Intelligent scheduling system and method for passing of unmanned logistics vehicles in park

Abstract

The invention discloses an intelligent scheduling system and method for the passage of unmanned logistics vehicles in a park, and particularly relates to the technical field of intelligent scheduling. All scheduling data are stored through a decentralized block chain network, each unmanned logistics vehicle is connected with a block chain, whether information circulation is necessary is judged according to the use scene of the unmanned logistics vehicle, and if yes, the unmanned logistics vehicle passes through the block chain; activating an information interaction mechanism, screening out vehicles needing to perform information interaction with the target unmanned logistics vehicle through interaction analysis, and obtaining a corresponding database combination from the block chain; based on the database combination, using a pre-trained neural network model to extract restriction features, and generating a restriction condition set; using a genetic algorithm to optimize the scheduling path of the target vehicle based on a limiting condition, and updating the scheduling path to the block chain network; whether the target vehicle and the surrounding vehicles have the necessity of information circulation can be accurately judged, an information interaction mechanism is started when necessary, and effective coordination work among the vehicles can be ensured.

Description

Technical Field

[0001] The present invention relates to the field of intelligent scheduling technology, and more specifically, to an intelligent scheduling system and method for the passage of unmanned logistics vehicles in a park. Background Art

[0002] With the development of unmanned driving technology and intelligent dispatching systems, unmanned logistics vehicles in the park have played an increasingly important role in logistics transportation, distribution and other tasks. The current unmanned logistics vehicle dispatching system mainly relies on centralized servers for dispatching and management. Although this model can meet basic dispatching needs, it is often inefficient and prone to single point failures when dealing with complex scenarios, such as collaborative work between vehicles, avoidance of path conflicts, and resource scheduling. In addition, during the vehicle dispatching process, there are certain challenges in the real-time and security of information sharing and data flow. If there is a lack of effective information interaction mechanisms and collaborative work methods between vehicles, path conflicts, waste of resources, and inefficient task execution may occur.

[0003] In order to deal with these problems, decentralized blockchain technology has been introduced into the dispatching system of unmanned logistics vehicles. Blockchain technology has the characteristics of data immutability, transparent information sharing and decentralization, but it also brings unnecessary information transmission. How to reduce communication overhead and efficiently plan routes in complex and changing dispatching scenarios is an urgent problem to be solved. Summary of the invention

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] An intelligent dispatching method for unmanned logistics vehicles in a park, comprising the following steps:

[0006] All dispatch data are stored in a decentralized blockchain network, and each unmanned logistics vehicle is connected to the blockchain network to determine the usage scenario of each unmanned logistics vehicle and whether there is a need for information flow for the target unmanned logistics vehicle;

[0007] When the target unmanned logistics vehicle needs information flow, the information interaction mechanism is activated, and the unmanned logistics vehicle that needs to interact with the target unmanned logistics vehicle is screened out through interaction analysis, and then the database combination corresponding to the information interaction is obtained from the blockchain network;

[0008] Based on the database combination, a pre-trained neural network model is used to extract restricted features and obtain a set of restricted conditions;

[0009] Using genetic algorithms, the dispatch path of the target unmanned logistics vehicle is optimized and planned based on a set of constraints, and the optimized dispatch path is updated in the blockchain network.

[0010] In a preferred embodiment, the decentralized blockchain network contains the individual data sets of all unmanned logistics vehicles, as well as a public data set.

[0011] In a preferred embodiment, the usage scenarios of each unmanned logistics vehicle are judged, and a conflict coefficient of the target unmanned logistics vehicle is generated. The conflict coefficient is compared with a preset conflict threshold. If the conflict coefficient is greater than the preset conflict threshold, it is judged that there is a need for information flow with the target unmanned logistics vehicle. If the conflict coefficient is less than or equal to the preset conflict threshold, it is judged that there is no need for information flow with the target unmanned logistics vehicle.

[0012] In a preferred embodiment, the logic for obtaining the conflict coefficient is:

[0013] Obtain the preset path vector and time vector of the target vehicle, mark the path vector as Ptarget, mark the time vector as Ttarget, mark the preset path vector corresponding to each other vehicle around as Pi, and mark the time vector as Ti, where i represents the vehicle number around the target vehicle;

[0014] For each surrounding vehicle i, calculate the conflict value Ci between it and the target vehicle:

[0015] Ci represents the conflict value between the surrounding vehicle i and the target vehicle, and then the maximum value of the conflict value Ci is taken as the conflict coefficient.

[0016] The database combination is composed of a restricted database consisting of the own databases of all unmanned logistics vehicles with which the target unmanned logistics vehicle interacts with information, and a public database.

[0017] In a preferred embodiment, the unmanned logistics vehicle that needs to interact with the target unmanned logistics vehicle through interactive analysis is selected as follows:

[0018] Obtain the frequency of the target unmanned logistics vehicle and the surrounding vehicle i in the past collaborative tasks and mark it as Hi, and the absolute value of the difference in task priority between the target vehicle and vehicle i and mark it as Yi, and then substitute it into the formula:

[0019]

[0020] α and β satisfy:

[0021] Poverlap i Toverlap represents the path overlap length between the target unmanned logistics vehicle and the surrounding vehicle i. irepresents the time overlap interval between the target unmanned logistics vehicle and the surrounding vehicle i, Ptotal represents the path length of the target unmanned logistics vehicle, Ttotal represents the time length of the target unmanned logistics vehicle, CNR represents the cooperation necessity index, w1 and w2 are both preset proportional coefficients and are not zero;

[0022] When the collaboration necessity index CNR is greater than the preset information flow threshold, there is a need for information flow between the target unmanned logistics vehicle and the surrounding vehicles i.

[0023] In a preferred embodiment, the neural network model is a convolutional neural network model.

[0024] In a preferred embodiment, using a genetic algorithm to optimize the scheduling path of a target unmanned logistics vehicle based on a set of constraints and updating the optimized scheduling path in the blockchain network refers to:

[0025] Initialize the population: The set of constraints extracted from the databases of all unmanned logistics vehicles with which the target unmanned logistics vehicle interacts and the public database is used as the initial scheduling conditions, and the initial scheduling path plan is randomly generated as the initial population for path planning;

[0026] Fitness evaluation: The set of restriction conditions is used as the constraint conditions in fitness evaluation, and the fitness evaluation formula is established at the same time: ZYj represents the utilization efficiency of resource type j of the target unmanned logistics vehicle, m represents the total amount of resource types of the target unmanned logistics vehicle, fj represents the preset influence coefficient corresponding to resource type j of the target unmanned logistics vehicle, and F is the fitness value;

[0027] Selection operation: Use the roulette wheel selection method to select offspring as new parents;

[0028] Crossover operation: randomly exchange data in chromosomes of different parents;

[0029] Mutation operation: randomly select data of different offspring chromosomes for adjustment;

[0030] Iteration and termination conditions: When the pre-set termination conditions are reached, the chromosome with the highest fitness is selected from the final population for decoding to obtain the optimal scheduling path solution.

[0031] In a preferred embodiment, an intelligent dispatching system for unmanned logistics vehicles in a park includes:

[0032] Storage module, which stores all scheduling data in a decentralized blockchain network;

[0033] The scenario judgment module connects each unmanned logistics vehicle to the blockchain network, judges the usage scenario of each unmanned logistics vehicle, and determines whether there is a need for information flow for the target unmanned logistics vehicle;

[0034] Interaction screening module: when there is a need for information flow between the target unmanned logistics vehicles, the information interaction mechanism is activated, and the unmanned logistics vehicles that need to interact with the target unmanned logistics vehicles are screened out through interaction analysis;

[0035] The feature extraction module obtains the database combination corresponding to the information interaction from the blockchain network, and based on the database combination, uses the pre-trained neural network model to extract restricted features to obtain a set of restriction conditions;

[0036] The scheduling module uses a genetic algorithm to optimize the scheduling path of the target unmanned logistics vehicle based on a set of constraints and updates the optimized scheduling path in the blockchain network.

[0037] Technical effects and advantages of the present invention:

[0038] The present invention introduces a decentralized blockchain network and stores all dispatch data in the blockchain, achieving transparent sharing and immutability of data, and avoiding the single point failure problem in the traditional centralized system. Through the application of blockchain technology, the dispatch system can process and share information in real time during multi-vehicle collaborative work and path planning, improving the security of the system and the accuracy of dispatch.

[0039] The present invention combines neural networks and genetic algorithms to efficiently dispatch and optimize the paths of unmanned logistics vehicles in complex and changeable park environments. Neural networks are used to extract constraints from multidimensional data, and genetic algorithms are used to intelligently plan paths based on constraints, so that the dispatching process can quickly adapt to different task requirements and effectively avoid potential path conflicts and resource waste.

[0040] The present invention uses an intelligent dispatching system to accurately dispatch vehicles according to their resource status, reduce excessive energy consumption, and improve resource utilization. This is particularly important in multi-vehicle collaborative tasks, ensuring that each vehicle's task execution can be completed efficiently while maximizing the use of available infrastructure resources.

[0041] Through the collaboration necessity index calculation method of the present invention, the system can accurately determine whether there is a necessity for information flow between the target vehicle and surrounding vehicles, and start the information interaction mechanism when necessary to ensure that the vehicles can effectively coordinate their work and avoid task conflicts or resource competition, which greatly improves the efficiency of multi-vehicle collaborative work.

[0042] The dispatching system of the present invention can flexibly respond to dispatching problems under the circumstances of complex tasks and changing park environments, provide personalized path planning and optimization solutions, and adapt to the dispatching needs of unmanned logistics vehicles in a variety of complex scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0044] Figure 1 This is a schematic diagram of an intelligent scheduling method for the passage of unmanned logistics vehicles in a park in the present invention.

[0045] Figure 2 This is a schematic diagram of an intelligent dispatching system for the passage of unmanned logistics vehicles in a park in the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Reference Figure 1 - Figure 2 The following embodiments are obtained:

[0048] Embodiment 1: An intelligent dispatching method for unmanned logistics vehicles in a park, comprising the following steps:

[0049] All dispatch data is stored in a decentralized blockchain network, and each unmanned logistics vehicle is connected to the blockchain network. The use scenarios of each unmanned logistics vehicle are judged to determine whether the target unmanned logistics vehicle needs information flow. The decentralized blockchain network ensures the security, transparency and immutability of dispatch data. All vehicles and dispatch systems can access the same data source in real time to ensure information consistency and prevent single point failures, thereby improving the reliability and efficiency of the dispatch system.

[0050] When there is a need for information flow between the target unmanned logistics vehicle, the information interaction mechanism is activated, and the unmanned logistics vehicle that needs to interact with the target unmanned logistics vehicle is screened out through interactive analysis, and then the database combination corresponding to the information interaction is obtained from the blockchain network; in each scheduling scenario, judging whether the target vehicle needs to interact with other vehicles can avoid unnecessary information transmission and reduce communication overhead. If the risk of conflict is high, information flow is necessary to ensure that the coordinated scheduling between vehicles is more efficient. Through analysis and screening, vehicles that may have path or time conflicts with the target vehicle interact with information. This information interaction mechanism helps improve the collaborative ability of vehicles, avoid route conflicts, resource conflicts, etc., and improve the accuracy and efficiency of vehicle scheduling. Extract the data set of the target vehicle and its interacting vehicles from the blockchain network to ensure that the scheduling algorithm can be calculated based on the latest, real, and reliable data to avoid information lag or errors.

[0051] Based on the database combination, a pre-trained neural network model is used to extract restriction features and obtain a set of restriction conditions. By extracting a set of restriction conditions from a complex database combination, the convolutional neural network (CNN) can more effectively extract the multidimensional features that affect vehicle scheduling, ensuring that the scheduling algorithm takes into account various possible restriction conditions, such as traffic conditions, equipment status, etc., making the scheduling plan more accurate.

[0052] Using genetic algorithms, the dispatch path of the target unmanned logistics vehicle is optimized and planned based on a set of constraints, and the optimized dispatch path is updated in the blockchain network. Genetic algorithms can select the optimal dispatch path from a large number of possible path solutions by simulating the natural selection process. Fitness evaluation is comprehensively optimized based on conditions such as the resource utilization efficiency of the target vehicle to ensure that the selected path can maximize resource utilization and reduce conflicts. Resource utilization efficiency, as an important indicator of fitness evaluation, can measure how the vehicle reasonably uses its energy, equipment and other resources during driving.

[0053] The decentralized blockchain network contains the data sets of all unmanned logistics vehicles and the public data sets. The decentralized blockchain network ensures data security through distributed storage. All data of unmanned logistics vehicles (such as status information, task progress, etc.) are stored in the blockchain to prevent data tampering or forgery, and improve the security and transparency of the system. Because the blockchain has the characteristics of being tamper-proof, all scheduling-related data sets (whether individual vehicle data or public data) can be shared safely and reliably.

[0054] Through blockchain, all vehicles and dispatch centers can access public data sets at any time, ensuring that each node can obtain global information (such as traffic conditions, road closures, environmental data, etc.), so as to use the most accurate and up-to-date information in dispatch decisions. The sharing of public data sets ensures that each node in the vehicle dispatch system uses consistent information to make decisions, avoiding information asymmetry.

[0055] Although individual vehicle data sets of unmanned logistics vehicles are stored on the blockchain, the decentralized nature allows the privacy of each vehicle to be protected through permission management. Each vehicle can choose to share its own data with other vehicles in specific circumstances, such as when there is a necessary scheduling interaction, without having to share all information in irrelevant situations. This protects privacy and enables data sharing when necessary. Since the blockchain network is decentralized, all vehicle data is backed up on multiple nodes, and the system will not cause data loss or inaccessibility due to a single point of failure. Even if some nodes or vehicles cannot connect to the network due to a failure, other nodes can still work normally, ensuring the stability and fault resistance of the scheduling system.

[0056] Data in a decentralized network is updated and synchronized in real time, which means that the status and public data of each vehicle can be quickly updated to all nodes in the network, ensuring that scheduling decisions are based on the latest data. During the scheduling process, vehicles can adjust their path planning at any time based on the latest status and environmental data to improve the real-time and responsiveness of scheduling. In some scenarios, multiple unmanned logistics vehicles need to work together, and a decentralized blockchain network can efficiently exchange data. By sharing individual vehicle data and public data, collaborative vehicles can quickly obtain each other's status and constraints, and make intelligent decisions based on global information to avoid route conflicts or resource competition.

[0057] Moreover, all vehicle and dispatch operation data are stored in the blockchain and can be traced and verified. During the dispatch process, if a problem or conflict occurs, the transparency of the blockchain records allows the dispatch center or manager to view the history of the data and trace the root cause of the problem. This improves the system's responsibility management and reduces the risk of dispatch errors or failures.

[0058] The usage scenarios of each unmanned logistics vehicle are judged, and the conflict coefficient of the target unmanned logistics vehicle is generated. The conflict coefficient is compared with the preset conflict threshold. If the conflict coefficient is greater than the preset conflict threshold, it is judged that there is a need for information flow of the target unmanned logistics vehicle. If the conflict coefficient is less than or equal to the preset conflict threshold, it is judged that there is no need for information flow of the target unmanned logistics vehicle. This mechanism can be understood as a smart contract. A smart contract is an automated protocol running on the blockchain that can automatically execute operations according to preset conditions.

[0059] Pre-condition: The comparison between the conflict coefficient and the threshold is used as the criterion for determining whether information transfer is required. This comparison logic can be used as a trigger condition in the smart contract.

[0060] Automatic execution: When the conflict coefficient is greater than the threshold, the system automatically determines and activates the necessity of information flow; when it is less than or equal to the threshold, no information flow is performed. This is a typical automatic execution logic in smart contracts.

[0061] No human intervention required: Smart contracts will automatically execute corresponding tasks when certain conditions are met. In this scenario, the system can automatically determine whether the target vehicle needs to exchange data with other vehicles by calculating the conflict coefficient without human intervention.

[0062] Decentralized execution: Because the data is stored in a decentralized blockchain network, this judgment mechanism can also be automatically executed on the blockchain through smart contracts without relying on the intervention of a central server or dispatch center.

[0063] Contract content: The triggering condition of the smart contract can be designed as follows: compare the conflict coefficient with the preset conflict threshold. If the conflict coefficient is greater than the preset conflict threshold, it is judged that there is a need for information flow of the target unmanned logistics vehicle; if the conflict coefficient is less than or equal to the preset conflict threshold, it is judged that there is no need for information flow of the target unmanned logistics vehicle.

[0064] The logic for obtaining the conflict coefficient is:

[0065] Obtain the preset path vector and time vector of the target vehicle, mark the path vector as Ptarget, mark the time vector as Ttarget, mark the preset path vector corresponding to each other vehicle around as Pi, and mark the time vector as Ti, where i represents the vehicle number around the target vehicle;

[0066] For each surrounding vehicle i, calculate the conflict value Ci between it and the target vehicle:

[0067] Ci represents the conflict value between the surrounding vehicle i and the target vehicle, and then the maximum value of the conflict value Ci is taken as the conflict coefficient.

[0068] The conflict coefficient C is an indicator that measures the potential spatial and temporal conflicts between the target vehicle and its surrounding vehicles. It is calculated comprehensively using the path vector (i.e., the vehicle's driving path) and the time vector (i.e., the vehicle's expected time). The calculation of the conflict coefficient reflects the overlap between the target vehicle and the surrounding vehicles in space and time. The larger the value, the higher the possibility of conflict. Through the vectors of the two dimensions of path and time, the system can evaluate the conflicts between vehicles in space (path) and time (time window). If the time overlaps between two vehicles on the same path to a large extent, the risk of conflict will increase accordingly. Formula Ci represents the calculation of the conflict value between the target vehicle and the surrounding vehicle i, combining the overlap between the paths and time between the vehicles. The formula calculates the dot product of the path vector and the time vector to determine how much overlap there is between the two vehicles in path and time, and at the same time, the denominator is normalized to avoid large numerical differences. The maximum value of all calculated conflict values ​​Ci is used as the conflict coefficient of the target vehicle. The maximum conflict value reflects the most serious conflict between the target vehicle and all surrounding vehicles. If the conflict value of a vehicle is very high, it means that the path and time overlap between the vehicle and the target vehicle is large, and special attention should be paid and avoidance measures may be taken. By calculating the conflict coefficient, the system can automatically determine whether the target vehicle needs to interact or collaborate with other vehicles. If the conflict coefficient is higher than the preset threshold, the system will trigger the information flow mechanism to optimize the path and time planning by exchanging data with other vehicles.

[0069] It should be noted that the definition of surrounding vehicles i is usually based on a comprehensive judgment of multiple dimensions including the spatial position, time, task relevance and communication range of the target vehicle. The specific definition method can be based on the following situations:

[0070] Based on distance: The surrounding vehicles i can be defined according to the physical distance from the target vehicle. Usually a perception range radius (such as 50 meters or 100 meters) is set, and all vehicles within this range are considered to be surrounding vehicles.

[0071] Based on regional division: the park can be divided into several areas, and all vehicles in the same area as the target vehicle are considered as surrounding vehicles. This can avoid too many vehicles being included in the calculation.

[0072] Time overlap: If the target vehicle and other vehicles have a large overlap in their driving time periods, these vehicles can be considered as surrounding vehicles. Even if the physical distance is far, if the two vehicles will meet in the same area at a certain point in time, these vehicles need to be included in the category of surrounding vehicles. Set a time window, such as vehicles that may meet within 10 minutes, and vehicles outside this time window are not considered surrounding vehicles.

[0073] Based on task association: If the target vehicle and other vehicles are performing the same task (such as collaboratively transporting goods or completing a task together), these vehicles will be regarded as surrounding vehicles even if there is a slight gap in space or time.

[0074] Mission priority: Certain mission-critical vehicles may require special attention during scheduling. Even if they are not within the perception range of the physical space, they can be considered as surrounding vehicles for early conflict detection and collaborative planning.

[0075] Based on the communication network: If the target vehicle can communicate directly with other vehicles (such as through the vehicle communication system or a dedicated wireless network within the park), then all vehicles within the communication range can be defined as surrounding vehicles.

[0076] Based on dynamic perception: Through on-board sensors and environmental perception systems, the target vehicle can detect in real time whether there are other vehicles around it. Based on these sensor data, it can determine in real time which vehicles are around it. For example, using laser radar, cameras, etc. to detect nearby vehicles in real time, so that the list of surrounding vehicles can be dynamically updated.

[0077] Based on road topology: In some complex campus scenarios, the definition of surrounding vehicles can be based on the structure of the road network. If the target vehicle and other vehicles will meet at the same intersection or junction, or are on the same road network segment, they are considered surrounding vehicles.

[0078] Based on preset rules: There may be specific rules in the park that stipulate that certain types of vehicles or vehicles for certain tasks are always considered as surrounding vehicles of each other. For example, loading and unloading vehicles and freight vehicles near the logistics center are predefined as surrounding vehicles by the system even if they are not within the physical range.

[0079] The database combination is composed of a restricted database consisting of the own databases of all unmanned logistics vehicles with which the target unmanned logistics vehicle interacts with information, and a public database. The database combination consists of two parts: Restricted database: The own databases of all unmanned logistics vehicles with which the target unmanned logistics vehicle interacts with information. Each vehicle has its own status information, task information, resource usage, etc., which contains various dynamic data of each vehicle when performing scheduling tasks. Public database: Contains global information that can be accessed by all vehicles, such as road conditions in the park, real-time traffic information, environmental changes, available resources (such as the use of charging stations or parking spaces), etc. This information can provide real-time reference for all vehicles.

[0080] Restricted database content (from the target vehicle and other vehicles that interact with it): Vehicle status: such as location, speed, remaining power, and current task progress. Task information: including the target location, priority, and task time window of the transportation task. Resource usage information: such as the current resource consumption status, the remaining capacity of the battery, and whether charging or maintenance is required. Historical data: the vehicle's previous dispatch path, the success rate of task execution, and the collaborative records with other vehicles.

[0081] Public database content: Road conditions: such as current traffic flow, congestion, road closures, etc. Environmental information: weather conditions, temperature, humidity and other environmental factors that affect vehicle performance. Resource availability: such as the location and availability of charging piles in the park, the number and distribution of parking spaces, etc. Infrastructure data: including road networks, restricted areas and accessible areas in the park.

[0082] Step 1: Construction of database combination. When the target unmanned logistics vehicle needs to plan or adjust the path, the system first identifies other unmanned logistics vehicles related to it through the information flow mechanism. The system combines the restriction databases of these vehicles (vehicle status, tasks, resource information, etc.) and the relevant information in the public database into a comprehensive database, which contains all information related to the scheduling of the target vehicle. The database combination ensures that the target vehicle fully considers the restrictions of itself, the surrounding collaborative vehicles and the environment when planning the path or task, forming a complete global view.

[0083] Step 2: The neural network model is a convolutional neural network model. A pre-trained convolutional neural network model is used. The convolutional neural network (CNN) model has been trained with historical data and has the ability to extract complex constraints. When the database combination is built, the data is input into the CNN model. Through layer-by-layer convolution and pooling operations, CNN can extract key constraints from the data, mainly including: risk assessment of vehicle task and path conflicts; assessment of the impact of environmental impacts (such as weather changes) on scheduling; and dependency analysis of multi-vehicle collaborative tasks. The CNN model can extract key constraints from complex multi-dimensional data, helping the scheduling system to quickly determine which factors the target vehicle needs to prioritize when planning the path, thereby improving scheduling efficiency and decision-making accuracy.

[0084] Step 3: Generate a set of constraints. Through the analysis of the convolutional neural network, the system will finally extract the key information into a "constraint set", which covers all the constraints that affect the scheduling of the target vehicle. Specifically, it includes: current road accessibility and path risk, task urgency and priority, availability of vehicle energy resources, and scheduling requirements of other collaborative vehicles. The constraint set is the core basis for optimizing the scheduling path. According to this set, the system ensures that potential risks can be avoided when scheduling vehicles, emergency tasks can be performed first, and resources can be used effectively to ensure the smooth progress of collaborative work. The training process of the convolutional neural network will not be repeated here. The convolutional neural network CNN can extract deep features from multidimensional data from different sources. The convolutional neural network can integrate data from different sources (such as vehicle status, task requirements, environmental changes, etc.) through a hierarchical structure to form a unified constraint. This ability makes CNN very adaptable in complex scenarios. Since there may be a large number of vehicles being scheduled at the same time in the park, the convolutional neural network can efficiently process a large amount of data, quickly extract key scheduling factors from it, and reduce manual intervention and calculation delays. Therefore, the pre-trained neural network model is selected to extract restricted features to obtain a set of constraints.

[0085] The unmanned logistics vehicles that need to interact with the target unmanned logistics vehicles through interactive analysis are:

[0086] Obtain the frequency of the target unmanned logistics vehicle and the surrounding vehicle i in the past collaborative tasks and mark it as Hi, and the absolute value of the difference in task priority between the target vehicle and vehicle i and mark it as Yi, and then substitute it into the formula:

[0087]

[0088] α and β satisfy:

[0089] Poverlap i Toverlap represents the overlap length of the paths of the target unmanned logistics vehicle and the surrounding vehicle i, that is, the part where the two vehicles travel on the same path. i represents the time overlap interval of the target unmanned logistics vehicle and the surrounding vehicle i, that is, the time that the two vehicles overlap on the same road section, Ptotal represents the path length of the target unmanned logistics vehicle, Ttotal represents the time length of the target unmanned logistics vehicle, CNR represents the cooperation necessity index, w1 and w2 are both preset proportional coefficients and are not zero, which are used to balance the impact of the conflict coefficient and the task history cooperation frequency on the cooperation necessity; Ci is the conflict value that measures the potential conflict between the target vehicle and the surrounding vehicles;

[0090] When the collaboration necessity index CNR is greater than the preset information flow threshold, there is a need for information flow between the target unmanned logistics vehicle and the surrounding vehicles i.

[0091] Two adaptive adjustment coefficients α and β are introduced into the formula to dynamically adjust their impact on the necessity of collaboration according to the overlap of time and path. If the time overlap of two vehicles is high, it means that they may use the same path at the same time. At this time, α will be larger, which increases the influence weight of time overlap; if the path overlap is higher, β will be larger, which increases the influence of path overlap. Through dynamic adjustment, the weight of the influence of time and path on the necessity of collaboration can be reasonably allocated to avoid a single factor dominating the decision. Adding conflict value can ensure that the system triggers the collaboration mechanism when there is a greater risk of conflict, avoiding potential conflict problems in path planning. Hi represents the collaboration history between vehicles. The higher the frequency, the richer the successful experience of collaboration, and the more necessary it is to continue the information flow. Yi represents the difference in task priority. The greater the difference, the lower the urgency of collaboration. The historical collaboration frequency and priority difference directly affect the necessity of collaboration, avoid high-priority tasks from being interfered with by low-priority tasks, and more efficiently handle vehicles that need close collaboration. The collaboration necessity index (CNR) is to determine whether information flow is necessary. By calculating CNR, the system can dynamically evaluate whether it is necessary for the target vehicle to interact with certain surrounding vehicles. If the CNR exceeds the set threshold, the system triggers the information flow mechanism to ensure smooth collaboration between vehicles. Dynamically balance time and path overlap: Through the adaptive adjustment of α and β, the system can dynamically balance the impact of time and path overlap according to actual conditions, avoid excessive influence of a single factor on decision-making, and improve the accuracy of judgment on the necessity of collaboration. The formula combines multiple factors such as conflict risk, historical collaboration experience, and task priority differences. It can intelligently determine which vehicles need to collaborate in complex vehicle scheduling scenarios, thereby improving scheduling efficiency, avoiding unnecessary information flow, and reducing system burden. This formula can provide accurate collaboration judgments in complex multi-vehicle collaboration scenarios, especially in high-density unmanned logistics vehicle environments such as campuses. It has high practical value.

[0092] Using genetic algorithms to optimize the scheduling path of the target unmanned logistics vehicle based on a set of constraints and updating the optimized scheduling path in the blockchain network means:

[0093] Initialize the population: The set of constraints extracted from the databases of all unmanned logistics vehicles with which the target unmanned logistics vehicle interacts and the public database is used as the initial scheduling conditions, and the initial scheduling path plan is randomly generated as the initial population for path planning; for example, assuming that there are multiple unmanned logistics vehicles transporting goods at the same time in the park, the process of initializing the population is that the system generates multiple possible initial path plans based on the current traffic conditions, vehicle power, road use restrictions, etc. For example, an unmanned logistics vehicle may have five different feasible paths, which will become candidate solutions in the initial population.

[0094] Fitness evaluation: The set of restriction conditions is used as the constraint conditions in fitness evaluation, and the fitness evaluation formula is established at the same time: ZYj represents the utilization efficiency of resource type j of the target unmanned logistics vehicle, m represents the total amount of resource types of the target unmanned logistics vehicle, fj represents the preset influence coefficient corresponding to resource type j of the target unmanned logistics vehicle, and F is the fitness value; for example: suppose that the unmanned logistics vehicle needs to consume electricity and use road resources when driving in the park. The fitness evaluation will calculate the pros and cons of the path based on the efficiency of the use of these resources. For example, if a path consumes less electricity and avoids peak traffic sections, the fitness value of this path will be higher. Assuming that a vehicle has low power consumption and avoids congestion, its fitness value F will be higher.

[0095] Selection operation: Use the roulette wheel selection method to select offspring as new parents; for example, in the process of scheduling route optimization, the path scheme with high fitness value (such as low power consumption and smooth traffic) is more likely to be selected as the parent of the next generation, and further crossover and mutation. This is similar to the process of biological evolution, where individuals with high fitness have a greater probability of passing on their genes.

[0096] Crossover operation: randomly exchange data in chromosomes of different parents; for example: suppose there are two path plans, one of which performs well in the first half of the path (such as low energy consumption), and the other is better in the second half of the path (such as avoiding traffic rush hour). Through the crossover operation, the system may generate a new path, the first half of which comes from the first path, and the second half comes from the second path.

[0097] Mutation operation: randomly select data from different offspring chromosomes for adjustment; for example, in a path plan, the system may randomly select a section of the path for adjustment, such as selecting an alternative route or changing the strategy for using certain resources. If the original path passes through a congested area, the mutation operation may randomly try to bypass this area, thereby possibly finding a better path.

[0098] Iteration and termination conditions: When the pre-set termination conditions are reached, the chromosome with the highest fitness is selected from the final population for decoding to obtain the optimal scheduling path plan. For example, suppose the system is preset to stop iteration when the fitness value reaches a certain threshold. If the fitness of a path plan in a certain iteration has met the goal (such as minimizing energy consumption and optimizing time), the system will terminate the optimization and select the path as the final scheduling plan.

[0099] In all optimization iterations, the chromosome with the highest fitness is decoded into the final scheduling path plan and updated to the blockchain network. By putting the optimized scheduling path on the chain, the data is ensured to be transparent and secure, and shared with relevant vehicles in real time when necessary. In the final selected scheduling plan, assuming that the path plan is not only optimal in resource utilization, but also avoids peak traffic and has the lowest power consumption, the system will update this path to the blockchain. Using genetic algorithms to optimize scheduling paths, the system can find the optimal solution from a large number of possible path plans through continuous selection, crossover, and mutation operations. This method can quickly adapt to complex scheduling environments and optimize the paths of unmanned logistics vehicles to achieve optimal resource utilization, time cost, etc. For example, in a park scenario, assuming that an unmanned logistics vehicle needs to deliver goods during peak hours, the genetic algorithm can generate multiple scheduling plans based on limiting factors such as vehicle traffic congestion, and find the optimal path through multiple iterations to ensure that the task can be completed smoothly and efficiently.

[0100] Embodiment 2: An intelligent dispatching system for unmanned logistics vehicles in a park, comprising:

[0101] Storage module, which stores all scheduling data in a decentralized blockchain network;

[0102] The scenario judgment module connects each unmanned logistics vehicle to the blockchain network, judges the usage scenario of each unmanned logistics vehicle, and determines whether there is a need for information flow for the target unmanned logistics vehicle;

[0103] Interaction screening module: when there is a need for information flow between the target unmanned logistics vehicles, the information interaction mechanism is activated, and the unmanned logistics vehicles that need to interact with the target unmanned logistics vehicles are screened out through interaction analysis;

[0104] The feature extraction module obtains the database combination corresponding to the information interaction from the blockchain network, and based on the database combination, uses the pre-trained neural network model to extract restricted features to obtain a set of restriction conditions;

[0105] The scheduling module uses a genetic algorithm to optimize the scheduling path of the target unmanned logistics vehicle based on a set of constraints and updates the optimized scheduling path in the blockchain network.

[0106] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0107] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0108] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0109] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0110] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An intelligent dispatching method for unmanned logistics vehicles in a park, characterized in that: The following steps are involved: All dispatch data are stored in a decentralized blockchain network, and each unmanned logistics vehicle is connected to the blockchain network to determine the usage scenario of each unmanned logistics vehicle and whether there is a need for information flow for the target unmanned logistics vehicle; When the target unmanned logistics vehicle needs information flow, the information interaction mechanism is activated, and the unmanned logistics vehicle that needs to interact with the target unmanned logistics vehicle is screened out through interaction analysis, and then the database combination corresponding to the information interaction is obtained from the blockchain network; Based on the database combination, a pre-trained neural network model is used to extract restricted features and obtain a set of restricted conditions; Using genetic algorithms, the dispatching path of the target unmanned logistics vehicle is optimized and planned based on a set of constraints, and the optimized dispatching path is updated in the blockchain network.

2. According to claim 1, the intelligent dispatching method for unmanned logistics vehicles in a park is characterized in that: The decentralized blockchain network contains the individual data sets of all unmanned logistics vehicles, as well as public data sets.

3. According to claim 2, the intelligent dispatching method for unmanned logistics vehicles in a park is characterized in that: The usage scenarios of each unmanned logistics vehicle are judged, and the conflict coefficient of the target unmanned logistics vehicle is generated. The conflict coefficient is compared with the preset conflict threshold. If the conflict coefficient is greater than the preset conflict threshold, it is judged that there is a need for information flow of the target unmanned logistics vehicle. If the conflict coefficient is less than or equal to the preset conflict threshold, it is judged that there is no need for information flow of the target unmanned logistics vehicle.

4. According to claim 3, the intelligent dispatching method for unmanned logistics vehicles in a park is characterized in that: The logic for obtaining the conflict coefficient is: Obtain the preset path vector and time vector of the target vehicle, mark the path vector as Ptarget, mark the time vector as Ttarget, mark the preset path vector corresponding to each other vehicle around as Pi, and mark the time vector as Ti, where i represents the vehicle number around the target vehicle; For each surrounding vehicle i, calculate the conflict value Ci between it and the target vehicle: Ci represents the conflict value between the surrounding vehicle i and the target vehicle, and then the maximum value of the conflict value Ci is taken as the conflict coefficient.

5. The intelligent dispatching method for unmanned logistics vehicles in a park according to claim 4 is characterized in that: The database combination is composed of a restricted database consisting of the own databases of all unmanned logistics vehicles with which the target unmanned logistics vehicle interacts with information, and a public database.

6. The intelligent dispatching method for unmanned logistics vehicles in a park according to claim 5 is characterized in that: The unmanned logistics vehicles that need to interact with the target unmanned logistics vehicles through interactive analysis are: Obtain the frequency of the target unmanned logistics vehicle and the surrounding vehicle i in the past collaborative tasks and mark it as Hi, and the absolute value of the difference in task priority between the target vehicle and vehicle i and mark it as Yi, and then substitute it into the formula: α and β satisfy: Poverlap i Toverlap represents the path overlap length between the target unmanned logistics vehicle and the surrounding vehicle i. i represents the time overlap interval between the target unmanned logistics vehicle and the surrounding vehicle i, Ptotal represents the path length of the target unmanned logistics vehicle, Ttotal represents the time length of the target unmanned logistics vehicle, CNR represents the cooperation necessity index, w1 and w2 are both preset proportional coefficients and are not zero; When the collaboration necessity index CNR is greater than the preset information flow threshold, there is a need for information flow between the target unmanned logistics vehicle and the surrounding vehicles i.

7. The intelligent dispatching method for unmanned logistics vehicles in a park according to claim 6 is characterized in that: The neural network model is a convolutional neural network model.

8. The intelligent dispatching method for unmanned logistics vehicles in a park according to claim 7 is characterized in that: Using genetic algorithms to optimize the scheduling path of the target unmanned logistics vehicle based on a set of constraints and updating the optimized scheduling path in the blockchain network means: Initialize the population: The set of constraints extracted from the databases of all unmanned logistics vehicles with which the target unmanned logistics vehicle interacts and the public database is used as the initial scheduling conditions, and the initial scheduling path plan is randomly generated as the initial population for path planning; Fitness evaluation: The set of restriction conditions is used as the constraint conditions in fitness evaluation, and the fitness evaluation formula is established at the same time: ZYj represents the utilization efficiency of resource type j of the target unmanned logistics vehicle, m represents the total amount of resource types of the target unmanned logistics vehicle, fj represents the preset influence coefficient corresponding to resource type j of the target unmanned logistics vehicle, and F is the fitness value; Selection operation: Use the roulette wheel selection method to select offspring as new parents; Crossover operation: randomly exchange data in chromosomes of different parents; Mutation operation: randomly select data of different offspring chromosomes for adjustment; Iteration and termination conditions: When the pre-set termination conditions are reached, the chromosome with the highest fitness is selected from the final population for decoding to obtain the optimal scheduling path solution.

9. An intelligent dispatching system for unmanned logistics vehicles in a park, implemented based on an intelligent dispatching method for unmanned logistics vehicles in a park as claimed in any one of claims 1 to 8, characterized in that: include: Storage module, which stores all scheduling data in a decentralized blockchain network; The scenario judgment module connects each unmanned logistics vehicle to the blockchain network, judges the usage scenario of each unmanned logistics vehicle, and determines whether there is a need for information flow for the target unmanned logistics vehicle; Interaction screening module: when there is a need for information flow between the target unmanned logistics vehicles, the information interaction mechanism is activated, and the unmanned logistics vehicles that need to interact with the target unmanned logistics vehicles are screened out through interaction analysis; The feature extraction module obtains the database combination corresponding to the information interaction from the blockchain network, and based on the database combination, uses the pre-trained neural network model to extract restricted features to obtain a set of restriction conditions; The scheduling module uses a genetic algorithm to optimize the scheduling path of the target unmanned logistics vehicle based on a set of constraints and updates the optimized scheduling path in the blockchain network.

Images