CN-122018522-A - Unmanned aerial vehicle cluster distributed task allocation method and system under communication limitation

CN122018522ACN 122018522 ACN122018522 ACN 122018522ACN-122018522-A

Abstract

The invention belongs to the technical field of unmanned aerial vehicle cluster cooperative control, and discloses a unmanned aerial vehicle cluster distributed task allocation method and system under communication limitation, wherein the method comprises the steps of constructing a communication prediction model; the method comprises the steps of carrying out priority calculation and importance evaluation on cache tasks of unmanned aerial vehicles, constructing a distributed task cache space based on the cache tasks subjected to the priority calculation and the importance evaluation, judging future communication states by the unmanned aerial vehicles by using a communication prediction model, carrying out multi-round negotiation comparison on each unmanned aerial vehicle by using a distributed auction mechanism, carrying out autonomous execution on the basis of the cache tasks subjected to the priority calculation and the importance evaluation and updating the execution states of the unmanned aerial vehicles if the future communication states are limited, carrying out completion degree judgment on a task area, and generating a task report if all local tasks in the task area are completed. The invention can improve the utilization efficiency of resources, and can quickly respond to environmental changes and newly discovered clues based on a dynamic task redistribution mechanism of real-time information.

Inventors

REN XUEFENG
HOU YANCHENG
LI LINA

Assignees

北京卓翼智能科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260121

Claims (10)

1. The unmanned aerial vehicle cluster distributed task allocation method under the communication limitation is characterized by comprising the following steps of: acquiring a preset task area, and constructing a communication prediction model by utilizing a preset digital elevation model and a long-short-term memory network model; Dividing a task area into a plurality of local tasks, taking the local tasks as cache tasks of the unmanned aerial vehicle, and carrying out priority calculation and importance evaluation on the cache tasks of the unmanned aerial vehicle by combining the unmanned aerial vehicle performance parameters and the historical execution results of the unmanned aerial vehicle, which are obtained in advance; Based on the buffer tasks after priority calculation and importance evaluation, constructing a distributed task buffer space, distributing trust scores to the buffer tasks, and adjusting the buffer tasks by utilizing a dynamic elimination mechanism according to the scoring results; The unmanned aerial vehicle judges future communication states by using a communication prediction model, if the future communication states are normal, based on the adjusted cache tasks, the unmanned aerial vehicle performs multi-round negotiation comparison on each unmanned aerial vehicle by using a distributed auction mechanism, determines the unmanned aerial vehicle executing the adjusted cache tasks, and updates the unmanned aerial vehicle executing state; if the future communication state is limited, the unmanned aerial vehicle autonomously executes based on the cache task after priority calculation and importance evaluation, and updates the execution state of the unmanned aerial vehicle; and judging the completion degree of the task area, if all local tasks in the task area are completed, generating a task report, otherwise, re-executing the unmanned aerial vehicle to judge the future communication state.
2. The method for distributing tasks of a cluster of unmanned aerial vehicles under communication restriction according to claim 1, wherein the steps of obtaining a preset task area and constructing a communication prediction model by using a preset digital elevation model and a long-short-term memory network model comprise: Extracting topographic data in the digital elevation model and establishing a mathematical model by utilizing a wireless propagation theory to characterize the influence of topography on communication; dividing a preset task area into grids, detecting the line-of-sight blocking condition of any two grids through a ray tracing algorithm, and establishing a position-communication quality mapping table; and taking the mathematical model and the current position data of the unmanned aerial vehicle as input features of a long-period memory network model, and constructing a communication prediction model by combining the movement track and the history of the unmanned aerial vehicle.
3. The method for distributing tasks among unmanned aerial vehicle clusters under communication limitation according to claim 2, wherein the steps of extracting the topographic data from the digital elevation model and establishing a mathematical model to characterize the effect of the topographic data on the communication by using the wireless propagation theory comprise: Extracting altitude data, gradient data and slope data in the topographic data; Calculating an initial attenuation value of the reference signal according to a free space path loss formula; and calculating the additional attenuation value of the mountain body to the reference signal by using the blade diffraction model in the wireless propagation theory.
4. The method for distributing distributed tasks of unmanned aerial vehicle clusters under communication limitation according to claim 1, wherein the steps of constructing a distributed task cache space based on the cache tasks after priority calculation and importance evaluation, distributing trust scores to the cache tasks, and adjusting the cache tasks by using a dynamic elimination mechanism comprise: The distributed task cache space comprises a main cache task, a standby cache task and a neighbor cache task; Distributing a trust degree score to each cache task according to the reliability of the cache task source and a historical execution result, wherein the historical execution result is the success rate and quality evaluation record of the past execution of the same type task of each unmanned aerial vehicle; And according to the storage space limitation of the unmanned aerial vehicle and the importance of the cache task, performing self-adaptive elimination on the cache task by utilizing a dynamic elimination mechanism.
5. The method for distributing distributed tasks of unmanned aerial vehicle clusters under communication restriction according to claim 4, wherein the distributed task buffer space comprises a main buffer task, a standby buffer task and a neighbor buffer task, and the method comprises the following steps: The main cache task is used for storing a current executing task and a high-priority task to be executed, the main cache task is high in trust degree grading, the standby cache task is used for storing an alternative scheme task and a medium-priority task, the standby cache task is medium in trust degree grading, the neighbor cache task is used for storing an area search state and neighbor unmanned aerial vehicle task information, and the neighbor cache task is low in trust degree grading; Executing a hierarchical compression strategy on the cache task, wherein the main cache task adopts a lossless compression algorithm, the complete task parameters are reserved, the standby cache task adopts a lossy compression algorithm, key field information is reserved, and the neighbor cache task only reserves task summary information; when the unmanned aerial vehicle detects that the cache utilization rate exceeds a preset merging threshold, the tasks of the same type which are adjacent in space and similar in time are synthesized into task clusters and stored; When the unmanned aerial vehicle detects that the storage occupancy rate is less than 50%, a layered compression strategy is used for the cache tasks, when the storage occupancy rate is between 50% and 80%, medium compression and selective combination are carried out on all the cache tasks, when the storage occupancy rate is greater than 80%, high compression and deep combination are carried out on all the cache tasks, and key task information is reserved.
6. The method for distributing tasks to unmanned aerial vehicle clusters under communication limitation according to claim 5, wherein the main buffer task is used for storing a currently executed task and a task to be executed with high priority, the main buffer task is a high trust score, the standby buffer task is used for storing an alternative task and a task with medium priority, the standby buffer task is a medium trust score, the neighbor buffer task is used for storing a regional search state and neighbor unmanned aerial vehicle task information, and the neighbor buffer task is a low trust score, the method comprising: when the unmanned aerial vehicle receives the neighbor unmanned aerial vehicle task information through the neighbor cache task, performing trust degree comparison, and if the neighbor cache task is higher than the current main cache task, lifting the neighbor cache task to be the main cache task for execution; and degrading the current main buffer task into a standby buffer task, and sharing the original standby buffer task to the neighbor buffer task.
7. The method for distributing tasks among unmanned aerial vehicle clusters under communication limitation according to claim 4, wherein the adaptively eliminating the cache tasks by using a dynamic elimination mechanism according to the storage space limitation of the unmanned aerial vehicle and the importance of the cache tasks comprises: Eliminating the unallocated cache tasks with the residence time of more than 60 minutes in the neighbor cache tasks; Calculating a cache task by using a comprehensive value judgment formula, and if the cache task is smaller than an elimination threshold value, performing elimination treatment; When the unmanned aerial vehicle detects that the buffer utilization rate exceeds a preset threshold, sequencing the buffer tasks according to the comprehensive value of the tasks, and eliminating the buffer tasks with the comprehensive value at the lowest preset proportion; When a plurality of unmanned aerial vehicles detect that the overlapping rate between task areas corresponding to the cache tasks exceeds a preset threshold, the cache tasks are subjected to comprehensive value comparison, only the cache task with the highest comprehensive value is reserved, and if the execution failure times of the cache tasks reach the preset times threshold, the cache tasks are subjected to elimination processing.
8. The method for distributing distributed tasks of unmanned aerial vehicle clusters under communication limitation according to claim 1, wherein the unmanned aerial vehicle judges future communication states by using a communication prediction model, if the future communication states are normal, based on the adjusted cache tasks, the unmanned aerial vehicle executing the adjusted cache tasks is determined by performing multiple rounds of negotiation comparison on each unmanned aerial vehicle by using a distributed auction mechanism, and updating the execution states comprises: when the unmanned aerial vehicle detects the adjusted cache task, calculating the bidding value by combining the self parameters; A distributed experience pool is pre-configured for each unmanned aerial vehicle, and an unmanned aerial vehicle body Q value table is built by using the experience pool; When communication is established between unmanned aerial vehicles, the unmanned aerial vehicles judge whether to use bidding values to participate in distributed auction negotiation based on the Q value table; And if the plurality of unmanned aerial vehicles select the same cache tasks and/or search areas to overlap, processing each unmanned aerial vehicle by a space-time conflict detection method and a priority comparison method.
9. The method for distributing distributed tasks of a cluster of unmanned aerial vehicles under communication restriction according to claim 8, wherein when communication is established between unmanned aerial vehicles, the unmanned aerial vehicle determining whether to use bidding value to participate in distributed auction negotiation based on a Q-table comprises: when the unmanned aerial vehicle performs bidding, exchanging update information of respective Q value tables, and performing optimal bidding decision by the unmanned aerial vehicle of the receiver by adopting a weighted fusion strategy; introducing a load balance factor in the bidding process, and when the load of the adjacent unmanned aerial vehicle is too high, promoting a caching task to flow to a region with lower load; Setting initial negotiation rounds, calculating load variance after each round of negotiation, if the variance load is larger than a preset threshold value and the negotiation rounds are smaller than the maximum negotiation rounds, carrying out the next round of negotiation, and if the continuous two rounds of load variance dropping rate is smaller than the dropping value, terminating the negotiation.
10. A distributed task distribution system for a cluster of unmanned aerial vehicles under communication constraints, the system comprising: the communication prediction module is used for acquiring a preset task area and constructing a communication prediction model by utilizing a preset digital elevation model and a long-short-term memory network model; The calculation evaluation module is used for dividing the task area into a plurality of local tasks, taking the local tasks as cache tasks of the unmanned aerial vehicle, and carrying out priority calculation and importance evaluation on the cache tasks of the unmanned aerial vehicle by combining the pre-acquired unmanned aerial vehicle performance parameters and historical execution results of the unmanned aerial vehicle; The cache space module is used for constructing a distributed task cache space based on the cache tasks after priority calculation and importance evaluation, distributing trust scores to the cache tasks, and adjusting the cache tasks by utilizing a dynamic elimination mechanism according to the scoring results; The distributed auction module is used for judging future communication states by the unmanned aerial vehicle by using a communication prediction model, carrying out multi-round negotiation comparison on each unmanned aerial vehicle by using a distributed auction mechanism based on the adjusted cache tasks if the future communication states are normal, determining the unmanned aerial vehicle which executes the adjusted cache tasks, and updating the execution states of the unmanned aerial vehicles; The autonomous execution module is used for automatically executing the unmanned aerial vehicle based on the buffer task after priority calculation and importance evaluation if the future communication state is limited, and updating the execution state of the unmanned aerial vehicle; and the progress judging module is used for judging the completion degree of the task area, generating a task report if all local tasks in the task area are completed, and executing the unmanned aerial vehicle again to judge the future communication state if not.

Description

Unmanned aerial vehicle cluster distributed task allocation method and system under communication limitation Technical Field The invention relates to the technical field of unmanned aerial vehicle cluster cooperative control, in particular to a unmanned aerial vehicle cluster distributed task distribution method and system under communication limitation. Background With the development of unmanned aerial vehicle technology, unmanned aerial vehicle's application area is wider and wider, for example lies in using unmanned aerial vehicle to replace manpower resources to operate under the complicated topography environment such as mountain region search and rescue, forest fire prevention, border patrol, can improve the operating efficiency, has also guaranteed personal safety. However, existing unmanned aerial vehicle cluster systems commonly employ a centralized architecture, relying on a stable communication link for task coordination. In complex terrain environments such as mountainous regions, communication is frequently interrupted due to terrain shielding, the system cannot work normally, a search area is pre-allocated before a task starts in a traditional method, task allocation cannot be dynamically adjusted according to clues found in real time or environmental changes, and resource allocation is unreasonable. When communication is limited, each unmanned aerial vehicle lacks effective information sharing mechanism, appears many unmanned aerial vehicles repeatedly searching the same regional problem easily, extravagant time and energy. Therefore, how to provide a distributed task allocation method and system for unmanned aerial vehicle clusters under communication limitation is a problem to be solved at present. Disclosure of Invention The embodiment of the invention provides a distributed task allocation method for unmanned aerial vehicle clusters under communication limitation, which aims to solve the technical problems in the prior art. According to a first aspect of an embodiment of the present invention, a method for distributing distributed tasks of a unmanned aerial vehicle cluster under communication limitation is provided. In one embodiment, a method for distributing distributed tasks of a unmanned aerial vehicle cluster under communication limitation includes: acquiring a preset task area, and constructing a communication prediction model by utilizing a preset digital elevation model and a long-short-term memory network model; Dividing a task area into a plurality of local tasks based on a preset task area, taking the local tasks as cache tasks of the unmanned aerial vehicle, and carrying out priority calculation and importance evaluation on the cache tasks of the unmanned aerial vehicle by combining the unmanned aerial vehicle performance parameters and the historical execution results of the unmanned aerial vehicle, which are obtained in advance; Constructing a distributed task cache space based on the cache task, distributing trust scores to the cache task, and adjusting the cache task by utilizing a dynamic elimination mechanism according to the scoring results; The unmanned aerial vehicle judges future communication states by using a communication prediction model, if the future communication states are normal, based on the adjusted cache tasks, the unmanned aerial vehicle performs multi-round negotiation comparison on each unmanned aerial vehicle by using a distributed auction mechanism, determines the unmanned aerial vehicle executing the adjusted cache tasks, and updates the unmanned aerial vehicle executing state; if the future communication state is limited, the unmanned aerial vehicle autonomously executes based on the cache task after priority calculation and importance evaluation, and updates the execution state of the unmanned aerial vehicle; and judging the completion degree of the task area, if all local tasks in the task area are completed, generating a task report, otherwise, re-executing the unmanned aerial vehicle to judge the future communication state. In one embodiment, obtaining a pre-set task area and constructing a communication prediction model using a pre-configured digital elevation model and a long-short term memory network model comprises: Extracting topographic data in the digital elevation model and establishing a mathematical model by utilizing a wireless propagation theory to characterize the influence of topography on communication; dividing a preset task area into grids, detecting the line-of-sight blocking condition of any two grids through a ray tracing algorithm, and establishing a position-communication quality mapping table; and taking the mathematical model and the current position data of the unmanned aerial vehicle as input features of a long-period memory network model, and constructing a communication prediction model by combining the movement track and the history of the unmanned aerial vehicle. In one embodiment, extracting terrain data from a d