Multi-task planning method and system for unmanned equipment

Abstract

The invention relates to a multi-task planning method and system for unmanned equipment, belongs to the technical field of unmanned equipment task planning, and solves the problem that complex logistics task planning in the prior art cannot give consideration to planning efficiency and dynamic adaptability requirements. The method comprises the following steps: dividing according to a plurality of time points related to all tasks to obtain at least one time zone; according to the sequence of the time zones, dynamically determining a first priority of the involved tasks in each time zone, and sequentially performing single-task planning on one or more tasks according to the first priority; wherein in each single task planning, the second priority of one or more available unmanned aerial vehicles is dynamically determined, and the delivery fragment of each available unmanned aerial vehicle in the current task in the current time slice area is traversed and simulated according to the second priority; and summarizing and optimizing all the delivery fragments to obtain an overall execution plan of all the current tasks. According to the technical scheme, flexible planning and dynamic optimization of multiple tasks are effectively achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned equipment task planning, and particularly relates to a multi-task planning method and system for unmanned equipment. BACKGROUND

[0002] An unmanned equipment is a self-operated equipment relying on a self-provided program control device or using radio remote control to operate, which includes unmanned aerial vehicles, unmanned vehicles, unmanned ships and various bionic unmanned equipment, and is mainly applied to logistics, transportation, firefighting, agriculture, shooting, surveying, search and rescue, reconnaissance and combat fields of enterprises, institutions, individuals or armies. The autonomous ability of the unmanned equipment mainly depends on its task planning and execution ability. Taking the unmanned aerial vehicle logistics field as an example, a large logistics system often needs to manage and control multiple equipment (such as a group of unmanned aerial vehicles), multiple tasks and complex environments, and needs to comprehensively consider factors such as equipment performance, task demand, environmental conditions, multiple equipment cooperation and multiple task scheduling, so as to plan one or more optimal or suboptimal flight paths for each unmanned aerial vehicle from a starting point to a target point, and ensure that the unmanned aerial vehicle efficiently and successfully completes the flight task and safely returns.

[0003] Traditional unmanned aerial vehicle task planning generally relies on a dispatch algorithm for pre-planning or an intelligent model for on-site training. In the dispatch algorithm, the unmanned aerial vehicle parameters and execution tasks are taken as inputs, and path planning algorithms (such as A*, Dijkstra, genetic algorithm, etc.), task allocation algorithms (such as contract net, market mechanism-based, etc.), formation control algorithms (such as virtual structure method, behavior-based method, etc.) and other algorithms are sequentially processed to directly output the plan of the unmanned aerial vehicle flight activities. The pre-planning method has the advantages of fast calculation and quick response to task allocation demand, but has the disadvantage that it cannot timely consider the dynamic changes on site in a complex unknown environment, and lacks the ability to flexibly adapt to business. The intelligent model is to sample a limited number of unmanned aerial vehicle activities by simulating the unmanned aerial vehicle, the task and the on-site conditions, compare and analyze multiple simulation schemes, select the optimal scheme as an example, and finally form a planning scheme for all tasks and equipment by relying on the example samples. The on-site training method has the advantages of being close to business and being able to better adapt to uncertainties, emergencies and complexities that may be encountered during implementation, but has the disadvantage of slow execution speed. The simulation process has an exponential increase in the amount of calculation with the increase of the unmanned equipment, the passing points and the task quantity, resulting in difficulty in timely forming a planning scheme.

[0004] It can be seen that both types of multi-task multi-unmanned aerial vehicle logistics planning schemes in the prior art have obvious defects, and cannot balance the task planning efficiency and dynamic adaptability, and are difficult to be applied to major or urgent logistics projects. SUMMARY

[0005] In view of the above analysis, the embodiments of the present application aim to provide a multi-task planning method and system for unmanned equipment to solve the problem that the existing complex logistics task planning cannot balance the planning efficiency and dynamic adaptability requirements.

[0006] In one aspect, the embodiments of the present application provide a multi-task planning method for unmanned equipment, which comprises the steps of:

[0007] obtaining relevant information of all current tasks and relevant information of all available resources, and dividing at least one time slice area according to multiple time points of the relevant current all tasks;

[0008] determining a first priority of one or more tasks involved in each time slice area according to the time sequence of the time slice area, and sequentially performing single task planning on the one or more tasks according to the order of the first priority;

[0009] In each single task planning, a second priority of one or more available unmanned aerial vehicles in the all available resources is dynamically determined, and each of the available unmanned aerial vehicles is simulated in a delivery segment in the current time slice area and the current task according to the order of the second priority.

[0010] All the delivery segments are summarized and optimized to obtain a total execution plan for the current all tasks.

[0011] Based on the further improvement of the above method, the simulation of each of the available unmanned aerial vehicles in the delivery segment in the current time slice area and the current task comprises the steps of:

[0012] Based on the information of the current available resources and the second priority, the number of flights of each of the available unmanned aerial vehicles in the current time slice area for the current task is determined;

[0013] The series of actions of each of the available unmanned aerial vehicles are simulated in sequence, and each action is associated with the corresponding time and resources to form a task delivery segment of each of the available unmanned aerial vehicles;

[0014] When each of the task delivery segments is simulated, the information of all the tasks and all the available resources is updated.

[0015] Based on the further improvement of the above method, the summary and optimization of all the delivery segments comprises the steps of:

[0016] An unmanned aerial vehicle task allocation matrix C is constructed according to all the delivery segments:

[0017]

[0018] wherein, c ijrepresents the number of transportation of the UAV i to the destination j, n represents the total number of UAVs, and m represents the total number of destinations;

[0019] determine a plurality of constraint conditions according to the relevant information of the current total tasks and the relevant information of the total available resources;

[0020]

[0021] wherein a i represents the carrying capacity of the UAV i, M j represents the total demand of the destination j, represents the task duration of the UAV i, T ij represents the time for the UAV i to arrive at the destination j, represents the endurance time of the UAV i;

[0022] substitute the allocation matrix C into the plurality of constraint conditions, and determine whether the allocation matrix C is established according to the time requirement and the mutual exclusivity of resources;

[0023] set an optimization target of total task duration, identify one or more unreasonable c ij values and automatically adjust, and solve the allocation matrix C that meets the optimization target of the total task duration.

[0024] Based on further improvement of the above method, the optimization target of the total task duration is the minimum total task duration, and when the allocation matrix C makes the task duration of each UAV i as close as possible, the optimization target of the total task duration is determined to be met.

[0025] Based on further improvement of the above method, the first priority is determined according to the task urgency calculated according to the task end time point and the task start time point, and the second priority is determined according to the comprehensive determination of the carrying capacity, the enabled state and the endurance capacity of the UAV.

[0026] Based on further improvement of the above method, the method further comprises:

[0027] real-time receive a new task indication, update the relevant information of the current total tasks according to the new task indication, and adjust the division of the time slice area;

[0028] from the first time slice area that changes, sequentially adjust the single task planning.

[0029] Based on further improvement of the above method, the simulation of each of the delivery segments is determined by prior simulation sampling of the UAV activity site.

[0030] In another aspect, an embodiment of the present application provides a multi-task planning system of an unmanned device, the system comprising:

[0031] a time slice division module configured to obtain relevant information of all current tasks and relevant information of all available resources, and divide at least one time slice according to multiple time points of the relevant information of all current tasks;

[0032] a single task planning module configured to dynamically determine a first priority of one or more tasks involved in each time slice according to a time sequence of the time slices, and sequentially perform single task planning on the one or more tasks according to the first priority;

[0033] The single task planning module comprises a delivery segment planning module configured to dynamically determine a second priority of one or more available UAVs in the all available resources in each single task planning, and simulate delivery segments of each available UAV in a current task in a current time slice according to the second priority.

[0034] a whole planning module configured to aggregate and optimize all the delivery segments to obtain a total execution plan of the current all tasks.

[0035] Based on the further improvement of the above system, the delivery segment planning module comprises:

[0036] a flight sortie estimation module configured to determine a flight sortie of each available UAV in a current task in a current time slice based on the information of the current available resources and the second priority;

[0037] an action simulation module configured to simulate a series of actions of each flight sortie of each available UAV, and associate each action with corresponding time and resources to form a task delivery segment of each available UAV;

[0038] an information updating module configured to update all task and all available resource information when each task delivery segment is simulated.

[0039] Based on the further improvement of the above system, the whole planning module comprises:

[0040] a distribution matrix construction module configured to construct a UAV task distribution matrix C according to all the delivery segments:

[0041]

[0042] wherein c ij represents the number of times of transportation of UAV i to destination j, n is the total number of UAVs, and m is the total number of destinations.

[0043] a constraint setting module configured to determine multiple constraint conditions according to the relevant information of all current tasks and the relevant information of all available resources.

[0044]

[0045] wherein, a i is the carrying capacity of the unmanned aerial vehicle i, M j is the total demand of the destination j, is the task duration of the unmanned aerial vehicle i, T ij is the time for the unmanned aerial vehicle i to reach the destination j, is the endurance time of the unmanned aerial vehicle i;

[0046] a constraint verification module for substituting the allocation matrix C into the plurality of constraint conditions, and determining whether the allocation matrix C is established according to the time requirement and the mutual exclusivity principle of resources;

[0047] an optimization module for setting an optimization target of total task duration, identifying one or more unreasonable c ij values and automatically adjusting, solving the allocation matrix C satisfying the optimization target of the total task duration.

[0048] Compared with the prior art, the embodiment of the present application solves the multiple drawbacks of the conventional task planning method, such as not being universal and not being efficient enough, and according to the technical solution provided by the present application, the multi-task planning of the unmanned aerial vehicle can be simulated and optimized under the assistance of the model constraint algorithm, the running load of the simulation part is effectively reduced after dimension reduction, the universal ability of the algorithm part is strengthened, and the user experience of the task planning is effectively improved.

[0049] In the present application, the above-mentioned technical solutions can be combined with each other to realize more preferred combination solutions. Other features and advantages of the present application will be described in the subsequent specification, and some advantages will become apparent from the specification, or will be understood by implementing the present application. The purposes and other advantages of the present application can be realized and obtained from the contents specifically pointed out in the specification and the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0050] The accompanying drawings are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings:

[0051] Figure 1 is a flow chart of the multi-task planning method of the unmanned device in an embodiment of the present application;

[0052] Figure 2 is a nested flowchart of the multi-layer cycle of the multi-task planning method of the unmanned device in a preferred embodiment of the present application;

[0053] Figure 3This is a schematic diagram illustrating the simulation effect of multiple UAV delivery segments during single-task planning in a preferred embodiment of the present invention. Detailed Implementation

[0054] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0055] A specific embodiment of the present invention discloses a multi-task planning method for unmanned equipment, such as... Figure 1 As shown, the method includes the following steps:

[0056] S11, obtain relevant information of all current tasks and relevant information of all available resources, and divide at least one time segment based on multiple relevant time points of all current tasks.

[0057] In embodiments of the present invention, for each task, the task-related information includes at least one or more of the following: task start time (the earliest start time specified by the task requester), task end time (the latest end time specified by the task requester), task delivery destination, type and quantity of items required by the task, etc.; the available resource-related information includes at least one or more of the following: warehouse locations and corresponding inventory status, type, quantity and status of available drones, location and occupancy status of charging / fueling points, location and occupancy status of docking points, airspace occupancy status, etc.

[0058] After obtaining relevant information for all current tasks, the multiple start and end times related to all current tasks are arranged in chronological order into a time series. The time interval determined by each pair of adjacent time points in the time series is taken as a time slice, resulting in one or more time slices.

[0059] For example, such as Figure 3 As shown, suppose there are two tasks, and their start and end times are arranged in sequence as a time series consisting of four moments: A, B, C, and D. Moments A and B can determine a time slice t. AB Time points B and C can determine a time slice t. BC Time points C and D can determine a time slice t. CD, and 3 continuous time slices are obtained. Due to the uncertainty of the tasks, one time slice can correspond to 0, 1 or multiple tasks, depending on the overlap of the execution time periods of multiple tasks. The time slice with 0 tasks is obviously better to handle, and no task planning is needed, only the whole resources are arranged to prepare as much as possible (manage the power of the unmanned aerial vehicle, the docking position and the warehouse materials, etc.). For the time slice with 1 task, single task planning is directly performed; for the time slice with multiple tasks, the task priorities of the tasks are determined first, and multiple single task planning is performed according to the task priority order. In this way, the embodiment of the present application reduces the multi-task planning to multiple single task planning, thereby reducing the system complexity and realizing efficient and real-time planning calculation.

[0060] S12, according to the time sequence of the time slice, dynamically determining the first priority of one or more tasks involved in each time slice, and sequentially performing single task planning on the one or more tasks according to the order of the first priority;

[0061] In each single task planning, the second priority of one or more available unmanned aerial vehicles in the whole available resources is dynamically determined, and the delivery segment of each available unmanned aerial vehicle in the current task in the current time slice is simulated according to the order of the second priority.

[0062] In the embodiment of the present application, because the task end time point represents the latest end time specified by the task party, it is considered as a hard requirement first, so the priority of the task with the earlier task end time point is higher; for multiple tasks with the same end time point, because the later the start time point is, the more urgent the task appears (or the less execution time is given), the priority of the task with the later start time point is further higher; for multiple tasks with the same end time point and start time point, in principle, the priority does not need to be distinguished, and in actual operation, the priority can be randomly allocated, or further according to the amount of task demand, the distance of the destination and other factors to further distinguish the priority.

[0063] In embodiments of this invention, multi-task planning is implemented using nested multi-level loop planning. This involves sequentially traversing resources and tasks according to different priority levels to achieve real-time scheduling of multiple tasks in complex environments. Specifically, the outermost loop traverses time slices, analyzing each time slice in the divided time order to determine if there are any unplanned tasks. The middle loop traverses single-task planning, analyzing each task in the first priority order to determine if it can accommodate more resources within its current time slice (full or no resources available). The innermost loop traverses drone resources, assigning each drone in the second priority order and determining its activation count under the current task in the current time slice. Planning ends from the inside out: when the innermost drone traversal loop is complete, the current single-task planning is complete; when the middle single-task planning loop is complete, the current single-time-slice planning is complete; and when the outermost single-time-slice planning loop is complete, the entire multi-task planning is finished.

[0064] Specifically, such as Figure 2 As shown, the information acquisition and time segment division are the same as in step S11 above. Then, in step S121, the time segments are selected in chronological order, and each task in the time segment is set and selected according to the task priority (hereinafter referred to as the first priority) mentioned above. Figure 2 The preferred embodiment executes nested multi-level loops:

[0065] S122: Select the highest priority unprocessed task in the current time zone as the current task. For all available drones for the current task, first set the priority according to the load capacity, activation status and endurance.

[0066] In this invention, multiple drones of different models are used to jointly complete task planning. To ensure smooth scheduling, single-task planning first determines the priority of all available drones (hereinafter referred to as the second priority). In a preferred embodiment of this invention, drones with higher payload capacity have higher priority. Among multiple drones with the same payload capacity (e.g., the same model), those that are already activated and have weaker endurance (lower battery or fuel) have higher priority. Drones with the same payload but not yet activated are ranked after activated drones of the same payload level and their priority is randomly assigned. Drones with weaker payload capacity have lower priority. In this way, drones with larger payloads are given priority in this embodiment of the invention, aiming to meet task requirements with as few sorties and in as short a time as possible. At the same time, as few drones as possible are activated to reduce planning complexity and reserve as many resources as possible to cope with dynamically changing task requirements or complex situations. Finally, as many activated drones as possible are planned for use to determine their endurance limits as soon as possible and arrange for them to be removed from the plan in a timely manner.

[0067] S123, according to the second priority of available drones, traverse and simulate the planning of their mission delivery segments for the current task within the current time segment;

[0068] For drones with the highest priority (largest payload, activated, and shortest range), their mission delivery segments within the current time zone are directly calculated. This calculation will simultaneously trigger dynamic updates to mission and available resource information.

[0069] S124. If the current task requirement cannot be met, determine whether all available drones have been traversed. If there are still available drones, return to step S123 to calculate the task delivery segment of the current drone in this time slot according to the priority order of the drones and the updated task and available resource information. Until all available drones have been traversed, the current task requirement has been met, or at least one key resource of the task in this time slot has been fully loaded, proceed to step S125.

[0070] by Figure 3 To illustrate, let's assume that in the current time region t... BC If the first drone with the highest priority can complete a 200kg delivery, and this drone is currently available and the current mission requirement is greater than 200kg, then this drone will be dispatched first. The plan includes a series of actions for the drone, such as arriving at the delivery warehouse from its current location, landing and parking using take-off and landing resources, receiving materials using the delivery warehouse docking position, taking off using take-off and landing resources, flying from the delivery warehouse to the destination, landing and parking using take-off and landing resources, delivering materials using the destination docking position, taking off using take-off and landing resources, and returning to the designated location. Each action, along with its corresponding time and resources, will form a mission delivery segment. The mission and available resource information will be updated (i.e., the mission and resource status after excluding the resources occupied by this mission delivery segment, such as the remaining mission requirement, remaining warehouse inventory, available time periods for each docking position, and available time periods for airspace). Based on the updated information, the second-priority drone performs interleaving calculations (i.e., calculates the mission execution probability using the resource-free time period left by the first drone), determining that it can complete 4 deliveries of 30kg each within its current time slot. It then plans these 4 delivery segments sequentially (the mission actions for each segment are similar to those described above and will not be detailed further), and subsequently updates the mission and available resource information. Based on the updated information, the third-priority drone performs interleaving calculations again, determining that it can also complete 4 deliveries of 30kg each within its current time slot. Similarly, it plans these 4 delivery segments sequentially and updates the mission and available resource information.

[0071] It should be noted that, Figure 3The task delivery segment in the time slice is only exemplary to represent the whole time occupation, and in actual planning, the time arrangement of each specific action can be split, such as 2 minutes for receiving materials, 1 minute for taking off and ascending, 10 minutes for flying to the destination, and so on. That is, in actual planning, each task delivery segment is a combination of a series of action times, and through the planning of subdivided action times, more precise mutual exclusion condition recognition, resource allocation and time management can be achieved. However, in the planning of the overall goal, for example, the delivery arrangement of the time slice and the constraints and optimization of the task allocation matrix C in the subsequent text, the whole task delivery segment is processed, so as to avoid too high system complexity and too much computing resources. In the embodiment of the present application, the specific action combination and action time arrangement of the task delivery segment can be determined by prior simulation sampling of the unmanned aerial vehicle activity site, and by splitting the complete task into each delivery segment of the time slice, the segment data is determined by using the simulation sampling result of each action in advance, so as to realize the combination of the advantages of planning algorithm and field simulation. This mode obviously improves the flexibility of simulation sampling, greatly improves the planning speed of the system, and can ensure that the pre-planned task delivery segment can also match the field conditions, and realizes the consideration of task planning efficiency and dynamic adaptability.

[0072] In the preferred embodiment of the present application, the planning of the task delivery segment can be appropriately more than the range of the time slice or the task demand, and a certain number of delivery segments are planned, such as Figure 3 As shown in the time slice t BC One delivery segment is additionally planned for the second and third unmanned aerial vehicles, and through this mode, the system can predict the carrying capacity of the unmanned aerial vehicle and resources, so as to prevent some adverse situations that may occur in the future (such as unmanned aerial vehicle failure, loss of endurance, etc.), and flexibly adjust the planning according to the prior simulation. When the unmanned aerial vehicle performs the current task, it needs to confirm that the unmanned aerial vehicle still has the endurance to return to the charging / refueling point after the flight plan is completed, otherwise it needs to be charged or refueled before performing the next task. For the unmanned aerial vehicle that does not have endurance and exits the planning, its state is automatically set as unavailable, and the system will arrange it to charge or supplement fuel (occupy charging point / fuel supplement point resources); after completing the charging or fuel supplement, the state of the unmanned aerial vehicle is automatically updated as available but not enabled, and the second priority is set accordingly.

[0073] The UAVs are sequentially traversed according to the priorities of the UAVs to simulate the mission delivery segments of the tasks, until at least one of the deadline conditions of the current task in the time slice occurs. In the preferred embodiment of the present application, the deadline conditions include, but are not limited to, the current task demand being satisfied, all available UAVs being traversed, or at least one critical resource being full (for example, all receiving capacities of the resources such as the destination docking positions, take-off and landing resources, and / or airspace in the current time slice are full and cannot accommodate more UAVs). At this point, a single task planning for the current time slice is completed, and the judgment of step S125 is entered.

[0074] S125, whether all tasks in the current time slice have been traversed, if there are still unprocessed tasks, return to step S122, and repeat the above steps S122-S125 to implement the next single task planning according to the next priority in the multi-layer loop (based on the remaining available resource information, and according to the priority of the UAV, each UAV is traversed to plan its mission delivery segment in the current time slice for the task); if all tasks involved in the time slice have been traversed, or at least one critical resource in the current time slice is full (for example, all available UAVs have been dispatched, all warehouse inventories have been shipped, etc.), the judgment of step S126 is entered.

[0075] S126, whether all time slices have been traversed, if there are still unprocessed time slices, return to step S121, and repeat the above steps S121-S126 to implement one or more single task planning for the next time slice according to the next priority in the multi-layer loop; if all time slices have been traversed, or at least one critical resource in the system is full, step S13 is entered.

[0076] S13, the delivery segments are summarized and optimized to obtain the overall execution plan for the current tasks.

[0077] In the embodiments of the present application, many resources are mutually exclusive resources, such as docking positions, charging positions, delivery positions, etc., which can only satisfy one demand in a certain time period, and therefore, whether in single task planning or in the summary of multiple tasks and multiple time slices, the constraint conditions such as resource mutual exclusion need to be considered. In addition, in some cases, some data in the result after the multi-layer loop may be abnormally high, such as a certain type of UAV being overused while there are a large number of resources not being effectively utilized, and therefore, it is necessary to set a certain target to select a reasonable number and type of UAVs from a plurality of different numbers of UAV types for the overall planning of multiple tasks.

[0078] Specifically, in a preferred embodiment of the present application, the summarizing and optimizing of all the delivery segments includes the steps of:

[0079] According to all the delivery segments, a UAV task allocation matrix C is constructed:

[0080]

[0081] wherein c ij represents the number of transportation of the UAV i to the destination j, n is the total number of UAVs, and m is the total number of destinations;

[0082] determine a plurality of constraint conditions according to the relevant information of the current all tasks and the relevant information of all available resources;

[0083]

[0084] wherein a i is the carrying capacity of the UAV i, M j is the total demand of the destination j, is the task duration of the UAV i, T ij is the time of the UAV i to the destination j, is the endurance time of the UAV i;

[0085] substitute the distribution matrix C into the plurality of constraint conditions, and determine whether the distribution matrix C is established according to the time requirement and the mutual exclusivity of resources;

[0086] set an optimization target of total task duration, identify one or more unreasonable c ij values and automatically adjust to solve the distribution matrix C that meets the optimization target of the total task duration.

[0087] wherein the optimization target of the total task duration is the minimum total task duration, when the distribution matrix C makes the task duration of each UAV i as close as possible, it is determined that the optimization target of the total task duration is met.

[0088] The specific optimization process can use a search iteration method to adjust the distribution matrix. The iterative adjustment can use a particle swarm optimization or a simulated annealing algorithm to generate a new c ij value at each step, and then evaluate the performance of the new matrix according to the algorithm rules, determine the convergence mode, and perform the next iteration. In the particle swarm optimization, the particles are updated according to the individual and group historical optimal positions. The simulated annealing algorithm searches by controlling parameters and gradually converges to the global optimal solution. Both of them can obtain at least one set of distribution matrix that meets the optimization target or is closest to the optimization target within a controllable time with a reasonable cost.

[0089] Further, the embodiments of the present application also support dynamic updating and adjustment in the task planning and execution scene. Specifically:

[0090] ​Real-time receive new task indication, update the relevant information of all current tasks according to the new task indication and adjust the division of time slice area; from the first changed time slice area, adjust the single task planning in turn.

[0091] If the allocation matrix C cannot be obtained or the allocation matrix C cannot satisfy all the constraint conditions in a certain planning process, it means that the current resource condition cannot meet all the task requirements, and at least one of the task time and / or material requirements needs to be adjusted immediately, and then the planning adjustment is carried out based on the adjusted task condition. Generally, a small amount of change in resources will not affect the overall planning of the task, but if a large-scale abnormal event occurs, such as extreme weather, large-area electromagnetic interference, damaged warehouse or destination, etc., the allocation matrix C cannot be established or the data is abnormal, at this time, the system alarm will also be triggered to notify manual intervention.

[0092] The embodiment of the application realizes flexible planning and dynamic optimization of multiple tasks to the greatest extent. During or before task execution, tasks can be added or modified, several flight activities can be reasonably modified, UAV resources, inventory resources, transfer location resources, delivery resources, refueling / charging resources, location fixed / movable configuration parameters can be changed at will, and actual flight restrictions can be added. Regardless of the state or information update, the overall execution plan can be adjusted. The technical scheme of the application can identify the changed information in real time, adjust multiple cycles as necessary, and manage the tasks and resource information involved in the cycle process by mutual exclusion and constraint, so as to supplement and perfect the related information, and efficiently and timely complete the re-planning.

[0093] On the other hand, the embodiment of the application provides a multi-task planning system of an unmanned device, which comprises:

[0094] A time slice area division module is configured to obtain the relevant information of all current tasks and the relevant information of all available resources, and divide at least one time slice area according to the relevant multiple time points of the current all tasks;

[0095] A single task planning module is configured to dynamically determine the first priority of one or more tasks involved in each time slice area according to the time sequence of the at least one time slice area, and sequentially plan the one or more tasks according to the first priority.

[0096] The single-task planning module comprises a delivery segment planning module, which is configured to dynamically determine a second priority of one or more available UAVs in the total available resources in each single-task planning, and simulate delivery segments of each available UAV in the current task in the current time slice according to the order of the second priority.

[0097] The overall planning module is configured to aggregate and optimize all the delivery segments to obtain an overall execution plan of the current total task.

[0098] Based on the further improvement of the above system, the delivery segment planning module comprises:

[0099] The flight sorties estimation module is configured to determine the flight sorties of each available UAV for the current task in the current time slice based on the information of the current available resources and the second priority.

[0100] The action simulation module is configured to simulate a series of actions of each flight sortie of each available UAV in sequence, and associate each action with corresponding time and resources to form a task delivery segment of each available UAV.

[0101] The information updating module is configured to update the information of all tasks and all available resources when each task delivery segment is simulated.

[0102] Based on the further improvement of the above system, the overall planning module comprises:

[0103] The allocation matrix construction module is configured to construct a UAV task allocation matrix C according to all the delivery segments:

[0104]

[0105] wherein, c ij represents the number of transportation of UAV i to destination j, n is the total number of UAVs, and m is the total number of destinations;

[0106] The constraint setting module is configured to determine a plurality of constraint conditions according to the related information of the current total task and the related information of the total available resources:

[0107]

[0108] wherein, a i is the carrying capacity of UAV i, M j is the total demand of destination j, is the task duration of UAV i, T ij is the time of UAV i to destination j, is the endurance time of UAV i;

[0109] A constraint verification module is configured to substitute the allocation matrix C into the plurality of constraint conditions, and determine whether the allocation matrix C is correct according to time requirements and resource mutual exclusion principles.

[0110] An optimization module is configured to set an optimization target of total task duration, identify one or more unreasonable c ij values and automatically adjust, and solve the allocation matrix C that satisfies the optimization target of the total task duration.

[0111] In summary, the embodiments of the present application solve the problems of conventional task planning methods, such as lack of generalization and lack of efficiency. According to the technical solutions provided by the present application, the multi-task planning of the unmanned aerial vehicle can simulate and select the optimal core business part of each task under the assistance of the model constraint algorithm. The single-task planning after dimension reduction effectively reduces the running load of the simulation part and enhances the general ability of the algorithm part, thereby effectively improving the user experience of the task planning.

[0112] Those skilled in the art can understand that all or part of the processes of the above-mentioned embodiments can be completed by a computer program instructing related hardware, and the program can be stored in a computer readable storage medium. The computer readable storage medium includes a magnetic disk, an optical disk, a read-only memory, a random access memory, etc.

[0113] The above description is only a preferred embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A multi-task planning method for unmanned equipment, characterized in that, The method includes the following steps: Obtain relevant information about all current tasks and all available resources, and divide the current tasks into at least one time segment based on multiple relevant time points. Based on the time sequence of the time slots, the first priority of one or more tasks involved in each time slot is dynamically determined, and the one or more tasks are planned one by one in sequence according to the first priority. In each single-task planning, the second priority of one or more available drones among all available resources is dynamically determined, and the delivery segment of each available drone in the current task of the current time segment is simulated according to the order of the second priority. All the delivered segments are summarized and optimized to obtain the overall execution plan for all current tasks.

2. The method according to claim 1, characterized in that, The simulation of each available UAV's delivery segment in the current time slot and current task includes the following steps: Based on the information of currently available resources and the second priority, determine the number of sorties each available UAV can fly for the current task within the current time slot; The series of actions of each sortie of each available UAV are simulated in sequence, and each action is associated with the corresponding time and resources to form a mission delivery segment for each available UAV. Update all tasks and all available resource information as each of the aforementioned task delivery segments is simulated and completed.

3. The method according to claim 1, characterized in that, The process of summarizing and optimizing all the delivered segments includes the following steps: Construct the UAV task allocation matrix C based on all the aforementioned delivery segments: Among them, c ij This represents the number of times drone i travels to destination j, where n is the total number of drones and m is the total number of destinations. Based on the relevant information of all current tasks and all available resources, several constraints are determined: Among them, a i For the payload capacity of drone i, M j Let the total demand for destination j be , For the mission duration of drone i, T ij Let i be the time it takes for drone i to travel to destination j. The flight time of drone i; Substitute the allocation matrix C into the multiple constraints, and determine whether the allocation matrix C holds true based on time requirements and the principle of mutual exclusion of resources; Set an optimization target for the total task duration, and identify one or more unreasonable c ij The value is automatically adjusted to solve for the allocation matrix C that satisfies the optimization objective of the total task duration.

4. The method according to claim 3, characterized in that, The optimization objective for the total task duration is to minimize the total task duration, when the allocation matrix C minimizes the task duration of each UAV i. When the time is as close as possible, it is determined that the optimization target of the total task duration is met.

5. The method according to claim 1, characterized in that, The first priority is determined based on the urgency of the task calculated from the task end time and the task start time, while the second priority is determined based on the drone's payload capacity, activation status, and endurance.

6. The method according to claim 1, characterized in that, The method further includes: Receive new task instructions in real time, update the relevant information of all current tasks and adjust the division of time slots according to the new task instructions; Starting with the first time segment that changes, adjust the single-task planning sequentially.

7. The method according to claim 1 or 2, characterized in that, The simulations of each delivery segment were determined in advance through on-site simulation sampling of UAV activities.

8. A multi-task planning system for unmanned equipment, characterized in that, The system includes: The time-segmentation module is used to obtain relevant information of all current tasks and all available resources, and to divide at least one time segment based on multiple relevant time points of all current tasks. The single-task planning module is used to dynamically determine the first priority of one or more tasks involved in each time slot according to the time sequence of the time slot, and to perform single-task planning on the one or more tasks in sequence according to the first priority. The single-task planning module includes a delivery segment planning module, which is used to dynamically determine the second priority of one or more available drones among all available resources in each single-task planning, and to traverse and simulate the delivery segments of each available drone in the current task in the current time segment according to the order of the second priority. The overall planning module is used to summarize and optimize all the delivery segments to obtain the overall execution plan for all current tasks.

9. The system according to claim 8, characterized in that, The delivery segment planning module includes: The sortie estimation module is used to determine, based on information about currently available resources and the second priority, the number of sorties each available UAV can fly for the current mission within the current time slot; The motion simulation module is used to sequentially simulate a series of actions for each sortie of each available UAV, and associate each action with the corresponding time and resources to form a mission delivery segment for each available UAV. The information update module is used to update all task and all available resource information when each of the task delivery segments is simulated and completed.

10. The system according to claim 8, characterized in that, The overall planning module includes: The allocation matrix construction module is used to construct the UAV mission allocation matrix C based on all the delivery segments: Among them, c ij This represents the number of times drone i travels to destination j, where n is the total number of drones and m is the total number of destinations. The constraint setting module is used to determine multiple constraints based on relevant information of all current tasks and all available resources: Among them, a i For the payload capacity of drone i, M j Let the total demand for destination j be , For the mission duration of drone i, T ij Let i be the time it takes for drone i to travel to destination j. The flight time of drone i; The constraint verification module is used to substitute the allocation matrix C into the multiple constraint conditions and determine whether the allocation matrix C is valid based on time requirements and the mutual exclusion principle of resources. The optimization module is used to set the optimization target for the total task duration and identify one or more unreasonable values. ij The value is automatically adjusted to solve for the allocation matrix C that satisfies the optimization objective of the total task duration.

Images