of abnormal item information or identity verification failure.

[0010] The automatic replenishment and recycling module consists of an inventory detection and warning submodule, an automatic replenishment mechanism submodule, and an item recycling processing submodule, which are used to notify management personnel to replenish stocks, automatically replenish consumables, and automatically mark and recycle damaged items;

[0011] The data analysis and optimization module consists of a data collection and organization submodule, a behavior analysis and prediction submodule, and a system optimization suggestion submodule. It is used to collect and organize various data on batch delivery and pickup, predict the batch delivery and pickup needs of different user groups in different time periods, and provide suggestions for system improvement and optimization.

[0012] Preferably, the order parsing and grouping submodule is connected to the user terminal device to obtain batch delivery and pickup order information, and uses the K-Means clustering algorithm to group the orders according to factors such as item size, weight, fragility and destination address; the optimal path planning submodule establishes a three-dimensional map model of the locker group, uses the A* algorithm, and plans the optimal delivery and pickup path for the courier or user based on the order grouping information and the locker usage status, and optimizes it by considering factors such as actual walking obstacles.

[0013] Preferably, the time estimation and scheduling submodule collects and analyzes the operation time data of previous batch delivery and pickup tasks, establishes a time prediction model, estimates the time required for this task, and reasonably schedules the box resources according to the time and locker usage status.

[0014] Preferably, the three-dimensional visualization display submodule uses 3D modeling software to create an accurate three-dimensional model of the locker, synchronizes the locker compartment usage status information in the central server in real time, distinguishes by different colors or logos, and supports users to perform interactive operations such as zooming and rotating; the remote operation simulation submodule allows users to simulate batch delivery and retrieval operations on the virtual model, input operation instructions, the system performs simulation calculations and feasibility checks, and generates actual operation instructions based on the final results; the real-time synchronization update submodule installs sensors on the smart locker to monitor the opening and closing of the box door and the storage status of items in real time, transmits the data to the central server, and the server updates the database information and synchronizes it to the virtual model.

[0015] Preferably, the multimodal recognition technology submodule integrates barcode scanning, QR code recognition, RFID recognition and image recognition technologies; the batch identity authentication submodule supports couriers or users to upload multiple identity information at one time, connects with a third-party identity authentication platform, uses an encryption algorithm for batch identity authentication, and promptly feeds back the verification results; the abnormal information warning submodule formulates abnormal information judgment rules, monitors the data in the identification and verification process in real time, and immediately issues an alarm once an abnormality is found, and records detailed information.

[0016] Preferably, the inventory monitoring and early warning submodule installs weight sensors, liquid level sensors, etc. in the locker to monitor the inventory of consumables in real time, and automatically issues a replenishment notification when the inventory is lower than the set threshold; the item recovery and processing submodule automatically marks items that have not been taken away for a long time or are damaged, uses robotic arms and other equipment to remove them from the box, places them in the designated recycling area, and classifies them; the data collection and collation submodule collects batch delivery and retrieval related data from multiple data sources such as smart lockers, user-end devices, and central servers, cleans and converts them, and stores them in a database.

[0017] Preferably, the behavior analysis and prediction submodule uses data mining algorithms and machine learning models to analyze users' delivery and pickup behavior patterns and predict future delivery and pickup needs; the system optimization suggestion submodule provides suggestions for system function improvement, box layout adjustment, service process optimization, etc. based on the data analysis results.

[0018] A method for delivering and retrieving items in a locker batch delivery and retrieval system, comprising the following steps:

[0019] 1. Batch submission

[0020] S1. Task initiation and order upload: The courier logs in to the client device, selects "batch delivery", and uploads the order by scanning the code or importing the file. After verification by the device, the order is encrypted and transmitted to the central server.

[0021] S2, order processing and route planning: The central server receives orders. The order parsing and grouping submodule uses the K-Means clustering algorithm to group items by their characteristics and addresses. The optimal route planning submodule uses the A* algorithm combined with the locker status to plan the route, taking into account obstacles and other factors, and then feeds back to the user device.

[0022] S3: Item identification and identity verification: The courier arrives at the locker group and places the item in the identification area. The multimodal recognition technology submodule quickly reads the information. At the same time, the courier enters their identity information into the device. The batch identity verification submodule connects with the third-party platform for encrypted verification. The abnormal information warning submodule detects anomalies and issues alarms.

[0023] S4: Item placement and progress update. After verification, the central server issues a compartment opening instruction. The courier places the item according to the route and number. After each compartment is placed, the device confirms the progress and provides real-time feedback. The server updates the compartment status.

[0024] S5, consumables monitoring and replenishment, the inventory monitoring and early warning submodule monitors the consumables inventory through sensors, and sends a replenishment notification when the inventory is below the threshold. When consumables are available, the automatic replenishment mechanism submodule delivers consumables according to instructions;

[0025] S6: Delivery completion and information synchronization. After delivery is complete, the courier confirms the delivery on the device, and the device sends a signal to the central server. The server updates the box status, synchronizes the information to the third-party logistics system, and generates a delivery report.

[0026] 2. Batch Pickup

[0027] S7, Task initiation and identity verification: The user logs in to the device and selects "Batch Pickup", enters the pickup code or scans the code to upload the order, and verifies the identity through fingerprint, face, etc. The device encrypts and transmits the information to the central server for verification. If the server fails, the user is prompted to proceed;

[0028] S8, Order Processing and Route Planning: The central server checks the location of items and bins. The order parsing and grouping submodule groups items by location and distribution. The optimal route planning submodule uses the A* algorithm to plan routes, combines historical data to estimate pickup times, and displays the route and time on the device.

[0029] S9: The central server opens the bin according to the path. The user follows the navigation and prompts to retrieve the item. After taking the item, the user confirms the receipt on the device. The system closes the door of the bin that has been taken and opens the next door to guide the user to retrieve the item.

[0030] S10, exception handling and feedback. If there is an exception such as a box door failure or mismatched items during pickup, the exception information warning submodule will alarm. The user can provide feedback on the device, and the central server will notify the management staff to handle the problem.

[0031] S11, Pickup completion and information update: After the user confirms the pickup, the device sends a signal to the central server, which updates the bin status, synchronizes the information to the third-party logistics system, and generates a pickup report;

[0032] S12. Data analysis and system optimization. The data analysis and optimization module collects pickup data, the data collection and organization submodule cleans and organizes the data, the behavior analysis and prediction submodule uses algorithms for analysis, and the system optimization suggestion submodule provides improvement suggestions and feedback to the development and operation team.

[0033] The present invention has the following beneficial effects:

[0034] 1. This invention uses the batch task intelligent planning module with the help of the K-Means clustering algorithm and the A* algorithm to automatically complete order grouping and route planning, effectively reducing the walking distance and operation time of couriers and users. The multimodal recognition technology submodule and the batch identity verification submodule can quickly process item information and identity verification, significantly improving the overall efficiency of batch delivery and pickup.

[0035] 2. The virtual locker mapping module of the present invention presents a virtual map through 3D modeling. Users can perform interactive operations such as zooming and rotating, and can also simulate the operation of dropping and picking items to familiarize themselves with the process in advance. At the same time, the system provides real-time information synchronization and intelligent guidance, allowing users to grasp the status of the locker at any time, greatly improving the convenience and smoothness of operation and enhancing user satisfaction.

[0036] 3. The time estimation and scheduling submodule of the present invention rationally arranges bin resources through a time prediction model to avoid idle or wasted resources. The inventory monitoring and early warning submodule monitors consumable inventory in real time, the automatic replenishment mechanism submodule replenishes consumables in a timely manner, and the item recycling submodule classifies and recycles damaged items, comprehensively optimizing resource management and reducing operating costs.

[0037] 4. This invention integrates multiple recognition technologies with a rapid identification and verification module, interfaces with third-party authentication platforms, and employs encryption algorithms to ensure the accuracy and security of identity verification. The abnormal information warning submodule monitors in real time and immediately issues an alarm upon detection, providing comprehensive security for batch delivery and pickup operations and effectively reducing risk. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a system block diagram of a locker batch delivery and retrieval system proposed by the present invention;

[0039] Figure 2 A batch delivery flow chart of a batch delivery method for a locker batch delivery system proposed by the present invention;

[0040] Figure 3 A batch retrieval flow chart of a batch delivery and retrieval method for a locker batch delivery and retrieval system proposed by the present invention;

[0041] Figure 4 This is the code diagram of the K-Means clustering algorithm for the locker batch delivery and retrieval system proposed by the present invention;

[0042] Figure 5 This is the code diagram of the A* algorithm of the locker batch delivery and retrieval system proposed by the present invention. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0044] Reference Figure 1 A locker batch delivery and retrieval system includes a user terminal device, a central server, and a smart locker group. The user terminal device communicates with the central server via a network, and the central server communicates with the smart locker group via a network. The central server includes a batch task intelligent planning module, a virtual locker mapping module, a fast identification and verification module, an automatic replenishment and recycling module, and a data analysis and optimization module.

[0045] The batch task intelligent planning module consists of an order analysis and grouping submodule, an optimal route planning submodule, and a time estimation and scheduling submodule, which are used to analyze and group orders, plan delivery and pickup routes, estimate delivery and pickup times, and reasonably schedule container resources.

[0046] The order parsing and grouping submodule is connected to the user terminal device to obtain batch delivery and pickup order information, and uses the K-Means clustering algorithm to group the orders according to factors such as item size, weight, fragility and destination address.

[0047] The optimal path planning submodule establishes a three-dimensional map model of the locker group and uses the A* algorithm to plan the optimal delivery and pickup path for couriers or users based on order grouping information and locker usage status, and optimizes it by considering factors such as actual walking obstacles.

[0048] The time estimation and scheduling submodule collects and analyzes the operation time data of previous batch delivery and pickup tasks, establishes a time prediction model, estimates the time required for this task, and reasonably schedules the box resources according to the time and locker usage status.

[0049] The virtual locker mapping module consists of a 3D visualization display submodule, a remote operation simulation submodule, and a real-time synchronization update submodule. It is used to create a virtual map of the locker, facilitate users to simulate operations in the virtual map, and allow users to keep abreast of the latest locker status at any time.

[0050] The three-dimensional visualization display submodule uses 3D modeling software to create an accurate three-dimensional model of the locker, synchronizes the locker compartment usage status information in the central server in real time, distinguishes by different colors or logos, and supports users to perform interactive operations such as zooming and rotating; the remote operation simulation submodule allows users to simulate batch delivery and retrieval operations on the virtual model, input operation instructions, and the system performs simulation calculations and feasibility checks, and generates actual operation instructions based on the final results; the real-time synchronization update submodule installs sensors on the smart locker to monitor the opening and closing of the box door and the storage status of items in real time, and transmits the data to the central server. The server updates the database information and synchronizes it to the virtual model.

[0051] The rapid identification and verification module consists of a multimodal identification technology submodule, a batch identity verification submodule, and an abnormal information warning submodule. It is used to quickly and accurately identify batch item information, support batch user identity verification, and issue warnings in the event of abnormal item information or identity verification failure.

[0052] The multimodal identification technology submodule integrates barcode scanning, QR code recognition, RFID recognition and image recognition technologies; the batch identity authentication submodule supports couriers or users to upload multiple identity information at one time, connects with a third-party identity authentication platform, uses an encryption algorithm for batch identity authentication, and promptly feeds back the verification results; the abnormal information warning submodule formulates abnormal information judgment rules, monitors the data in the identification and verification process in real time, and immediately issues an alarm once an abnormality is found, and records detailed information.

[0053] The automatic replenishment and recycling module consists of an inventory detection and warning submodule, an automatic replenishment mechanism submodule, and an item recycling processing submodule, which are used to notify management personnel to replenish stocks, automatically replenish consumables, and automatically mark and recycle damaged items;

[0054] The inventory monitoring and early warning submodule installs weight sensors, liquid level sensors, etc. in the lockers to monitor the inventory of consumables in real time. When the inventory falls below the set threshold, it automatically issues a replenishment notification; the item recovery and processing submodule automatically marks items that have not been taken away for a long time or are damaged, uses robotic arms and other equipment to remove them from the box, places them in the designated recycling area, and classifies them; the data collection and collation submodule collects batch delivery and retrieval related data from multiple data sources such as smart lockers, user-end devices, and central servers, cleans and converts them, and stores them in the database.

[0055] The data analysis and optimization module consists of a data collection and organization submodule, a behavior analysis and prediction submodule, and a system optimization suggestion submodule. It is used to collect and organize various data on batch delivery and pickup, predict the batch delivery and pickup needs of different user groups in different time periods, and provide suggestions for system improvement and optimization.

[0056] The behavior analysis and prediction submodule uses data mining algorithms and machine learning models to analyze users' delivery and pickup behavior patterns and predict future delivery and pickup needs; the system optimization suggestion submodule provides suggestions for system function improvement, box layout adjustment, service process optimization, etc. based on the data analysis results.

[0057] Reference Figure 2-3 A batch delivery and retrieval method for a locker batch delivery and retrieval system includes the following steps:

[0058] 1. Batch submission

[0059] S1. Task initiation and order upload: The courier logs in to the client device, selects "batch delivery", and uploads the order by scanning the code or importing the file. After verification by the device, the order is encrypted and transmitted to the central server.

[0060] S2, order processing and route planning: The central server receives orders. The order parsing and grouping submodule uses the K-Means clustering algorithm to group items by their characteristics and addresses. The optimal route planning submodule uses the A* algorithm combined with the locker status to plan the route, taking into account obstacles and other factors, and then feeds back to the user device.

[0061] S3: Item identification and identity verification: The courier arrives at the locker group and places the item in the identification area. The multimodal recognition technology submodule quickly reads the information. At the same time, the courier enters their identity information into the device. The batch identity verification submodule connects with the third-party platform for encrypted verification. The abnormal information warning submodule detects anomalies and issues alarms.

[0062] S4: Item placement and progress update. After verification, the central server issues a compartment opening instruction. The courier places the item according to the route and number. After each compartment is placed, the device confirms the progress and provides real-time feedback. The server updates the compartment status.

[0063] S5, consumables monitoring and replenishment, the inventory monitoring and early warning submodule monitors the consumables inventory through sensors, and sends a replenishment notification when the inventory is below the threshold. When consumables are available, the automatic replenishment mechanism submodule delivers consumables according to instructions;

[0064] S6: Delivery completion and information synchronization. After delivery is complete, the courier confirms the delivery on the device, and the device sends a signal to the central server. The server updates the box status, synchronizes the information to the third-party logistics system, and generates a delivery report.

[0065] 2. Batch Pickup

[0066] S7, Task initiation and identity verification: The user logs in to the device and selects "Batch Pickup", enters the pickup code or scans the code to upload the order, and verifies the identity through fingerprint, face, etc. The device encrypts and transmits the information to the central server for verification. If the server fails, the user is prompted to proceed;

[0067] S8, Order Processing and Route Planning: The central server checks the location of items and bins. The order parsing and grouping submodule groups items by location and distribution. The optimal route planning submodule uses the A* algorithm to plan routes, combines historical data to estimate pickup times, and displays the route and time on the device.

[0068] S9: The central server opens the bin according to the path. The user follows the navigation and prompts to retrieve the item. After taking the item, the user confirms the receipt on the device. The system closes the door of the bin that has been taken and opens the next door to guide the user to retrieve the item.

[0069] S10, exception handling and feedback. If there is an exception such as a box door failure or mismatched items during pickup, the exception information warning submodule will alarm. The user can provide feedback on the device, and the central server will notify the management staff to handle the problem.

[0070] S11, Pickup completion and information update: After the user confirms the pickup, the device sends a signal to the central server, which updates the bin status, synchronizes the information to the third-party logistics system, and generates a pickup report;

[0071] S12. Data analysis and system optimization. The data analysis and optimization module collects pickup data, the data collection and organization submodule cleans and organizes the data, the behavior analysis and prediction submodule uses algorithms for analysis, and the system optimization suggestion submodule provides improvement suggestions and feedback to the development and operation team.

[0072] Example 1

[0073] This example describes in detail how to apply the formula and process of the K-Means clustering algorithm to the order parsing and grouping submodule:

[0074] 1. Data Preparation

[0075] First, the order parsing and grouping submodule obtains batch delivery and pickup order information from the third-party logistics system or user-end device. For each order, factors such as item size (such as length, width, and height), weight, fragility (which can be quantified as a numerical value, such as non-fragile is 0 and fragile is 1), and destination address (which can be converted into latitude and longitude coordinates through geocoding) can be combined into a multidimensional vector x. i , where =1,2,3...K.

[0076] Assume that each order is represented by a four-dimensional vector: x i =(x i1 ,x i2 ,x i3 ,x i4 ), which correspond to the size of the item (a comprehensive value that can be normalized), weight, fragility and the latitude and longitude of the destination (which can be converted into relative values ​​from the starting point, etc.).

[0077] 2. Initialize cluster centers

[0078] The K-Means algorithm requires the number of clusters to be determined in advance. This can be set according to actual business needs, such as different transportation routes, locker areas, etc. Randomly select K from N order data points as the initial cluster centers μ j , where j = 1, 2,…, K.

[0079] 3. Assign data points to clusters

[0080] For each order data point x i , calculate it to each cluster center μ j The distance d(x i ,μ j ), the commonly used distance metric is Euclidean distance, the formula is:

[0081]

[0082] Here 4 is the dimension of the vector, corresponding to the four features mentioned above.

[0083] Then the data point is assigned to the cluster C with the nearest cluster center. j In Chinese, that is:

[0084]

[0085] 4. Update cluster centers

[0086] For each cluster C j , calculate the mean of all data points in the cluster and use it as the new cluster center μ j , the formula is:

[0087]

[0088] Among them, |C j | represents cluster C j The number of data points in .

[0089] 5. Repeat steps 3 and 4

[0090] The process of assigning data points to clusters and updating cluster centers is repeated until the cluster centers no longer change or the maximum number of iterations is reached. At this point, each order data point has been assigned to a cluster, completing the grouping of orders.

[0091] And the code example of K-Means clustering algorithm is as follows Figure 4 shown.

[0092] Through the above steps and code examples, the K-Means clustering algorithm is applied to the order parsing and grouping submodule, which enables the function of grouping orders based on factors such as item size, weight, fragility, and destination address.

[0093] Example 2

[0094] This example describes in detail how to apply the formula and process of the A* algorithm to the optimal path planning submodule:

[0095] 1. Constructing the search space

[0096] Node Definition: In the 3D map model of the locker cluster, each locker location and key locations along the corridor (such as turns and intersections) are defined as nodes. Each node can be represented by 3D coordinates (x, y, z) to represent its position in space.

[0097] Edge definition: If two nodes can be directly passed between them (i.e., there are no obstacles blocking the way), then an edge is established between the two nodes. The weight of the edge can be determined based on factors such as the actual distance between the nodes and the travel time.

[0098] 2. Initialization parameters

[0099] Starting node and target node: Based on the order grouping information and the user's delivery and pickup requirements, the starting node (such as the courier's current location or the user's entrance to the locker area) and the target node (such as the location of the locker where the delivery needs to be made) are determined.

[0100] Heuristic function h(n): Select an appropriate heuristic function to estimate the cost from node n to the target node. In three-dimensional space, commonly used heuristic functions include Manhattan distance and Euclidean distance.

[0101] Manhattan distance: For two nodes n1 = (x1, y1, z1) and n2 = (x2, y2, z2), the Manhattan distance formula is:

[0102] h(n1)=|x1-x2|+|y1-y2|+|z1-z2|

[0103] Euclidean distance: The Euclidean distance formula is:

[0104]

[0105] 3. Calculate g(n), h(n), and f(n)

[0106] Calculation of g(n): g(n) represents the actual cost from the starting node to node n. In path planning, g(n) can be the sum of the weights of the edges from the starting node to node n. Suppose that from the starting node s to node n, a series of nodes n1, n2, ..., n k , the edge weights are w1,w2,…,w k ,but:

[0107]

[0108] Calculation of h(n): Based on the selected heuristic function, calculate the estimated cost h(n) from node n to the target node.

[0109] Calculation of f(n): f(n) represents the total estimated cost of node n, and is calculated according to the formula f(n) = g(n) + h(n).

[0110] 4. Node expansion and path search

[0111] Open List and Closed List: Use the Open List to store nodes to be expanded, and the Closed List to store nodes that have been expanded. Initially, the Open List only contains the starting node, and the Closed List is empty.

[0112] Node expansion: Select the node n with the smallest f(n) value from the open list for expansion. For each adjacent node m of node n, calculate its g(m), h(m), and f(m) values.

[0113] If node m is not in the open list or the closed list, add it to the open list and set its parent node to n.

[0114] If node m is already in the open list, compare the newly calculated g(m) value with the original g(m) value. If the new g(m) value is smaller, update g(m) and f(m) values, and update its parent node to n.

[0115] If node m is already in the closed list and the newly calculated g(m) value is smaller, it is removed from the closed list, added to the open list, and the g(m), f(m) values ​​and parent node are updated.

[0116] Path search: Repeat the above node expansion process until the open list is empty or the target node is found. If the target node is found, the optimal path from the start node to the target node can be obtained by backtracking to the target node's parent node.

[0117] 5. Consider actual walking obstacles

[0118] Obstacle handling: When building a 3D map model, actual obstacles (such as walls, pillars, etc.) are marked in the map. When expanding a node, if an adjacent node is blocked by an obstacle, it will not be considered as an expandable node, thus preventing the path from passing through the obstacle.

[0119] Path optimization: After obtaining the initial path, you can further consider the convenience and efficiency of actual walking and optimize the path. For example, avoid excessive tortuosity and reduce the number of turns.

[0120] And the code example of A* algorithm is as follows Figure 5 shown.

[0121] Through the above steps and code examples, the A* algorithm is applied to the optimal path planning submodule. This allows the courier or user to plan the optimal delivery and pickup path based on order grouping information and locker usage status, and to optimize the route by taking into account factors such as actual walking obstacles.

[0122] Example 3

[0123] This embodiment details how to apply the time prediction model to the time estimation and scheduling submodule:

[0124] 1. Data Collection and Preprocessing

[0125] Data Collection: The time estimation and scheduling submodule collects the operation time data of previous batch delivery and pickup tasks from the system's historical records, and also records relevant influencing factors such as order quantity, item type, locker distribution, etc. Suppose there are n pieces of historical data collected, and the operation time of the i-th piece of data is y i , the corresponding influencing factor vector is x i =(x i1 ,x i2 ,…,x ip ), where p is the number of influencing factors.

[0126] Data preprocessing: Clean the collected data and remove outliers and missing values. Then standardize the data and convert the influencing factor vector x i Convert to z i , so that the mean of each feature is 0 and the standard deviation is 1. The standardization formula is:

[0127]

[0128] where μ j is the standard deviation of the jth feature.

[0129] 2. Select a time prediction model and apply the formula

[0130] Linear regression model

[0131] Model formula: The linear regression model assumes that there is a linear relationship between the operation time y and the influencing factor vector x. Its formula is:

[0132] y=β0+β1x1+β2x2+…+β p x p +∈

[0133] where β0,β1,…,β p is the parameter of the model, ∈ is the error term, and it obeys a normal distribution with mean 0.

[0134] Parameter estimation: Use the least squares method to estimate the model parameters β. Let x be an n×(p+1) matrix, where the first column is all 1 and the remaining columns are the standardized influencing factor vector z i ; y is the n-dimensional operation time vector. Then the estimated value of parameter β is:

[0135]

[0136] Time estimation: For a new batch delivery task, the influencing factor vector is x new , after normalization, we get z new , z new Substitute into the linear regression model to get the estimated operation time

[0137]

[0138] ARIMA model

[0139] Model formula: The general form of the ARIMA (Autoregressive Integrated Moving Average) model is Φ(B)(1-B) d Y t =Θ(B)∈ t , where Y t is the time series (operation time), B is the backward shift operator, Φ(B)=1-φ1B-φ2B 2 -…-φ p B p is an autoregressive polynomial, Θ(B)=1-θ1B-θ2B 2 -…-θ q B q is the sliding average polynomial, p is the autoregressive order, d is the difference order, and q is the sliding average order.

[0140] Parameter estimation: Use methods such as maximum likelihood estimation to estimate the model parameters φ1,…,φ p and θ1,…,θ q This can be achieved using statistical software such as the statsmodels library in Python.

[0141] Time estimation: Based on the estimated parameters and historical operation time data, the future operation time is predicted. For the new time point t+h, the prediction formula of the ARIMA model can be used to estimate.

[0142] Neural network model (taking a simple multi-layer perceptron as an example)

[0143] Model structure: A multilayer perceptron consists of an input layer, a hidden layer, and an output layer. Assume the input layer has p neurons (corresponding to p influencing factors), the hidden layer has m neurons, and the output layer has 1 neuron (corresponding to the operation time).

[0144] Forward propagation formula: The calculation from the input layer to the hidden layer is:

[0145]

[0146] where w ij is the weight from the i-th neuron in the input layer to the j-th neuron in the hidden layer, b j is the bias of the jth neuron in the hidden layer, and σ is the activation function (such as the Sigmoid function).

[0147] The calculation from the hidden layer to the output layer is:

[0148]

[0149] where v j is the weight from the jth neuron in the hidden layer to the output layer, and c is the bias of the output layer.

[0150] Parameter training: Use backpropagation algorithm and gradient descent method to update the model parameters w ij ,v j ,b j ,c, so that the error between the model's predicted value and the actual value is minimized.

[0151] Time estimation: For new batch delivery and pickup tasks, the influencing factor vector is input into the trained neural network model, and the estimated operation time is obtained through forward propagation calculation.

[0152] 3. Schedule container resources based on estimated time

[0153] Resource scheduling strategy: based on estimated operation time Based on the current locker usage status, a reasonable locker resource scheduling strategy can be developed. For example, if the estimated delivery time is long and the number of orders is large, adjacent lockers can be prioritized to facilitate the courier or user's operation; if the estimated delivery time is short, free lockers can be flexibly allocated.

[0154] Dynamic Adjustment: During actual operations, we monitor the progress and actual operation time in real time. If the actual operation time deviates significantly from the estimated time, we adjust the scheduling strategy for the container resources in a timely manner to improve system efficiency and resource utilization.

[0155] Through the above steps, the formula of the time prediction model is applied to the time estimation and scheduling sub-module, realizing the estimation of the time required for batch delivery and pickup tasks and the reasonable scheduling of box resources.

[0156] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A locker batch delivery and retrieval system, comprising a user terminal device, a central server, and a smart locker group, wherein the user terminal device communicates with the central server via a network, and the central server communicates with the smart locker group via a network, characterized in that: The central server includes a batch task intelligent planning module, a virtual locker mapping module, a rapid identification and verification module, an automatic replenishment and recycling module, and a data analysis and optimization module; The batch task intelligent planning module consists of an order analysis and grouping submodule, an optimal route planning submodule, and a time estimation and scheduling submodule, which are used to analyze and group orders, plan delivery and pickup routes, estimate delivery and pickup times, and reasonably schedule container resources. The virtual locker mapping module consists of a 3D visualization display submodule, a remote operation simulation submodule, and a real-time synchronization update submodule. It is used to create a virtual map of the locker, facilitate users to simulate operations in the virtual map, and allow users to keep abreast of the latest locker status at any time. The rapid identification and verification module consists of a multimodal identification technology submodule, a batch identity verification submodule, and an abnormal information warning submodule. It is used to quickly and accurately identify batch item information, support batch user identity verification, and issue warnings in the event of abnormal item information or identity verification failure. The automatic replenishment and recycling module consists of an inventory detection and warning submodule, an automatic replenishment mechanism submodule, and an item recycling processing submodule, which are used to notify management personnel to replenish stocks, automatically replenish consumables, and automatically mark and recycle damaged items; The data analysis and optimization module consists of a data collection and organization submodule, a behavior analysis and prediction submodule, and a system optimization suggestion submodule. It is used to collect and organize various data on batch delivery and pickup, predict the batch delivery and pickup needs of different user groups in different time periods, and provide suggestions for system improvement and optimization.

2. A locker batch delivery and retrieval system according to claim 1, characterized in that: The order parsing and grouping submodule is connected to the user terminal device to obtain batch delivery and pickup order information, and uses the K-Means clustering algorithm to group the orders according to factors such as item size, weight, fragility, and destination address; the optimal path planning submodule establishes a three-dimensional map model of the locker group, uses the A* algorithm, and plans the optimal delivery and pickup path for the courier or user based on the order grouping information and the locker usage status, and optimizes it by considering factors such as actual walking obstacles.

3. The locker batch delivery and retrieval system according to claim 1, characterized in that: The time estimation and scheduling submodule collects and analyzes the operation time data of previous batch delivery and pickup tasks, establishes a time prediction model, estimates the time required for this task, and reasonably schedules the box resources according to the time and locker usage status.

4. The locker batch delivery and retrieval system according to claim 1, characterized in that: The 3D visualization display submodule uses 3D modeling software to create an accurate 3D model of the locker, synchronizes the locker compartment usage status information with the central server in real time, distinguishes them by different colors or logos, and supports interactive operations such as zooming and rotating. The remote operation simulation submodule allows users to simulate batch delivery and retrieval operations on the virtual model, input operation instructions, and the system performs simulation calculations and feasibility checks, and generates actual operation instructions based on the final results. The real-time synchronization update submodule installs sensors on the smart lockers to monitor the opening and closing of the doors and the storage of items in real time, and transmits the data to the central server. The server updates the database information and synchronizes it to the virtual model.

5. The locker batch delivery and retrieval system according to claim 1, characterized in that: The multimodal identification technology submodule integrates barcode scanning, QR code recognition, RFID recognition and image recognition technologies; the batch identity authentication submodule supports couriers or users to upload multiple identity information at one time, connects with a third-party identity authentication platform, uses an encryption algorithm for batch identity authentication, and promptly feeds back the verification results; the abnormal information warning submodule formulates abnormal information judgment rules, monitors the data in the identification and verification process in real time, and immediately issues an alarm once an abnormality is found, and records detailed information.

6. The locker batch delivery and retrieval system according to claim 1, characterized in that: The inventory monitoring and early warning submodule installs weight sensors, liquid level sensors, etc. in the lockers to monitor the inventory of consumables in real time. When the inventory falls below the set threshold, it automatically issues a replenishment notification; the item recovery and processing submodule automatically marks items that have not been taken away for a long time or are damaged, uses robotic arms and other equipment to remove them from the box, places them in the designated recycling area, and classifies them; the data collection and collation submodule collects batch delivery and retrieval related data from multiple data sources such as smart lockers, user-end devices, and central servers, cleans and converts them, and stores them in the database.

7. The locker batch delivery and retrieval system according to claim 1, characterized in that: The behavior analysis and prediction submodule uses data mining algorithms and machine learning models to analyze users' delivery and pickup behavior patterns and predict future delivery and pickup needs; The system optimization suggestion submodule provides suggestions for system function improvement, box layout adjustment, service process optimization, etc. based on the data analysis results.

8. A batch delivery method for a locker batch delivery system according to any one of claims 1 to 7, characterized in that: The following steps are involved:

1. Batch submission S1. Task initiation and order upload: The courier logs in to the client device, selects "batch delivery," and uploads the order by scanning a QR code or importing a file. After verification by the device, the order is encrypted and transmitted to the central server. S2, order processing and route planning: The central server receives orders. The order parsing and grouping submodule uses the K-Means clustering algorithm to group items by their characteristics and addresses. The optimal route planning submodule uses the A* algorithm combined with the locker status to plan the route, taking into account obstacles and other factors, and then feeds back to the user device. S3: Item identification and identity verification: The courier arrives at the locker group and places the item in the identification area. The multimodal recognition technology submodule quickly reads the information. At the same time, the courier enters their identity information into the device. The batch identity verification submodule connects with the third-party platform for encrypted verification. The abnormal information warning submodule detects anomalies and issues alarms. S4: Item placement and progress update. After verification, the central server issues a compartment opening instruction. The courier places the item according to the route and number. After each compartment is placed, the device confirms the progress and provides real-time feedback. The server updates the compartment status. S5, consumables monitoring and replenishment, the inventory monitoring and early warning submodule monitors the consumables inventory through sensors, and sends a replenishment notification when the inventory is below the threshold. When consumables are available, the automatic replenishment mechanism submodule delivers consumables according to instructions; S6: Delivery completion and information synchronization. After delivery is complete, the courier confirms the delivery on the device, and the device sends a signal to the central server. The server updates the box status, synchronizes the information to the third-party logistics system, and generates a delivery report.

2. Batch Pickup S7: Task initiation and identity verification: The user logs in to the device and selects "Batch Pickup." Enter the pickup code or scan the code to upload the order. Identity verification is performed using fingerprint or facial recognition. The device encrypts and transmits the information to the central server for verification. If the server fails, the user is prompted to proceed. S8, Order Processing and Route Planning: The central server checks the location of items and bins. The order parsing and grouping submodule groups items by location and distribution. The optimal route planning submodule uses the A* algorithm to plan routes, combines historical data to estimate pickup times, and displays the route and time on the device. S9: The central server opens the bin according to the path. The user follows the navigation and prompts to retrieve the item. After taking the item, the user confirms the receipt on the device. The system closes the door of the bin that has been taken and opens the next door to guide the user to retrieve the item. S10, exception handling and feedback. If there is an exception such as a box door failure or mismatched items during pickup, the exception information warning submodule will alarm. The user can provide feedback on the device, and the central server will notify the management staff to handle the problem. S11, Pickup completion and information update: After the user confirms the pickup, the device sends a signal to the central server, which updates the bin status, synchronizes the information to the third-party logistics system, and generates a pickup report; S12. Data analysis and system optimization. The data analysis and optimization module collects pickup data, the data collection and organization submodule cleans and organizes the data, the behavior analysis and prediction submodule uses algorithms for analysis, and the system optimization suggestion submodule provides improvement suggestions and feedback to the development and operation team.

Images