Campus express delivery management method and system based on unmanned vehicle

Abstract

The invention provides a campus express distribution management method and system based on an unmanned vehicle, and relates to the field of express distribution management, and the method comprises the steps: building a first distribution map according to the position of an express station and the position of a distribution point; determining an optimal unmanned vehicle configuration scheme according to the first distribution map and the historical express distribution demand of the distribution point; determining a charging station configuration scheme according to the first distribution map, the historical express distribution demand of the distribution point and the optimal unmanned vehicle configuration scheme; establishing a second distribution map according to the position of each charging station and the first distribution map; determining a path monitoring scheme according to the second distribution map; acquiring real-time path state information according to the path monitoring scheme; acquiring a real-time delivery order; and according to the second delivery map, the real-time path state information and the real-time delivery order, carrying out express delivery management on the unmanned vehicle, and having the advantages of realizing unmanned delivery of campus express delivery and improving the express delivery efficiency.

Description

Technical Field

[0001] The present invention relates to the field of express delivery management, and particularly to a campus express delivery management method and system based on driverless vehicles. Background Art

[0002] With the continuous improvement of people's living standards and the rapid development of the Internet, people's lifestyles have also changed rapidly. Nowadays, online shopping has become one of the main consumption channels in life. As a crucial link in online shopping, the express delivery industry has been booming. To adapt to the high-speed development of the express delivery industry and meet the usage needs of users, "Internet + intelligent logistics" has become the main development trend. Contemporary college students are also the mainstream group of online shopping. For the delivery method of the last mile of express delivery, it is mainly carried out manually by delivery personnel. This not only requires a large number of delivery personnel, increasing labor costs, but also in order to improve work efficiency, delivery personnel usually need to ride two-wheeled or three-wheeled motorcycles. Riding motorcycles not only increases traffic congestion but also poses a safety hazard to the delivery personnel themselves, lacking safety. At the same time, when using the method of manual delivery by delivery personnel, it is difficult to ensure the security of user information, and there is a lack of flexibility in delivery time, affecting delivery efficiency.

[0003] Therefore, it is necessary to provide a campus express delivery management method and system based on driverless vehicles to achieve unmanned delivery of campus express and improve express delivery efficiency. Summary of the Invention

[0004] The present invention provides a campus express delivery management method based on driverless vehicles, including: obtaining the locations of express stations, delivery points, and the historical express delivery demands of the delivery points; establishing a first delivery map according to the locations of the express stations and the delivery points, wherein the first delivery map includes express station nodes, delivery point nodes, and path nodes from any express station node to any delivery point node; determining an optimal driverless vehicle configuration plan according to the first delivery map and the historical express delivery demands of the delivery points; determining a charging station configuration plan according to the first delivery map, the historical express delivery demands of the delivery points, and the optimal driverless vehicle configuration plan, wherein the charging station configuration plan at least includes the location of each charging station and the charging configuration; establishing a second delivery map according to the location of each charging station and the first delivery map, wherein the second delivery map includes express station nodes, delivery point nodes, charging station nodes, path nodes from any express station node to any delivery point node, and path nodes from any delivery point node to any charging station node; determining a path monitoring plan according to the second delivery map; obtaining real-time path status information according to the path monitoring plan; obtaining real-time delivery orders; and performing express delivery management on the driverless vehicles according to the second delivery map, the real-time path status information, and the real-time delivery orders.

[0005] Further, determining an optimal driverless vehicle configuration plan according to the first delivery map and the historical express delivery demands of the delivery points includes: for any two delivery points, determining the single-node maximum path similarity of the two delivery points corresponding to each express station node according to the first delivery map, and calculating the global path similarity of the two delivery points according to the single-node maximum path similarity of the two delivery points corresponding to each express station node; for any two delivery points, calculating the express delivery demand correlation parameter of the two delivery points according to the historical express delivery demands of the two delivery points; dividing multiple delivery points into multiple delivery point units according to the global path similarity and the express delivery demand correlation parameter of any two delivery points; and for each delivery point unit, determining the optimal driverless vehicle configuration plan according to the historical express delivery demands of the delivery points included in the delivery point unit and the first delivery map, wherein the optimal driverless vehicle configuration plan includes the optimal number of driverless vehicles corresponding to each delivery point unit.

[0006] Further, determining the optimal number of driverless vehicles corresponding to the distribution point unit includes: for each distribution point included in the distribution point unit, determining the initial current express delivery demand of the distribution point according to the historical express delivery demand of the distribution point; according to the express delivery demand correlation parameter between any two distribution points, correcting the initial current express delivery demand of each distribution point included in the distribution point unit to determine the current express delivery demand of each distribution point included in the distribution point unit; according to the current express delivery demand of each distribution point included in the distribution point unit, determining the unit express delivery demand corresponding to the distribution point unit; according to the current express delivery demand of each distribution point included in the distribution point unit and the positions of the distribution points included in the distribution point unit, determining the central position corresponding to the distribution point unit; establishing a set of constraint conditions, where the set of constraint conditions at least includes a transportation speed constraint and a loading constraint; according to the central position corresponding to the distribution point unit, the first distribution map, the unit express delivery demand corresponding to the distribution point unit, and the set of constraint conditions, determining the optimal number of driverless vehicles corresponding to the distribution point unit.

[0007] Further, the set of constraint conditions further includes a battery life constraint and a charging efficiency constraint; according to the first distribution map, the historical express delivery demand of the distribution point, and the optimal driverless vehicle configuration plan, determining a charging station configuration plan, including: for each distribution point unit, determining the charging demand corresponding to the distribution point unit according to the central position corresponding to the distribution point unit, the first distribution map, the unit express delivery demand corresponding to the distribution point unit, the optimal number of driverless vehicles corresponding to the distribution point unit, and the battery life constraint; merging multiple distribution point units according to the central position and charging demand corresponding to each distribution point unit to determine multiple distribution point groups; for each distribution point group, determining the position and charging configuration of the charging station corresponding to the distribution point group according to the central position and charging demand of each distribution point unit included in the distribution point group.

[0008] Further, according to the second distribution map, determining a path monitoring plan, including: based on the second distribution map, determining multiple candidate monitoring sections; obtaining the historical traffic flow data of each candidate monitoring section; for any two candidate monitoring sections, calculating the traffic flow correlation parameter between the two candidate monitoring sections according to the historical traffic flow data of the two candidate monitoring sections; determining the connectivity relationship between any two candidate monitoring sections; through a clustering algorithm, dividing the multiple candidate monitoring sections into multiple section clusters according to the traffic flow correlation parameter between any two candidate monitoring sections and the connectivity relationship between any two candidate monitoring sections; determining a path monitoring plan according to the multiple section clusters, where the path monitoring plan includes multiple path monitoring points, and one path monitoring point corresponds to one section cluster.

[0009] Further, according to the traffic correlation parameters of any two candidate monitoring sections and the connectivity relationship between any two candidate monitoring sections, the multiple candidate monitoring sections are divided into multiple section clusters, including: S11. Initialize the number of section clusters as K, where K is a positive integer; S12. Determine K centroids from the multiple candidate monitoring sections according to the number of section clusters; S13. For each candidate monitoring section, assign the candidate monitoring section to the section cluster where the centroid with the largest traffic correlation parameter and the connectivity relationship satisfying the preset connectivity relationship condition is located; S14. Determine whether each candidate monitoring section has been assigned. If so, execute S15; if not, execute S13; S15. For each section cluster, calculate the intra-cluster features according to the traffic correlation parameters of any two candidate monitoring sections included in the section cluster; S16. Determine whether the intra-cluster features of each section cluster meet the preset intra-cluster feature requirements. If so, execute S18; if not, execute S17; S17. Take the section clusters whose intra-cluster features do not meet the preset intra-cluster feature requirements as abnormal section clusters, update the centroids of the abnormal section clusters according to the traffic correlation parameters of any two candidate monitoring sections included in the abnormal section clusters, and execute S13; S18. Calculate the global clustering features according to the traffic correlation parameters of the K candidate monitoring sections corresponding to the K centroids; S19. Determine whether the global clustering features meet the preset global clustering feature requirements. If so, complete the clustering; if not, update the number K of section clusters according to the global clustering features, and execute S12.

[0010] Further, according to the second delivery map, real-time path status information, and real-time delivery orders, express delivery management is performed on the unmanned vehicle, including: determining the real-time delivery order corresponding to each delivery point unit according to the real-time delivery order; for each delivery point unit, generating a delivery task corresponding to the delivery point unit according to the loading constraint and the real-time delivery order corresponding to the delivery point unit, where one delivery task corresponds to one unmanned vehicle; for each of the delivery tasks, generating a delivery path corresponding to the delivery task according to the second delivery map and the real-time path status information, and controlling the unmanned vehicle corresponding to the delivery task to perform express delivery according to the delivery path corresponding to the delivery task.

[0011] Further, generating a delivery path corresponding to the delivery task according to the second delivery map and the real-time path status information includes: using a path generation model according to the second delivery map and the real-time path status information, where the path optimization model is a deep learning model, and the reward function of the path optimization model is related to the path length and the traffic of path monitoring points.

[0012] Further, for express delivery management of the unmanned vehicle based on the second delivery map, real-time path status information, and real-time delivery orders, it further includes: for each delivery task, predicting the remaining power of the unmanned vehicle when it completes the delivery task according to the initial power of the unmanned vehicle, real-time path status information, the delivery path corresponding to the delivery task, and the loading capacity; judging whether to charge according to the remaining power of the unmanned vehicle; if so, generating a charging path according to the charging station configuration plan and real-time status information of each charging station; after the unmanned vehicle completes the delivery task, controlling the unmanned vehicle to charge according to the charging path.

[0013] The present invention provides a campus express delivery management system based on an unmanned vehicle, which applies the above-mentioned campus express delivery management method based on an unmanned vehicle, and includes: a data acquisition module for acquiring the locations of express stations, delivery points, and historical express delivery demands of delivery points; a map establishment module for establishing a first delivery map according to the locations of the express stations and the delivery points, where the first delivery map includes express station nodes, delivery point nodes, and path nodes from any express station node to any delivery point node; a scheme determination module for determining an optimal unmanned vehicle configuration plan according to the first delivery map and the historical express delivery demands of delivery points; the scheme determination module is further used for determining a charging station configuration plan according to the first delivery map, the historical express delivery demands of delivery points, and the optimal unmanned vehicle configuration plan, where the charging station configuration plan at least includes the location of each charging station; the map establishment module is further used for establishing a second delivery map according to the location of each charging station and the first delivery map, where the second delivery map includes express station nodes, delivery point nodes, charging station nodes, path nodes from any express station node to any delivery point node, and path nodes from any delivery point node to any charging station node; the scheme determination module is further used for determining a path monitoring plan according to the second delivery map; a path monitoring module for acquiring real-time path status information according to the path monitoring plan; an order acquisition module for acquiring real-time delivery orders; a delivery management module for performing express delivery management on the unmanned vehicle according to the second delivery map, real-time path status information, and real-time delivery orders.

[0014] Compared with the prior art, the campus express delivery management method and system based on an unmanned vehicle provided by the present invention at least have the following beneficial effects:

[0015] 1. The driverless vehicle can navigate autonomously, locate in real-time, and plan routes, thus completing the express delivery task quickly and accurately. This significantly shortens the delivery time and improves the delivery efficiency. Through intelligent scheduling, the driverless vehicle can make dynamic adjustments according to the real-time path status information and real-time delivery orders, further optimizing the delivery route and reducing the waiting and empty running time. The driverless vehicle can replace some human labor and reduce the labor cost expenditure. Especially in scenarios such as campuses where the population is dense but the delivery demands are scattered, the application of driverless vehicles can significantly reduce the number of couriers employed and their labor intensity. Driverless vehicle delivery can reduce the delivery time, improve the operation efficiency, and thus lower the overall operation cost. It realizes intelligent management, including functions such as intelligent scheduling, route optimization, regular maintenance, and data analysis. This improves the intelligent level of the overall logistics system and reduces the management difficulty and cost.

[0016] 2. By comprehensively considering the historical express delivery demands and path similarities of delivery points, the number of driverless vehicles required for each delivery point unit can be determined more accurately. This avoids waste and shortage of resources and ensures the efficiency and accuracy of express delivery. Dividing the delivery points into multiple units and configuring driverless vehicles according to the demands of each unit can adjust resources more flexibly and improve resource utilization rate. This helps to reduce the operation cost and improve the overall economic benefits.

[0017] 3. By merging delivery point units and determining reasonable delivery point groups, the configuration of charging stations can be optimized, reducing the number of unnecessary charging stations and lowering the construction and operation costs. Determining the location and configuration of charging stations according to the charging demands of each delivery point group can ensure the efficient use of charging stations and avoid resource waste. Shorten the delivery time: By optimizing the charging strategy of driverless vehicles and the configuration of charging stations, the charging time of driverless vehicles can be shortened, thus shortening the overall delivery time and improving the user experience. Ensuring that the driverless vehicle can operate continuously and stably can improve the reliability of express delivery services and enhance users' trust and satisfaction with campus express delivery services.

[0018] 4. By using a clustering algorithm to divide multiple candidate monitoring road sections into multiple road section clusters and determine path monitoring points, the traffic flow conditions of each road section on campus can be grasped more accurately. This helps the driverless vehicle avoid congested road sections and choose a smoother route when selecting a delivery path, thus improving the delivery efficiency. The path monitoring solution can monitor the traffic conditions of each road section on campus in real time, including abnormal conditions such as congestion and accidents. The driverless vehicle can adjust the delivery route in a timely manner according to the real-time monitoring results, avoid getting stuck in traffic, and ensure the timeliness and accuracy of express delivery. Through the path monitoring solution, potential traffic risks such as road construction and traffic accidents can be detected and warned in a timely manner. The driverless vehicle can avoid risk areas in advance according to the warning information to ensure the safety and reliability of the delivery process. In case of emergencies such as traffic congestion or accidents, the path monitoring solution can quickly provide relevant information to help the driverless vehicle adjust the delivery route and ensure the continuity and stability of express delivery. At the same time, this also helps the campus management department take timely measures to relieve traffic pressure and restore campus traffic order. By using a clustering algorithm to divide the candidate monitoring road sections into multiple road section clusters and determining path monitoring points according to the flow correlation parameters and connectivity relationships of the road section clusters, the monitoring resources can be allocated more reasonably. This helps reduce the monitoring cost and improve the monitoring efficiency. The path monitoring solution can focus on monitoring the road sections with large traffic flow on campus to improve the monitoring accuracy. This helps to grasp the traffic conditions on campus more accurately and provide more reliable data support for the delivery path planning of the driverless vehicle.

[0019] 5. Generate corresponding delivery tasks based on the real-time delivery orders of each delivery point unit. This ensures that the driverless vehicle can quickly respond to order demands, reduce waiting time, and improve delivery efficiency. By using a deep learning model as the path optimization model and combining the second delivery map and real-time path status information, the optimal delivery path is generated. This path optimization not only considers the path length but also the traffic flow at the path monitoring points, enabling it to avoid congested sections and improve delivery speed and accuracy. When generating delivery tasks, this method takes into account the loading constraints to ensure that the load of each driverless vehicle is within its carrying capacity. This avoids delivery delays or driverless vehicle failures caused by overloading and improves delivery efficiency. It can dynamically adjust the delivery path according to the real-time path status information. When encountering abnormal situations such as traffic congestion and road construction, the driverless vehicle can quickly adjust the route to ensure the continuity and stability of delivery. By predicting the remaining power of the driverless vehicle when it completes the delivery task and generating a charging path based on the charging station configuration plan and the real-time status information of each charging station, this method can ensure that the driverless vehicle is charged in time when the power is insufficient. This avoids delivery interruptions caused by power depletion and enhances the reliability of the system. It can real-time monitor the traffic conditions on campus, including abnormal situations such as congestion and accidents. In case of an emergency, the driverless vehicle can quickly adjust the delivery route according to the real-time monitoring results and the path optimization model to ensure the timeliness and accuracy of express delivery. Generate delivery tasks according to the real-time delivery orders and loading constraints and assign them to the corresponding driverless vehicles. This ensures the reasonable allocation and utilization of driverless vehicles and avoids resource waste. By optimizing the charging management, this method can ensure that the driverless vehicle performs tasks when the power is sufficient, reducing unnecessary charging times and charging costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] This specification will be further described by way of exemplary embodiments, which will be described in detail through the accompanying drawings. These embodiments are not restrictive. In these embodiments, the same numbers represent the same structures, where:

[0021] Figure 1 is a flowchart showing the method for managing campus express delivery based on driverless vehicles according to some embodiments of this specification;

[0022] Figure 2 is a flowchart showing the method for determining the optimal number of driverless vehicles corresponding to a delivery point unit according to some embodiments of this specification;

[0023] Figure 3 is a flowchart showing the method for dividing multiple candidate monitoring sections into multiple section clusters according to some embodiments of this specification;

[0024] Figure 4It is a schematic diagram of the modules of a campus express delivery management system based on driverless vehicles shown in some embodiments of this specification. Detailed implementation manners

[0025] To more clearly illustrate the technical solutions of the embodiments of this specification, the accompanying drawings required for the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some examples or embodiments of this specification. For those of ordinary skill in the art, without creative efforts, this specification can also be applied to other similar scenarios based on these drawings. Unless obvious from the language context or otherwise stated, the same reference numerals in the drawings represent the same structure or operation.

[0026] Figure 1 It is a schematic flowchart of a campus express delivery management method based on driverless vehicles shown in some embodiments of this specification. As Figure 1 shown, the campus express delivery management method based on driverless vehicles may include the following steps.

[0027] Step 110, obtain the locations of the express stations, the locations of the delivery points, and the historical express delivery demands of the delivery points.

[0028] Among them, the historical express delivery demands of the delivery points may include the number of express deliveries that need to be delivered from each express station to the delivery point at the delivery point in multiple historical periods (for example, one day).

[0029] Step 120, establish a first delivery map according to the locations of the express stations and the delivery points.

[0030] Among them, the first delivery map includes express station nodes, delivery point nodes, and path nodes from any express station node to any delivery point node.

[0031] Step 130, determine the optimal driverless vehicle configuration plan according to the first delivery map and the historical express delivery demands of the delivery points.

[0032] In some embodiments, step 130 specifically includes:

[0033] For any two delivery points, according to the first delivery map, determine the single-node maximum path similarity of the two delivery points corresponding to each express station node, and calculate the global path similarity of the two delivery points according to the single-node maximum path similarity of the two delivery points corresponding to each express station node;

[0034] For any two delivery points, calculate the express delivery demand correlation parameter of the two delivery points according to the historical express delivery demands of the two delivery points;

[0035] Divide multiple delivery points into multiple delivery point units according to the global path similarity between any two delivery points and the express delivery demand correlation parameter.

[0036] For each delivery point unit, determine the optimal unmanned vehicle configuration plan according to the historical express delivery demand of the delivery points included in the delivery point unit and the first delivery map, where the optimal unmanned vehicle configuration plan includes the optimal number of unmanned vehicles corresponding to each delivery point unit.

[0037] Specifically, for any two delivery points, all paths from each delivery point to each express station node can be determined according to the first delivery map. For the same express station node, the similarity of any path from any two delivery points to the express station node can be calculated. For example, for delivery point A1, delivery point A2, and express point B1, the paths from delivery point A1 to express point B1 include paths C11, C12, C13, and the paths from delivery point A2 to express station node B1 include paths C21, C22, C23. The similarity between C11 and C21, C22, C23 can be calculated, the similarity between C12 and C21, C22, C23 can be calculated, and the similarity between C13 and C21, C22, C23 can be calculated. First, take the maximum value of the similarity between C11 and C21, C22, C23, the maximum value of the similarity between C12 and C21, C22, C23, and the maximum value of the similarity between C13 and C21, C22, C23. Then, take the maximum value among the maximum value of the similarity between C11 and C21, C22, C23, the maximum value of the similarity between C12 and C21, C22, C23, and the maximum value of the similarity between C13 and C21, C22, C23 as the single-node maximum path similarity of delivery point A1, delivery point A2 corresponding to express station node B1.

[0038] For example, the global path similarity between two delivery points can be calculated according to the following formula:

[0039]

[0040] where S ((i,j),2) is the global path similarity between the i-th delivery point and the j-th delivery point, S (((i,j),n),1) is the single-node maximum path similarity between the i-th delivery point and the j-th delivery point corresponding to the n-th express station node, and N is the total number of express station nodes.

[0041] The express delivery demand correlation parameter between two delivery points can be calculated according to the following formula:

[0042]

[0043] where C (i,j) is the express delivery demand correlation parameter between the i-th delivery point and the j-th delivery point, C ((i,j),n)is the correlation parameter of the express delivery demand corresponding to the i-th delivery point and the j-th delivery point for the n-th express station node, N (i,t,n) is the number of express deliveries that need to be delivered from each n-th express station to the i-th delivery point in the t-th historical period, N (j,t,n) is the number of express deliveries that need to be delivered from each n-th express station to the j-th delivery point in the t-th historical period, and T is the total number of sampled historical periods.

[0044] For any two delivery points, the global path similarity and the correlation parameter of the express delivery demand of the two delivery points can be weighted and summed to calculate the combined parameter of the two delivery points. According to the combined parameter of any two delivery points, multiple delivery points can be clustered by a clustering algorithm (for example, the K-means clustering algorithm), and the multiple delivery points can be divided into multiple delivery point units.

[0045] Figure 2 is a schematic flowchart of determining the optimal number of unmanned vehicles corresponding to a delivery point unit according to some embodiments of the present specification, as Figure 2 shown. In some embodiments, determining the optimal number of unmanned vehicles corresponding to a delivery point unit includes:

[0046] For each delivery point included in the delivery point unit, according to the historical express delivery demand of the delivery point, determine the initial current express delivery demand of the delivery point;

[0047] According to the correlation parameter of the express delivery demand of any two delivery points, correct the initial current express delivery demand of each delivery point included in the delivery point unit to determine the current express delivery demand of each delivery point included in the delivery point unit;

[0048] According to the current express delivery demand of each delivery point included in the delivery point unit, determine the unit express delivery demand corresponding to the delivery point unit;

[0049] According to the current express delivery demand of each delivery point included in the delivery point unit and the positions of the delivery points included in the delivery point unit, determine the central position corresponding to the delivery point unit;

[0050] Establish a set of constraint conditions, where the set of constraint conditions at least includes a transportation speed constraint and a loading constraint;

[0051] According to the central position corresponding to the delivery point unit, the first delivery map, the unit express delivery demand corresponding to the delivery point unit, and the set of constraint conditions, determine the optimal number of unmanned vehicles corresponding to the delivery point unit.

[0052] Specifically, the initial current express delivery demand of the delivery point can be determined by a demand prediction model according to the historical express delivery demand of the delivery point, where the demand prediction model can be a long short-term memory network model.

[0053] The initial current express delivery demand of each delivery point included in the delivery point unit can be corrected according to the following process:

[0054] S21. For any two delivery points included in the delivery point unit, the express delivery demand correlation parameter before update of the two delivery points can be calculated according to the historical express delivery demands of the two delivery points;

[0055] S22. For any two delivery points included in the delivery point unit, the express delivery demand correlation parameter after update of the two delivery points can be calculated according to the historical express delivery demands and the initial current express delivery demands of the two delivery points;

[0056] S23. For any two delivery points included in the delivery point unit, calculate the difference in the express delivery demand correlation parameter between the express delivery demand correlation parameter before update and the express delivery demand correlation parameter after update of the two delivery points;

[0057] S24. Determine whether there is at least two delivery points with the difference in the express delivery demand correlation parameter greater than the express delivery demand correlation parameter difference threshold. If so, complete the correction. If not, execute S25;

[0058] S25. Correct the initial current express delivery demand of the delivery point according to the difference in the express delivery demand correlation parameter of any two delivery points included in the delivery point unit, and generate the current express delivery demand of each delivery point included in the delivery point unit;

[0059] S26. For any two delivery points included in the delivery point unit, the express delivery demand correlation parameter after update of the two delivery points can be calculated according to the historical express delivery demands and the current express delivery demands of the two delivery points;

[0060] S27. For any two delivery points included in the delivery point unit, calculate the difference in the express delivery demand correlation parameter between the express delivery demand correlation parameter before update and the express delivery demand correlation parameter after update of the two delivery points;

[0061] S28. Determine whether there is at least two delivery points with the difference in the express delivery demand correlation parameter greater than the express delivery demand correlation parameter difference threshold. If so, complete the correction. If not, execute S29;

[0062] S29. Correct the current express delivery demand of the delivery point according to the difference in the express delivery demand correlation parameter of any two delivery points included in the delivery point unit, generate the current express delivery demand of each delivery point included in the delivery point unit, and execute S26.

[0063] For example, the initial current express delivery demand of the distribution points can be corrected through genetic algorithms, particle swarm optimization algorithms, etc., to generate the current express delivery demand of each distribution point included in the distribution point unit. It is also possible to correct the current express delivery demand of the distribution points through genetic algorithms, particle swarm optimization algorithms, etc., to generate the current express delivery demand of each distribution point included in the distribution point unit.

[0064] The unit express delivery demand corresponding to the distribution point unit can be the sum of the current express delivery demands of each distribution point included in the distribution point unit.

[0065] Based on the current express delivery demand of each distribution point included in the distribution point unit, the weight of each distribution point included in the distribution point unit can be determined. For example, the ratio of the current express delivery demand of the distribution point to the sum of the current express delivery demands of all distribution points included in the distribution point unit can be used as the weight of the distribution point. According to the weight of each distribution point included in the distribution point unit, the weighted sum of the positions of each distribution point included in the distribution point unit is calculated to determine the central coordinates, and the distribution point position closest to the central coordinates is used as the central position corresponding to the distribution point unit.

[0066] The transportation speed constraint can include the maximum speed corresponding to roads with different roughness levels.

[0067] The loading constraint can include the maximum express loading weight corresponding to roads with different roughness levels.

[0068] Based on the central position corresponding to the distribution point unit, the first distribution map, the unit express delivery demand corresponding to the distribution point unit, and the set of constraint conditions, the optimal number of unmanned vehicles corresponding to the distribution point unit can be determined through a mathematical model, specifically including:

[0069] 1. Define variables:

[0070] Let the number of unmanned vehicles be N.

[0071] Let the loading capacity of each unmanned vehicle be W (which needs to satisfy the loading constraint).

[0072] Let the delivery time be T (which needs to satisfy the transportation speed constraint).

[0073] 2. Establish the objective function:

[0074] The goal is to minimize the number of unmanned vehicles N while satisfying the express delivery demands of all distribution points.

[0075] 3. Establish the constraint conditions:

[0076] According to the transportation speed constraint, ensure that the transportation speed does not exceed the maximum speed corresponding to the road with the given roughness level.

[0077] Determine the maximum load capacity of each driverless vehicle according to the loading constraints to ensure that it does not exceed the maximum express delivery load weight of the road surface corresponding to the roughness.

[0078] 4. Solve the model

[0079] Data preprocessing: Calculate the distances and required times from each express delivery station to the distribution points according to the first delivery map and the central positions corresponding to the distribution point units.

[0080] Algorithm selection: Select linear programming, integer programming, heuristic algorithms, etc. to solve the model and determine the optimal number of driverless vehicles corresponding to each distribution point unit.

[0081] The driverless vehicle can include a vehicle body, a motor drive module, and a navigation module.

[0082] Specifically, the vehicle body determines the appearance, structure, and performance of the driverless vehicle. At the same time, the battery module is the energy core of the driverless delivery vehicle, providing power for the vehicle. It not only affects the endurance of the driverless vehicle but also relates to the performance, weight, and safety of the driverless vehicle. According to the express delivery requirements, customize a suitable chassis, frame, car body, and wheels to meet the project size, weight requirements, and the subsequent integration needs of motors, batteries, and other modules. Select a suitable battery type and capacity according to the power requirements of the driverless vehicle. Purchase the battery module and install and fix it to ensure the stability and safety of the battery.

[0083] The motor drive module is one of its core components, and they are jointly responsible for the movement and navigation functions of the driverless vehicle. The motor is the "heart" of the driverless vehicle, responsible for converting electrical energy into mechanical energy to drive the driverless vehicle to move. The drive module is an important module for the movement trajectory of the driverless vehicle, mainly composed of a reducer, drive and steering motors, etc. It is responsible for receiving instructions from the controller and rotating the drive motor to make the driverless vehicle move along the preset trajectory. Select a suitable motor and drive system according to the project requirements and the design parameters of the AGV vehicle body. Ensure that its performance, quality, and compatibility meet the project requirements. Write a motor drive control algorithm according to the control requirements of the AGV to ensure that the motor can accurately respond to the preset logic and instructions and achieve stable and efficient movement.

[0084] The navigation module of the driverless vehicle is one of its core components. It is responsible for determining the current position of the driverless vehicle, planning the movement path, and guiding the driverless vehicle to reach the destination. This project mainly uses lidar, cameras, and GNSS for driverless vehicle navigation. When using cameras for navigation or positioning, obvious markers need to be added to the environment, and markers are set every 15m. These markers will provide clear visual references for the cameras, thereby improving the positioning accuracy and navigation stability of the driverless vehicle. RFID and QR code readers can be added to implement magnetic navigation and QR code navigation functions.

[0085] The control module in the driverless vehicle is a core component and acts as the control center of the driverless vehicle. The control module is the "brain" of the driverless delivery vehicle, responsible for receiving sensor data, processing information, and issuing control instructions. According to the functional requirements and control logic of the driverless vehicle, corresponding control software is developed. Integrate the navigation module, motor drive module, etc. to achieve functions such as autonomous navigation, obstacle avoidance, and path planning of the driverless vehicle.

[0086] The technical indicators of the driverless vehicle can include: maximum speed: 5m / s, rated speed: 0.5m / s, vehicle body weight: 70kg, endurance: >50km under room temperature conditions, positioning accuracy: in the campus environment: average error <0.3m, anti-collision distance: 0.5m within the range of -30° to -30°.

[0087] Step 140, determine the charging station configuration plan according to the first delivery map, the historical express delivery requirements of the delivery points, and the optimal driverless vehicle configuration plan.

[0088] Among them, the charging station configuration plan at least includes the location of each charging station and the charging configuration. Specifically, the charging configuration can include the number of charging piles, etc.

[0089] In some embodiments, the constraint set further includes battery endurance constraints and charging efficiency constraints. Among them, the charging efficiency constraint can be the charging amount per unit time of a single charging pile, etc.

[0090] In some embodiments, determining the charging station configuration plan according to the first delivery map, the historical express delivery requirements of the delivery points, and the optimal driverless vehicle configuration plan includes:

[0091] For each delivery point unit, determine the charging demand corresponding to the delivery point unit according to the central position corresponding to the delivery point unit, the first delivery map, the unit express delivery requirements corresponding to the delivery point unit, the optimal number of driverless vehicles corresponding to the delivery point unit, and the battery endurance constraint;

[0092] Merge multiple delivery point units according to the central position and charging demand corresponding to each delivery point unit to determine multiple delivery point groups. Among them, in the delivery point group, the distance between the central positions corresponding to any two delivery point units is less than the distance threshold and the sum of the charging demands corresponding to each delivery point unit in the delivery point group is less than the charging demand threshold;

[0093] For each delivery point group, determine the location and charging configuration of the charging station corresponding to the delivery point group according to the central position and charging demand corresponding to each delivery point unit included in the delivery point group.

[0094] Specifically, the charging demand corresponding to the distribution point unit can be determined by a demand prediction model based on the central position corresponding to the distribution point unit, the first distribution map, the unit express delivery demand corresponding to the distribution point unit, the optimal number of unmanned vehicles corresponding to the distribution point unit, and the battery life constraint. Among them, the demand prediction model can be a convolutional neural network model.

[0095] The location and charging configuration of the charging station corresponding to the distribution point group can be determined by a configuration optimization model based on the central position and charging demand of each distribution point unit included in the distribution point group. Among them, the configuration optimization model can be a convolutional neural network model.

[0096] Step 150: Establish a second distribution map according to the location of each charging station and the first distribution map.

[0097] Among them, the second distribution map includes express station nodes, distribution point nodes, charging station nodes, path nodes from any express station node to any distribution point node, and path nodes from any distribution point node to any charging station node.

[0098] Step 160: Determine a path monitoring plan according to the second distribution map.

[0099] In some embodiments, step 160 specifically includes:

[0100] Based on the second distribution map, determine multiple candidate monitoring sections. Among them, the multiple candidate monitoring sections can be all sections in the second distribution map or sections where traffic flow is prone to change;

[0101] Obtain the historical traffic flow data of each candidate monitoring section. Among them, the historical traffic flow data of the candidate monitoring section can include the number of people in the candidate monitoring section at multiple historical time points;

[0102] For any two candidate monitoring sections, calculate the traffic flow correlation parameter between the two candidate monitoring sections according to the historical traffic flow data of the two candidate monitoring sections. Among them, the method of calculating the traffic flow correlation parameter between the two candidate monitoring sections is similar to the method of calculating the express delivery demand correlation parameter between the i-th distribution point and the j-th distribution point corresponding to the n-th express station node, which will not be elaborated here;

[0103] Determine the connectivity relationship between any two candidate monitoring sections. Among them, the connectivity relationship between the two candidate monitoring sections can be the minimum number of candidate monitoring sections connecting the two candidate monitoring sections;

[0104] Through a clustering algorithm, divide the multiple candidate monitoring sections into multiple section clusters according to the traffic flow correlation parameter between any two candidate monitoring sections and the connectivity relationship between any two candidate monitoring sections;

[0105] Determine a path monitoring scheme according to multiple road segment clusters, where the path monitoring scheme includes multiple path monitoring points, and one path monitoring point corresponds to one road segment cluster.

[0106] Figure 3 It is a schematic flow chart of dividing multiple candidate monitoring road segments into multiple road segment clusters shown in some embodiments of this specification. As Figure 3 shown, in some embodiments, through a clustering algorithm, according to the traffic association parameters of any two candidate monitoring road segments and the connectivity relationship between any two candidate monitoring road segments, divide multiple candidate monitoring road segments into multiple road segment clusters, including:

[0107] S11. Initialize the number of road segment clusters as K, where K is a positive integer;

[0108] S12. According to the number of road segment clusters, determine K centroids from multiple candidate monitoring road segments. For example, randomly select K candidate monitoring road segments as K centroids;

[0109] S13. For each candidate monitoring road segment, assign the candidate monitoring road segment to the road segment cluster where the centroid with the largest traffic association parameter and the connectivity relationship satisfies the preset connectivity relationship condition. By way of example only, the preset connectivity relationship condition may be that the minimum number of candidate monitoring road segments connecting two candidate monitoring road segments is less than the minimum candidate monitoring road segment number threshold;

[0110] S14. Determine whether each candidate monitoring road segment has been assigned. If so, execute S15; if not, execute S13;

[0111] S15. For each road segment cluster, calculate the intra-cluster feature according to the traffic association parameters of any two candidate monitoring road segments included in the road segment cluster;

[0112] S16. Determine whether the intra-cluster feature of each road segment cluster satisfies the preset intra-cluster feature requirement. If so, execute S18; if not, execute S17. By way of example only, the preset intra-cluster feature requirement may be that the intra-cluster feature of the road segment cluster is less than the intra-cluster feature threshold;

[0113] S17. Take the road segment cluster whose intra-cluster feature does not satisfy the preset intra-cluster feature requirement as an abnormal road segment cluster, update the centroid of the abnormal road segment cluster according to the traffic association parameters of any two candidate monitoring road segments included in the abnormal road segment cluster, and execute S13;

[0114] S18. Calculate the global clustering feature according to the traffic association parameters of the K candidate monitoring road segments corresponding to the K centroids;

[0115] S19. Determine whether the global clustering feature meets the preset global clustering feature requirement. If so, complete the clustering. If not, update the number K of road segment clusters according to the global clustering feature, and execute S12. By way of example only, the preset global clustering feature requirement may be that the global clustering feature is less than the global clustering feature threshold.

[0116] Specifically, the intra-cluster feature can be calculated according to the following formula:

[0117]

[0118] where σ k is the intra-cluster feature of the k-th road segment cluster, γ (e,f) is the traffic correlation parameter between the e-th candidate monitoring road segment and the f-th candidate monitoring road segment included in the k-th road segment cluster, and E is the total number of candidate monitoring road segments included in the k-th road segment cluster.

[0119] For each candidate monitoring road segment included in the abnormal road segment cluster, the mean value of the traffic correlation parameters between this candidate monitoring road segment and each other candidate monitoring road segment included in this abnormal road segment cluster can be calculated.

[0120] Take the candidate monitoring road segment with the largest mean value as the new centroid of this abnormal road segment cluster.

[0121] The global clustering feature can be calculated according to the following formula:

[0122]

[0123] where φ is the global clustering feature, and γ (g,h) is the traffic correlation parameter between the g-th centroid and the h-th centroid.

[0124] The number K of road segment clusters can be updated based on the following formula according to the global clustering feature:

[0125] K t+1 = K t + △K × φ

[0126] where K t+1 is the updated number of road segment clusters, K t is the number of road segment clusters before update, and △K is the preset step size.

[0127] Step 170. Obtain the real-time path status information according to the path monitoring plan.

[0128] Specifically, the real-time path status information may include the pedestrian flow at each path monitoring point.

[0129] Step 180. Obtain the real-time delivery orders.

[0130] Specifically, the user can initiate a delivery order through a mobile terminal. The delivery order may include the express waybill number, the express station, and the delivery point.

[0131] Step 190: Manage the express delivery of the driverless vehicle according to the second delivery map, the real-time path status information, and the real-time delivery order.

[0132] In some embodiments, step 190 specifically includes:

[0133] Determine the real-time delivery order corresponding to each delivery point unit according to the real-time delivery order.

[0134] For each delivery point unit, generate a delivery task corresponding to the delivery point unit according to the loading constraint and the real-time delivery order corresponding to the delivery point unit, where one delivery task corresponds to one driverless vehicle.

[0135] For each delivery task, generate a delivery path corresponding to the delivery task according to the second delivery map and the real-time path status information, and control the driverless vehicle corresponding to the delivery task to perform express delivery according to the delivery path corresponding to the delivery task.

[0136] In some embodiments, generating a delivery path corresponding to a delivery task according to the second delivery map and the real-time path status information includes:

[0137] Use a path generation model based on the second delivery map and the real-time path status information. The path optimization model is a deep learning model, and the reward function of the path optimization model is related to the path length and the traffic flow of the path monitoring points. The shorter the path length and the less the traffic flow of the path monitoring points passed through, the greater the reward value.

[0138] The driverless vehicle can identify and store the express information by scanning and recognizing the two-dimensional code or bar code on the package. The recognized express information can be automatically matched with the user's order and stored in the internal system of the driverless vehicle. The driverless vehicle is equipped with a warehousing management system for managing the express information in the warehouse. Connect all the driverless vehicles, the dispatching system, the warehousing management system, etc. to the campus network to achieve data sharing and real-time communication. Strengthen network security measures to ensure the security and reliability of data transmission. After the driverless vehicle arrives at the delivery point, the user can pick up the package by scanning the two-dimensional code on the driverless vehicle.

[0139] In some embodiments, managing the express delivery of the driverless vehicle according to the second delivery map, the real-time path status information, and the real-time delivery order further includes:

[0140] For each delivery task, based on the initial battery level of the driverless vehicle, the real-time path status information, the delivery path corresponding to the delivery task, and the loading capacity, predict the remaining battery level of the driverless vehicle when it completes the delivery task. According to the remaining battery level of the driverless vehicle, determine whether to charge. If so, generate a charging path based on the charging station configuration plan and the real-time status information of each charging station (for example, the number of remaining idle charging piles, the remaining charging time of the charging piles in working state, etc.). After the driverless vehicle completes the delivery task, control the driverless vehicle to charge according to the charging path.

[0141] Specifically, when the remaining battery level of the driverless vehicle is less than the remaining battery level threshold, it can be determined to charge. The charging path can be generated by a charging path generation model based on the charging station configuration plan, the real-time status information of each charging station, and the position of the driverless vehicle. Among them, the charging path generation model can be a deep learning model.

[0142] Figure 4 It is a schematic diagram of the modules of the campus express delivery management system based on driverless vehicles shown in some embodiments of this specification. As Figure 4 shown, the campus express delivery management system based on driverless vehicles may include a data acquisition module, a map building module, a scheme determination module, a path monitoring module, an order acquisition module, and a delivery management module.

[0143] The data acquisition module is used to acquire the positions of the express stations, the positions of the delivery points, and the historical express delivery demands of the delivery points.

[0144] The map building module is used to build a first delivery map based on the positions of the express stations and the delivery points. Among them, the first delivery map includes express station nodes, delivery point nodes, and path nodes from any express station node to any delivery point node.

[0145] The scheme determination module is used to determine the optimal driverless vehicle configuration plan according to the first delivery map and the historical express delivery demands of the delivery points.

[0146] The scheme determination module is also used to determine the charging station configuration plan according to the first delivery map, the historical express delivery demands of the delivery points, and the optimal driverless vehicle configuration plan. Among them, the charging station configuration plan at least includes the position of each charging station.

[0147] The map building module is also used to build a second delivery map according to the position of each charging station and the first delivery map. Among them, the second delivery map includes express station nodes, delivery point nodes, charging station nodes, path nodes from any express station node to any delivery point node, and path nodes from any delivery point node to any charging station node.

[0148] The scheme determination module is also used to determine the path monitoring plan according to the second delivery map.

[0149] A path monitoring module, configured to obtain real-time path status information according to a path monitoring scheme;

[0150] An order acquisition module, configured to obtain real-time delivery orders;

[0151] A delivery management module, configured to perform express delivery management on the unmanned vehicle according to a second delivery map, real-time path status information, and real-time delivery orders.

[0152] The campus express delivery management system based on an unmanned vehicle can be used to execute the campus express delivery management method based on an unmanned vehicle, which will not be elaborated herein.

[0153] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example rather than a limitation, alternative configurations of the embodiments of this specification can be considered to be consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. A campus express delivery management method based on unmanned vehicles, characterized in that: include: Obtain the location of express delivery stations, delivery point locations, and historical express delivery demand at delivery points; Establishing a first distribution map according to the express station location and the distribution point location, wherein the first distribution map includes express station nodes, distribution point nodes, and path nodes from any express station node to any distribution point node; Determining an optimal unmanned vehicle configuration plan based on the first delivery map and historical express delivery demand of the delivery point; Determine a charging station configuration plan according to the first delivery map, the historical express delivery demand of the delivery point and the optimal unmanned vehicle configuration plan, wherein the charging station configuration plan at least includes the location and charging configuration of each charging station; Establishing a second distribution map according to the location of each charging station and the first distribution map, wherein the second distribution map includes express station nodes, distribution point nodes, charging station nodes, path nodes from any express station node to any distribution point node, and path nodes from any distribution point node to any charging station node; Determining a route monitoring plan according to the second distribution map; According to the path monitoring scheme, obtaining real-time path status information; Get real-time delivery orders; The unmanned vehicle is managed for express delivery based on the second delivery map, real-time path status information and real-time delivery orders.

2. The campus express delivery management method based on unmanned vehicles according to claim 1 is characterized in that: According to the first delivery map and the historical express delivery demand of the delivery point, the optimal unmanned vehicle configuration scheme is determined, including: For any two delivery points, according to the first delivery map, determine the single-node maximum path similarity of the two delivery points corresponding to each courier station node, and calculate the global path similarity of the two delivery points according to the single-node maximum path similarity of the two delivery points corresponding to each courier station node; For any two delivery points, the express delivery demand correlation parameters of the two delivery points are calculated based on the historical express delivery demand of the two delivery points; According to the global path similarity of any two delivery points and the associated parameters of express delivery demand, multiple delivery points are divided into multiple delivery point units; For each delivery point unit, the optimal unmanned vehicle configuration plan is determined based on the historical express delivery demand of the delivery points included in the delivery point unit and the first delivery map, wherein the optimal unmanned vehicle configuration plan includes the optimal number of unmanned vehicles corresponding to each delivery point unit.

3. The campus express delivery management method based on unmanned vehicles according to claim 2 is characterized in that: Determine the optimal number of unmanned vehicles corresponding to the distribution point unit, including: For each delivery point included in the delivery point unit, determining an initial current express delivery demand of the delivery point according to a historical express delivery demand of the delivery point; According to the express delivery demand association parameters of any two delivery points, the initial current express delivery demand of each delivery point included in the delivery point unit is modified to determine the current express delivery demand of each delivery point included in the delivery point unit; Determining the unit express delivery demand corresponding to the delivery point unit according to the current express delivery demand of each delivery point included in the delivery point unit; Determine the central location corresponding to the delivery point unit according to the current express delivery demand of each delivery point included in the delivery point unit and the location of the delivery point included in the delivery point unit; Establishing a set of constraint conditions, wherein the set of constraint conditions at least includes a transport speed constraint and a loading constraint; The optimal number of unmanned vehicles corresponding to the delivery point unit is determined according to the central position corresponding to the delivery point unit, the first delivery map, the unit express delivery demand corresponding to the delivery point unit and the constraint condition set.

4. The campus express delivery management method based on unmanned vehicles according to claim 3 is characterized in that: The constraint condition set also includes battery life constraint and charging efficiency constraint; Determining a charging station configuration plan according to the first distribution map, the historical express delivery demand of the distribution point and the optimal unmanned vehicle configuration plan, including: For each delivery point unit, determine the charging demand corresponding to the delivery point unit according to the central location corresponding to the delivery point unit, the first delivery map, the unit express delivery demand corresponding to the delivery point unit, the optimal number of unmanned vehicles corresponding to the delivery point unit, and the battery life constraint; Merging multiple delivery point units according to the central location and charging requirements corresponding to each of the delivery point units to determine multiple delivery point groups; For each delivery point group, the location and charging configuration of the charging station corresponding to the delivery point group are determined according to the central location and charging demand corresponding to each of the delivery point units included in the delivery point group.

5. The campus express delivery management method based on unmanned vehicles according to any one of claims 1 to 4, characterized in that: Determine a path monitoring plan according to the second distribution map, including: Based on the second delivery map, determining a plurality of candidate monitoring sections; Obtaining historical traffic data for each candidate monitoring section; For any two candidate monitoring sections, the flow correlation parameters of the two candidate monitoring sections are calculated according to the historical flow data of the two candidate monitoring sections; Determine the connectivity relationship between any two candidate monitoring sections; By using a clustering algorithm, the plurality of candidate monitoring sections are divided into a plurality of section clusters according to flow association parameters of any two candidate monitoring sections and a connectivity relationship between any two candidate monitoring sections; A path monitoring scheme is determined according to the multiple road segment clusters, wherein the path monitoring scheme includes multiple path monitoring points, and one path monitoring point corresponds to one road segment cluster.

6. The campus express delivery management method based on unmanned vehicles according to claim 5 is characterized in that: By using a clustering algorithm, according to the flow association parameters of any two candidate monitoring sections and the connectivity relationship between any two candidate monitoring sections, the plurality of candidate monitoring sections are divided into a plurality of section clusters, including: S11, initializing the number of road segment clusters to K, where K is a positive integer; S12, determining K centroids from multiple candidate monitoring road sections according to the number of road section clusters; S13, for each candidate monitoring road section, assigning the candidate monitoring road section to a road section cluster where the centroid of which has the largest flow correlation parameter and whose connectivity relationship satisfies a preset connectivity relationship condition is located; S14, determine whether each candidate monitoring section has been allocated, if so, execute S15, if not, execute S13; S15, for each road segment cluster, calculating intra-cluster features according to flow correlation parameters of any two candidate monitoring road segments included in the road segment cluster; S16, determining whether the intra-cluster features of each road segment cluster meet the preset intra-cluster feature requirements, if so, executing S18, if not, executing S17; S17, taking a road section cluster whose intra-cluster features do not meet the preset intra-cluster feature requirements as an abnormal road section cluster, updating the centroid of the abnormal road section cluster according to the flow correlation parameters of any two candidate monitoring road sections included in the abnormal road section cluster, and executing S13; S18, calculating global clustering features according to the flow association parameters of the K candidate monitoring sections corresponding to the K centroids; S19, determining whether the global clustering feature meets the preset global clustering feature requirements, if so, completing clustering, if not, updating the number K of road segment clusters according to the global clustering feature, and executing S12.

7. The campus express delivery management method based on unmanned vehicles according to claim 5 is characterized in that: According to the second delivery map, the real-time route status information and the real-time delivery order, the unmanned vehicle is managed for express delivery, including: According to the real-time delivery order, determine the real-time delivery order corresponding to each delivery point unit; For each delivery point unit, a delivery task corresponding to the delivery point unit is generated according to the loading constraints and the real-time delivery order corresponding to the delivery point unit, where one delivery task corresponds to one unmanned vehicle; For each of the delivery tasks, a delivery path corresponding to the delivery task is generated according to the second delivery map and the real-time path status information, and according to the delivery path corresponding to the delivery task, the unmanned vehicle corresponding to the delivery task is controlled to perform express delivery.

8. The campus express delivery management method based on unmanned vehicles according to claim 7 is characterized in that: Generating a delivery path corresponding to the delivery task according to the second delivery map and the real-time path status information includes: The path generation model is based on the second distribution map and real-time path status information, wherein the path optimization model is a deep learning model, and the reward function of the path optimization model is related to the path length and the flow of the path monitoring points.

9. The campus express delivery management method based on unmanned vehicles according to claim 8 is characterized in that: According to the second delivery map, the real-time route status information and the real-time delivery order, the unmanned vehicle is managed for express delivery, further comprising: For each of the delivery tasks, the remaining power of the unmanned vehicle when completing the delivery task is predicted based on the initial power of the unmanned vehicle, the real-time path status information, and the delivery path and load capacity corresponding to the delivery task. Based on the remaining power of the unmanned vehicle, it is determined whether to charge. If so, a charging path is generated based on the charging station configuration plan and the real-time status information of each charging station. After the unmanned vehicle completes the delivery task, the unmanned vehicle is controlled to charge according to the charging path.

10. The campus express delivery management system based on unmanned vehicles is characterized by: The campus express delivery management method based on an unmanned vehicle according to any one of claims 1 to 9 is applied, comprising: The data acquisition module is used to obtain the locations of express delivery stations, delivery points and the historical express delivery needs of the delivery points; A map building module, used to build a first delivery map according to the express station location and the delivery point location, wherein the first delivery map includes express station nodes, delivery point nodes and path nodes from any express station node to any delivery point node; A solution determination module, used to determine the optimal unmanned vehicle configuration solution based on the first delivery map and the historical express delivery demand of the delivery point; The solution determination module is further used to determine a charging station configuration solution according to the first delivery map, the historical express delivery demand of the delivery point and the optimal unmanned vehicle configuration solution, wherein the charging station configuration solution at least includes the location of each charging station; The map establishment module is further used to establish a second distribution map according to the location of each charging station and the first distribution map, wherein the second distribution map includes express station nodes, distribution point nodes, charging station nodes, path nodes from any express station node to any distribution point node, and path nodes from any distribution point node to any charging station node; The solution determination module is further used to determine a path monitoring solution according to the second delivery map; A path monitoring module, used to obtain real-time path status information according to the path monitoring scheme; Order acquisition module, used to obtain real-time delivery orders; The delivery management module is used to manage the express delivery of the unmanned vehicle based on the second delivery map, real-time path status information and real-time delivery orders.

Images