CN111723985A - Unmanned aerial vehicle elastic computing method based on fog and cloud environment - Google Patents
Unmanned aerial vehicle elastic computing method based on fog and cloud environment Download PDFInfo
- Publication number
- CN111723985A CN111723985A CN202010554363.5A CN202010554363A CN111723985A CN 111723985 A CN111723985 A CN 111723985A CN 202010554363 A CN202010554363 A CN 202010554363A CN 111723985 A CN111723985 A CN 111723985A
- Authority
- CN
- China
- Prior art keywords
- task
- unmanned aerial
- scheme
- aerial vehicle
- fog
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06316—Sequencing of tasks or work
Abstract
An unmanned aerial vehicle elastic computing method based on a fog cloud environment comprises an unmanned aerial vehicle executing computing task scheduling process. In a system environment with the coexistence of fog computing and cloud computing, the scheduling center provides a corresponding resource scheduling strategy for the new request according to the type of the new request, namely, the unmanned aerial vehicle is used as the supplement of the computing capability or the communication capability in the system, so that various computing tasks can be completed before the deadline date, and the cost is reduced as much as possible.
Description
Technical Field
The invention relates to the fields of fog computing, cloud computing, unmanned aerial vehicles and scheduling optimization, in particular to an unmanned aerial vehicle elastic computing method based on a fog and cloud environment.
Background
With the continuous development of science and technology and the continuous progress of technology, the unmanned aerial vehicle is widely applied to the fields of disaster relief, surveying and mapping, aerial photography, intelligent manufacturing and the like with the advantages of rapidness, mobility, flexibility and the like. For example, in the aspect of disaster relief, forests are precious resources in the nature, and the research of forest fire detection technology has very important significance in forestry. Drones are increasingly used for forest fire monitoring and detection, with their high mobility and the ability to cover areas of different heights and positions at relatively low cost. On the other hand, it is also a profit and nothing more to use a drone instead of a human to do these high-risk tasks. In the field of intelligent manufacturing, the manufacturing industry takes the internet as infrastructure, realizes wide interconnection-throughout all links of manufacturing activities such as design, production, management, service and the like, obviously improves the production efficiency in the industry, solves the problem of uneven quality of products produced by manual work in the past, and ensures the quality and uniformity of products leaving factories.
The unmanned aerial vehicle has advantages in the background environment with large processing range, large data processing capacity and high network bandwidth occupancy rate, and how to make the unmanned aerial vehicle fully play the functions in the environment with fog computing and cloud computing at the same time is a considerable problem. In the cloud environment, the cloud nodes, the fog nodes and the unmanned aerial vehicles are sorted from large to small according to the computing capacity, the unmanned aerial vehicles, the fog nodes and the cloud nodes are sorted from large to small according to the communication capacity, and the cloud nodes (cloud server leasing and flow cost) with the largest cost consumption are selected. How to fully utilize the computing capacity and the communication capacity of the unmanned aerial vehicle to be matched with the fog nodes and the cloud nodes in the environment system so as to reduce the cost as much as possible while finishing the task request within the deadline time and achieve the win-win of efficiency economy, which is a problem to be solved urgently.
At present, few researches are conducted on the problem of flexible computing scheduling of unmanned aerial vehicles in a cloud environment, and the unmanned aerial vehicles are mostly only considered to be introduced into a cloud computing background. With the rapid development of the internet of things and 5G, the idea that the unmanned aerial vehicle is introduced into the cloud environment as a new computing resource is more practical and feasible, and therefore how to complete scheduling under the constraint condition by using cooperation of the three is a research focus.
Disclosure of Invention
Aiming at the problems in the background technology, the invention provides an unmanned aerial vehicle elastic computing method based on a fog and cloud environment.
An unmanned aerial vehicle elastic computing method based on a fog cloud environment comprises the following steps:
s01: the dispatching center receives a batch of new task requests collected by the data source, analyzes whether each request is a calculation intensive task or a data intensive task, stores the request into a new task set D1, and sets an initial value of task data freshness; simultaneously monitoring the task execution conditions in the set D2 to be secondarily optimized, namely simultaneously executing S02 and S12;
s02: the dispatching center judges whether the new task set D1 is empty, if the new task set D1 is empty, S01 is executed, and if not, S03 is executed;
s03: the scheduling center sets priorities for the tasks according to a given sequencing strategy, and the higher the task priority is, the higher the priority of the initial task allocation scheme is;
s04: the dispatching center judges all the calculation intensive tasks analyzed before and executes S06, and the rest are not calculation intensive tasks and execute S07;
s05: the dispatching center generates an initial task allocation scheme for all analyzed calculation intensive tasks, wherein a fog task allocation sub-scheme refers to allocating all calculation data to nearby corresponding fog equipment for processing; the unmanned aerial vehicle task allocation sub-scheme is that an unmanned aerial vehicle is used as a supplement to the computing power of fog equipment, and one or more unmanned aerial vehicles which are close to the fog equipment and have sufficient electric quantity are arranged by a dispatching center and fly to a specified place near the fog equipment to undertake the rest computing tasks; the cloud task allocation sub-scheme is that when the requirements of the deadline time cannot be met after the fog task allocation sub-scheme and the unmanned aerial vehicle task allocation sub-scheme are both scheduled and allocated, the rest data are sent to the cloud end for processing, and then S07 is executed;
s06: the dispatching center generates an initial task allocation scheme for all analyzed data intensive tasks, wherein an unmanned aerial vehicle task allocation sub-scheme refers to that an unmanned aerial vehicle is used as equipment with stronger communication capacity, and the dispatching center arranges one or more unmanned aerial vehicles which are close to the data source and have sufficient electric quantity to fly to a specified place near the data source to undertake calculation tasks; the fog task allocation sub-scheme is used for allocating the rest part of calculation data to nearby corresponding fog equipment for processing; the cloud task allocation sub-scheme is that when the requirements of the deadline time cannot be met after the fog task allocation sub-scheme and the unmanned aerial vehicle task allocation sub-scheme are both scheduled and allocated, the rest data are sent to the cloud end for processing, and then S07 is executed;
s07: the dispatching center sends an initial task allocation scheme to each data source, the fog equipment and the unmanned aerial vehicle, plans a path for the unmanned aerial vehicle, generates an unmanned aerial vehicle task scheme and updates available resource information in real time;
s08: the dispatching center assists each unmanned aerial vehicle to arrive at a designated place, monitors the condition that a data source, the fog equipment and the unmanned aerial vehicle execute a task allocation scheme, and moves the task out of a new task set D1;
s09: the dispatching center monitors whether all tasks have tasks with higher overdue risks, if not, the dispatching center executes S02, and if yes, the dispatching center executes S10;
s10: the scheduling center updates the data freshness of the tasks which are not executed according to the schedule, stores the initial task allocation scheme into the set D2 to be secondarily optimized, and then executes S01;
s11, the scheduling center judges whether the set D2 to be secondarily optimized is empty, if the set D2 is empty, S01 is executed, and if the set D2 is not empty, S12 is executed;
s12, if the freshness of the task data is lower than the threshold value, the scheduling center discards the task and releases the resource; the priority of the scheme to be optimized is set by the other tasks according to the data freshness from low to high, and the lower the data freshness, the higher the priority of the scheme to be optimized;
s13, the dispatching center judges all the calculation intensive tasks analyzed before and executes S14, and the rest are not calculation intensive tasks and execute S15;
s14, when detecting all tasks with the data transmission time being out of date when executing the calculation intensive tasks, the scheduling center arranges one or more unmanned aerial vehicles with idle time and sufficient electric quantity close to the data source to fly to a specified place near the data source, plans a path for the unmanned aerial vehicles, updates the unmanned aerial vehicle task scheme, updates the available information of resources in real time, and then executes S16;
s15, when detecting that all tasks with the calculation time exceeding the time period are executed during the execution of the data intensive tasks, the scheduling center arranges one or more unmanned aerial vehicles with idle time periods and sufficient electric quantity close to the fog equipment to fly to a specified place near the fog equipment, plans a path for the unmanned aerial vehicles, updates the unmanned aerial vehicle task scheme, updates the available resource information in real time, and then executes S16;
and S16, the dispatching center sends the secondary optimization scheme to the corresponding data source and the unmanned aerial vehicle or the corresponding fog equipment and the unmanned aerial vehicle, then moves out of the D2, monitors the execution condition of the secondary optimization scheme, and then executes S01 for circulation.
Further, in S01, the scheduling center receives a batch of new production task request information collected by data sources such as sensors and cameras on production equipment in the area, and one piece of production task request information includes information of a request task ID, a data source number ID, an analysis task type, a corresponding fog equipment number ID, task data information, a deadline, and data freshness after being analyzed; the scheduling center maintains an idle new task set table and stores newly arrived production task request information into the table; meanwhile, the scheduling center also maintains an available resource information table, wherein the available resource information table comprises an equipment number ID, position information and resource utilization information, a data source is represented by the beginning of D in the equipment number ID, the fog equipment is represented by the beginning of F, the unmanned aerial vehicle is represented by the beginning of U, and the cloud server is represented by the beginning of C; if the device number ID is a foggy device, the resource utilization information should only record the resource monitor information, and if the device number ID is an unmanned aerial vehicle, the resource utilization information should record the resource monitor information and the remaining flight time at the same time.
Further, in S01, the scheduling center stores each production task request into a new task set table, and the scheduling center sets an initial value of data freshness of 100% for each task, and the data freshness calculation method introduces a data freshness calculation formula, that is, a ratio of a difference between a deadline time and an executed time to the deadline time.
Further, in S03, the scheduling center sets the priority of the initial task allocation scheme according to the descending order of the deadline of each request task, where the priority ordering table needs to include the priority and the request task ID, and the smaller the deadline is, the higher the priority of the initial task allocation scheme is.
Further, in S03, the scheduling center sequentially generates an initial task allocation scheme for each task according to the available state of the current resource and the task priority, where the initial task allocation scheme includes a fog task allocation sub-scheme, an unmanned aerial vehicle task allocation sub-scheme, and a cloud task allocation sub-scheme.
Further, in S05 and S06, the scheduling center sequentially generates a fog task allocation sub-scheme, an unmanned aerial vehicle task allocation sub-scheme, and a cloud task allocation sub-scheme when generating an initial task allocation scheme for the compute-intensive task; when an initial task allocation scheme is generated for the data intensive tasks, an unmanned aerial vehicle task allocation sub-scheme, a fog task allocation sub-scheme and a cloud task allocation sub-scheme are sequentially generated; the generated initial task allocation scheme needs to include information such as a request task ID, a first processing device ID, a percentage of tasks undertaken by the first processing device, a second processing device ID, a percentage of tasks undertaken by the second processing device, a third processing device ID, a percentage of tasks undertaken by the third processing device, a predicted task completion duration, and the like.
Further, in S09, the scheduling center performs overdue risk assessment on all executing tasks, which is to calculate a processing speed according to the data amount and processing time that have been processed by the task, so as to determine whether the remaining data amount can be completed on time; the calculation method of the overdue risk value is the ratio of the overdue time length to the expected task completion time length, and when the overdue risk value is larger than a certain threshold value, the task is considered to have higher overdue risk.
Further, in S12, if the scheduling center finds that the data freshness threshold in the to-be-twice-optimized set D2 is lower than 5%, the task is discarded and all occupied resources of the task are released, and for the to-be-twice-optimized tasks whose data freshness thresholds are higher than 5%, the to-be-twice-optimized tasks are sorted from low to high according to the data freshness, and the order is the priority of the to-be-optimized scheme, so as to generate a to-be-twice-optimized task priority table, where the to-be-twice-optimized task priority table needs to include information such as priority, request task ID, and updated data freshness.
Further, the unmanned aerial vehicle executes an elastic calculation process in the operation process as follows:
the unmanned aerial vehicle receives the task allocation scheme and the unmanned aerial vehicle task scheme, goes to a first appointed place according to an appointed route in the scheme, and sends an arrival instruction to the dispatching center after arriving at the appointed place;
if the unmanned aerial vehicle arrives at the designated place on time, an arrival instruction is sent to the dispatching center, the corresponding data source or the fog equipment is connected to receive the calculation data, and a task allocation scheme is executed;
if the unmanned aerial vehicle does not arrive at the designated place on time or does not complete the calculation task on schedule, sending a message of not executing the task allocation scheme on schedule to the dispatching center, and waiting for the dispatching center to send the updated scheme;
if the unmanned aerial vehicle finishes the task allocation scheme, judging whether all the allocated tasks are finished according to a plan according to the unmanned aerial vehicle task scheme, if all the allocated tasks are not finished, sending a task completion response to the dispatching center by the unmanned aerial vehicle, and flying to the next designated place according to a designated route in the unmanned aerial vehicle task scheme updated in real time;
if the unmanned aerial vehicle finishes all the distributed tasks in the unmanned aerial vehicle task scheme, detecting the electric quantity of the unmanned aerial vehicle again, if the unmanned aerial vehicle detects that the electric quantity is lower than a threshold value, sending a charging instruction to a dispatching center, flying to a charging pile closest to the unmanned aerial vehicle for charging, and waiting for the dispatching center to send a new task distribution scheme and an unmanned aerial vehicle task scheme after charging is finished; otherwise, the scheduling center is directly waited to send a new task allocation scheme and an unmanned aerial vehicle task scheme.
The invention has the beneficial effects that: the invention provides an unmanned aerial vehicle elastic computing method based on a fog cloud environment, which comprises an elastic fog computing process and an unmanned aerial vehicle computing task executing process, wherein the unmanned aerial vehicle is used as a supplement of computing capability or a supplement of communication capability in a system, so that various computing tasks can be completed before an expiration date, and cost optimization and implementation are reduced as much as possible.
Drawings
Fig. 1 is a schematic diagram of a scheduling process of a scheduling center in an embodiment of the present invention.
Fig. 2 is a schematic flow chart of the unmanned aerial vehicle executing the calculation task in the embodiment of the present invention.
Fig. 3 is a schematic diagram of elasticity calculation performed when the unmanned aerial vehicle moves in the embodiment of the present invention.
FIG. 4 is a table of new task sets in an embodiment of the invention.
Fig. 5 is a table of available resource information in an embodiment of the present invention.
Fig. 6 is a priority ranking table in an embodiment of the invention.
FIG. 7 is a table of task allocation plans in an embodiment of the invention.
Fig. 8 is a table of the mission plan of the drone in an embodiment of the invention.
FIG. 9 is a task priority table to be secondarily optimized according to an embodiment of the present invention.
FIG. 10 is a table of secondary optimization task allocation schemes after updating according to an embodiment of the present invention.
Fig. 11 is an updated task plan table for the drone in an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is further explained in detail by combining the drawings in the specification.
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention and by taking intelligent manufacturing in a factory as an example, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without undue effort, are within the scope of the invention.
Fig. 1 is a schematic diagram of a scheduling flow of a scheduling center of an unmanned aerial vehicle flexible computing method, specifically including the following steps:
s01: the dispatching center receives a batch of new task requests collected by the data source, analyzes whether each request is a calculation intensive task or a data intensive task, stores the request into a new task set D1, and sets an initial value of task data freshness; and simultaneously monitoring the task execution conditions in the set D2 to be secondarily optimized, namely simultaneously executing S02 and S12.
The scheduling center receives a batch of new production task request information acquired by data sources (such as sensors, cameras and the like) on production equipment in an area, and one piece of production task request information should contain information such as a request task ID, a data source number ID, an analysis task type, a corresponding fog equipment number ID, task data information, expiration time, data freshness and the like after being analyzed. The scheduling center maintains an idle new task set table and stores newly arrived production task request information into the table; meanwhile, the scheduling center also maintains an available resource information table which should contain the equipment number ID, the position information and the resource utilization information, wherein the data source is represented by the beginning D in the equipment number ID, the fog equipment is represented by the beginning F, the unmanned aerial vehicle is represented by the beginning U, and the cloud server is represented by the beginning C; if the equipment number ID is fog equipment, the resource utilization information only needs to record resource monitor information, and if the equipment number ID is an unmanned aerial vehicle, the resource utilization information simultaneously records the resource monitor information and the remaining flight time; the location information is shown in fig. 3.
The scheduling center stores each production task request into a new task set table, the scheduling center sets a data freshness initial value of 100% for each task, and the data freshness calculation method can introduce a data freshness calculation formula, namely a ratio of difference between a cut-off duration and an executed duration to the cut-off duration. In this example, it is specified that when the data freshness is lower than 5%, the task data is regarded as stale data, and the task is discarded and all the resources occupied by the task are released.
For example, the scheduling center stores the request information of the new production Task in the new Task set table as shown in fig. 4, and the information of the currently available resources is shown in fig. 5 (for simplicity, the specific Task Data information of the request Task is represented by Task1 Data and Task2Data … …).
S02: the dispatching center judges whether the new task set D1 is empty, if the new task set D1 is empty, S01 is executed, and if not, S03 is executed.
If the new task set table (fig. 4) is not empty, S03 is executed.
S03: the scheduling center sets priorities for the tasks according to a given sorting strategy, and the higher the task priority is, the higher the priority of the initial task allocation scheme is.
The scheduling center can set the priority of the initial task allocation scheme according to the sequence of the deadline time of each request task from small to large, and the smaller the deadline time is, the higher the priority of the initial task allocation scheme is.
The priority ranking table needs to contain the priority, the requesting task ID. The resulting priority ranking table in this example is shown in fig. 6.
And the scheduling center sequentially generates an initial task allocation scheme for each task according to the available state of the current resources and the task priority level, wherein the initial task allocation scheme comprises a fog task allocation sub-scheme, an unmanned aerial vehicle task allocation sub-scheme and a cloud task allocation sub-scheme.
The scheduling center sequentially generates initial Task allocation schemes for Task2 and Task1 according to the available state (figure 5) of the current resources and the Task priority (figure 6), wherein the initial Task allocation schemes comprise a fog Task allocation sub-scheme, an unmanned aerial vehicle Task allocation sub-scheme and a cloud Task allocation sub-scheme.
S04: the dispatch center determines all the compute intensive tasks previously resolved and executes S06, while the rest are not compute intensive and execute S07.
According to FIG. 4, the requesting Task1 is parsed into a compute intensive Task, S05 is performed; the requesting Task2 resolves to a data intensive Task, executing S06.
S05: the dispatching center generates an initial task allocation scheme for all analyzed calculation intensive tasks, wherein a fog task allocation sub-scheme refers to allocating all calculation data to nearby corresponding fog equipment for processing; the unmanned aerial vehicle task allocation sub-scheme is that an unmanned aerial vehicle is used as a supplement to the computing power of fog equipment, and one or more unmanned aerial vehicles which are close to the fog equipment and have sufficient electric quantity are arranged by a dispatching center and fly to a specified place near the fog equipment to undertake the rest computing tasks; the cloud task allocation sub-scheme refers to sending the remaining data to the cloud end for processing when the requirements of the deadline time cannot be met after the fog task allocation sub-scheme and the unmanned aerial vehicle task allocation sub-scheme are both scheduled and allocated, and then executing S07.
The dispatching center generates an initial Task allocation scheme for the Task1, firstly generates a fog Task allocation sub-scheme, and predicts whether all calculation data are sent to the corresponding fog equipment F01 for calculation, and if the prediction result is that the F01 can bear 60% of the calculation Task of the Task1 at most, the rest 40% of calculation amount is allocated to the unmanned aerial vehicle for processing. Then, an unmanned aerial vehicle task allocation sub-scheme is generated, the dispatching center finds that the unmanned aerial vehicle which is close to the F01 and has enough electric quantity and idle time periods has U01 and U02, the two unmanned aerial vehicles just can meet the residual task requirements, the deadline is met, and the cloud task allocation sub-scheme does not need to be arranged. The scheduling center stores the planned initial task assignment into the task assignment table, and then performs S07.
The task allocation table needs to include a request task ID, a first processing device ID, a percentage of tasks undertaken by the first processing device, a second processing device ID, a percentage of tasks undertaken by the second processing device, a third processing device ID, a percentage of tasks undertaken by the third processing device, and a predicted task completion duration. The estimated task completion time is the maximum value of the fog task allocation sub-scheme completion time, the unmanned aerial vehicle task allocation sub-scheme completion time and the cloud task allocation sub-scheme completion time; the fog task allocation sub-scheme completion time length and the cloud task allocation sub-scheme completion time length comprise data transmission time length and calculation time length; the unmanned aerial vehicle task allocation sub-scheme completion time comprises data transmission time, calculation time and unmanned aerial vehicle flight time. The task allocation plan table for this example is shown in FIG. 7.
S06: the dispatching center generates an initial task allocation scheme for all analyzed data intensive tasks, wherein an unmanned aerial vehicle task allocation sub-scheme refers to that an unmanned aerial vehicle is used as equipment with stronger communication capacity, and the dispatching center arranges one or more unmanned aerial vehicles which are close to the data source and have sufficient electric quantity to fly to a specified place near the data source to undertake calculation tasks; the fog task allocation sub-scheme is used for allocating the rest part of calculation data to nearby corresponding fog equipment for processing; the cloud task allocation sub-scheme refers to sending the rest data to the cloud end for processing when the requirements of the deadline time cannot be met after the fog task allocation sub-scheme and the unmanned aerial vehicle task allocation sub-scheme are both scheduled and allocated.
The scheduling center generates an initial Task allocation scheme for the Task2, firstly, an unmanned aerial vehicle Task allocation sub-scheme is generated, the scheduling center predicts whether the unmanned aerial vehicle which sends all calculation data to the data source D02 and has enough electric quantity and free time period is feasible or not, and the prediction result is that only the unmanned aerial vehicle U03 can undertake 30% of the calculation Task of the Task 2. Then, a fog Task allocation sub-scheme is generated, the corresponding fog equipment F02 is expected to bear 55% of the Task2 computing tasks at most in the deadline, and the remaining 15% of the computing amount is allocated to the cloud end C01 for processing, so that the cloud Task allocation sub-scheme is generated. And the scheduling center stores the planned initial task allocation scheme into a task allocation scheme table. The task allocation plan table for this example is shown in FIG. 7.
S07: and the dispatching center sends an initial task allocation scheme to each data source, the fog equipment and the unmanned aerial vehicle, plans a path for the unmanned aerial vehicle, generates an unmanned aerial vehicle task scheme and updates the available resource information in real time.
And the scheduling center sends the initial Task distribution scheme of the request Task2 to U03, F02 and C01 according to the Task distribution scheme in the Task distribution table, and sends the initial Task distribution scheme of the request Task1 to F01, U01 and U02. Meanwhile, the scheduling center respectively generates planning paths for the unmanned aerial vehicles U01, U02 and U03 according to the current positions and the target positions of the unmanned aerial vehicles, and the planning paths are stored in an unmanned aerial vehicle task scheme table. The unmanned aerial vehicle task plan table needs to contain information such as an unmanned aerial vehicle device number ID, a destination number ID, multipoint route planning, predicted flight duration and the like. The planned route may have multiple destinations in turn for the drone in a batch of missions, and therefore multiple sub-routes. The mission plan table of the drone in this example is shown in fig. 8 (for simplicity, the multipoint route planning of the drone is represented by U01 route planning, U02 route planning … …, and the specific route is shown in example fig. 3).
S08: the dispatch center assists each drone in reaching the designated site and monitors the data source, the fog device, and the drone for the performance of the task assignment scheme and moves the task out of the new task set D1.
The dispatching center assists the unmanned aerial vehicle U03, the unmanned aerial vehicle U01 and the unmanned aerial vehicle U02 to reach a designated place along the multipoint path plan generated in the unmanned aerial vehicle Task scheme table, simultaneously monitors the condition that each data source, the fog device and the unmanned aerial vehicle execute each Task allocation scheme, and moves the Task1 and the Task2 out of the new Task table T (figure 4).
S09: the dispatch center monitors whether there are all tasks with higher risk of overrun, and if not, executes S02, and if so, executes S10.
The scheduling center monitors whether the tasks Task1 and Task2 have high overdue risks or not, and calculates the processing speed according to the data volume and the processing time which are already processed by the tasks Task1 and Task2 so as to judge whether the residual data volume can be completed on time or not. The calculation method of the overdue risk value can be the ratio of the overdue time length to the expected task completion time length, and when the overdue risk value is greater than 20%, the task is considered to have higher overdue risk. And monitoring and finding that the data transmission time of the Task1 exceeds the period in the Task execution process, the newly predicted Task completion time is 4.8 minutes, the calculation time of the Task2 exceeds the period, and the newly predicted Task completion time is 3.1 minutes, which have higher time-out risks, and executing S10.
S10: the dispatching center updates the data freshness of the tasks which are not executed according to the schedule, stores the initial task allocation scheme into the set D2 to be secondarily optimized, and then executes S01.
And the scheduling center calculates and updates the data freshness of the Task1 and the Task2 according to a data freshness operation formula, stores the initial Task allocation scheme of the tasks Task1 and Task2 in the Task allocation scheme table into the set to be secondarily optimized D2, and returns to execute S01.
S11: and the dispatching center judges whether the set D2 to be secondarily optimized is empty, if the D2 is empty, S01 is executed, and if the D2 is not empty, S12 is executed.
And the dispatching center constantly monitors the conditions of the new task set and the to-be-secondarily-optimized set D2, and executes S12 when the dispatching center finds that the secondary optimized set D2 is not empty.
S12: if the freshness of the task data is lower than the threshold value, the scheduling center discards the task and releases resources; and the other tasks set the priority of the scheme to be optimized according to the data freshness from low to high, and the lower the data freshness, the higher the priority of the scheme to be optimized.
If the dispatch center finds that the data freshness threshold value in the to-be-quadratic optimization set D2 is lower than 5%, the task is discarded and all the occupied resources are released. And for the tasks to be secondarily optimized with the rest data freshness threshold higher than 5%, sorting the tasks from low to high according to the data freshness, and taking the sorting as the priority of the scheme to be optimized. At present, the distribution schemes of the Task1 and the Task2 in the secondary set D2 to be optimized are to be optimized, and a Task priority table to be optimized for the second time is generated.
The task priority table to be optimized for the second time needs to contain information such as priority, task ID request, and data freshness of update. The priority table of the task to be secondarily optimized obtained in this example is shown in fig. 9.
S13: the dispatch center determines all the compute intensive tasks previously resolved and executes S14, while the rest are not compute intensive and execute S15.
The scheduling center judges whether all tasks requiring secondary optimization are calculation-intensive tasks or data-intensive tasks according to fig. 9 and the production task request information. The request Task1 is parsed into a compute intensive Task, and S14 is executed; the requesting Task1 resolves to a data intensive Task, executing S15.
S14: when detecting all tasks with the data transmission time exceeding the period during the execution of the calculation intensive tasks, the scheduling center arranges one or more unmanned aerial vehicles with idle time periods and sufficient electric quantity close to the data source to fly to a specified place near the data source, plans a path for the unmanned aerial vehicles, updates the unmanned aerial vehicle task scheme, updates the resource available information in real time, and then executes S16.
The scheduling center re-allocates Task scheduling for Task 1. The scheduling center finds that the data transmission time of the Task1 exceeds the period during execution, on the basis of an initial Task allocation scheme, checks whether an unmanned aerial vehicle with an idle time period and sufficient electric quantity exists near a data source D01, checks that the unmanned aerial vehicle U04 meets the requirements, plans a path for the unmanned aerial vehicle U04, updates an unmanned aerial vehicle Task allocation table, updates available resource information in real time, and then executes S16.
The secondary optimization task allocation plan table and the unmanned aerial vehicle task plan table after the update are respectively shown in fig. 10 and fig. 11.
S15: when detecting all tasks with the calculation time exceeding when executing the data intensive tasks, the scheduling center arranges one or more unmanned aerial vehicles with idle time periods and sufficient electric quantity close to the fog equipment to fly to a specified place near the fog equipment, plans a path for the unmanned aerial vehicles, updates an unmanned aerial vehicle task scheme and updates available resource information in real time.
The scheduling center re-allocates Task scheduling for Task 2. The scheduling center finds that the time of the Task2 is out of date when the Task is executed, on the basis of an initial Task allocation scheme, whether an unmanned aerial vehicle with an idle time period and sufficient electric quantity exists near the fog equipment F02 or not is checked, the unmanned aerial vehicle U01 is found to meet the requirements, the scheduling center plans a path for the unmanned aerial vehicle U01, the unmanned aerial vehicle Task allocation table is updated, and the resource available information is updated in real time.
The secondary optimization task allocation plan table and the unmanned aerial vehicle task plan table after the update are respectively shown in fig. 10 and fig. 11.
S16: and the dispatching center sends the secondary optimization scheme to the corresponding data source and the unmanned aerial vehicle or the corresponding fog equipment and the unmanned aerial vehicle, moves out of the D2, monitors the execution condition of the secondary optimization scheme, and then executes S01.
The dispatching center sends the updated secondary optimization scheme (figure 10) and the updated unmanned aerial vehicle Task scheme table (figure 11) to the corresponding data source and the unmanned aerial vehicle or the corresponding fog equipment and the unmanned aerial vehicle, namely, the updated Task distribution scheme of the request Task1 is sent to the data source D01 and the unmanned aerial vehicle U04; the updated Task assignment plan requesting the Task2 is sent to the fogger F02 and the drone U01. And then, the optimization solution is removed from the set D2 to be secondarily optimized, the execution condition of the secondary optimization scheme is monitored, and the execution is returned to S01.
The above description is only a preferred embodiment of the present invention, and the scope of the present invention is not limited to the above embodiment, but equivalent modifications or changes made by those skilled in the art according to the present disclosure should be included in the scope of the present invention as set forth in the appended claims.
Claims (9)
1. An unmanned aerial vehicle elastic computing method based on a fog and cloud environment is characterized in that: the calculation method comprises the following steps:
s01: the dispatching center receives a batch of new task requests collected by the data source, analyzes whether each request is a calculation intensive task or a data intensive task, stores the request into a new task set D1, and sets an initial value of task data freshness; simultaneously monitoring the task execution conditions in the set D2 to be secondarily optimized, namely simultaneously executing S02 and S12;
s02: the dispatching center judges whether the new task set D1 is empty, if the new task set D1 is empty, S01 is executed, and if not, S03 is executed;
s03: the scheduling center sets priorities for the tasks according to a given sequencing strategy, and the higher the task priority is, the higher the priority of the initial task allocation scheme is;
s04: the dispatching center judges all the calculation intensive tasks analyzed before and executes S06, and the rest are not calculation intensive tasks and execute S07;
s05: the dispatching center generates an initial task allocation scheme for all analyzed calculation intensive tasks, wherein a fog task allocation sub-scheme refers to allocating all calculation data to nearby corresponding fog equipment for processing; the unmanned aerial vehicle task allocation sub-scheme is that an unmanned aerial vehicle is used as a supplement to the computing power of fog equipment, and one or more unmanned aerial vehicles which are close to the fog equipment and have sufficient electric quantity are arranged by a dispatching center and fly to a specified place near the fog equipment to undertake the rest computing tasks; the cloud task allocation sub-scheme is that when the requirements of the deadline time cannot be met after the fog task allocation sub-scheme and the unmanned aerial vehicle task allocation sub-scheme are both scheduled and allocated, the rest data are sent to the cloud end for processing, and then S07 is executed;
s06: the dispatching center generates an initial task allocation scheme for all analyzed data intensive tasks, wherein an unmanned aerial vehicle task allocation sub-scheme refers to that an unmanned aerial vehicle is used as equipment with stronger communication capacity, and the dispatching center arranges one or more unmanned aerial vehicles which are close to the data source and have sufficient electric quantity to fly to a specified place near the data source to undertake calculation tasks; the fog task allocation sub-scheme is used for allocating the rest part of calculation data to nearby corresponding fog equipment for processing; the cloud task allocation sub-scheme is that when the requirements of the deadline time cannot be met after the fog task allocation sub-scheme and the unmanned aerial vehicle task allocation sub-scheme are both scheduled and allocated, the rest data are sent to the cloud end for processing, and then S07 is executed;
s07: the dispatching center sends an initial task allocation scheme to each data source, the fog equipment and the unmanned aerial vehicle, plans a path for the unmanned aerial vehicle, generates an unmanned aerial vehicle task scheme and updates available resource information in real time;
s08: the dispatching center assists each unmanned aerial vehicle to arrive at a designated place, monitors the condition that a data source, the fog equipment and the unmanned aerial vehicle execute a task allocation scheme, and moves the task out of a new task set D1;
s09: the dispatching center monitors whether all tasks have tasks with higher overdue risks, if not, the dispatching center executes S02, and if yes, the dispatching center executes S10;
s10: the scheduling center updates the data freshness of the tasks which are not executed according to the schedule, stores the initial task allocation scheme into the set D2 to be secondarily optimized, and then executes S01;
s11: the dispatching center judges whether the set D2 to be secondarily optimized is empty, if the set D2 is empty, S01 is executed, and if the set D2 is not empty, S12 is executed;
s12: if the freshness of the task data is lower than the threshold value, the scheduling center discards the task and releases resources; the priority of the scheme to be optimized is set by the other tasks according to the data freshness from low to high, and the lower the data freshness, the higher the priority of the scheme to be optimized;
s13: the dispatching center judges all the calculation intensive tasks analyzed before and executes S14, and the rest are not calculation intensive tasks and execute S15;
s14: when detecting all tasks with the data transmission time exceeding the period during the execution of the calculation intensive tasks, the scheduling center arranges one or more unmanned aerial vehicles with idle time periods and sufficient electric quantity close to the data source to fly to a specified place near the data source, plans a path for the unmanned aerial vehicles, updates the unmanned aerial vehicle task scheme, updates the resource available information in real time, and then executes S16;
s15: when detecting that all tasks with the time exceeding the calculation time are executed during the execution of the data intensive tasks, the scheduling center arranges one or more unmanned aerial vehicles with idle time periods and sufficient electric quantity close to the fog equipment to fly to a specified place near the fog equipment, plans a path for the unmanned aerial vehicles, updates an unmanned aerial vehicle task scheme, updates available resource information in real time, and then executes S16;
s16: and the dispatching center sends the secondary optimization scheme to the corresponding data source and the unmanned aerial vehicle or the corresponding fog equipment and the unmanned aerial vehicle, moves out of the D2, monitors the execution condition of the secondary optimization scheme, and then executes S01 for circulation.
2. The unmanned aerial vehicle elastic computing method based on the fog and cloud environment as claimed in claim 1, wherein: in S01, a scheduling center receives a batch of new production task request information collected by data sources such as sensors and cameras on production equipment in an area, and one piece of production task request information contains information of a request task ID, a data source number ID, an analysis task type, a corresponding fog equipment number ID, task data information, a cutoff time and data freshness after being analyzed; the scheduling center maintains an idle new task set table and stores newly arrived production task request information into the table; meanwhile, the scheduling center also maintains an available resource information table, wherein the available resource information table comprises an equipment number ID, position information and resource utilization information, a data source is represented by the beginning of D in the equipment number ID, the fog equipment is represented by the beginning of F, the unmanned aerial vehicle is represented by the beginning of U, and the cloud server is represented by the beginning of C; if the device number ID is a foggy device, the resource utilization information should only record the resource monitor information, and if the device number ID is an unmanned aerial vehicle, the resource utilization information should record the resource monitor information and the remaining flight time at the same time.
3. The unmanned aerial vehicle elastic computing method based on the fog and cloud environment as claimed in claim 1, wherein: in S01, the scheduling center stores each production task request into a new task set table, and the scheduling center sets an initial data freshness value of 100% for each task, and the data freshness calculation method introduces a data freshness calculation formula, that is, a ratio of a difference between a deadline time and an executed time to the deadline time.
4. The unmanned aerial vehicle elastic computing method based on the fog and cloud environment as claimed in claim 1, wherein: in S03, the scheduling center sets the priority of the initial task allocation scheme according to the descending order of the deadline of each request task, the priority order table needs to include the priority and the request task ID, and the smaller the deadline is, the higher the priority of the initial task allocation scheme is.
5. The unmanned aerial vehicle elastic computing method based on the fog and cloud environment as claimed in claim 1, wherein: in the step S03, the scheduling center sequentially generates an initial task allocation scheme for each task according to the available state of the current resources and the task priority level, wherein the initial task allocation scheme comprises a fog task allocation sub-scheme, an unmanned aerial vehicle task allocation sub-scheme and a cloud task allocation sub-scheme.
6. The unmanned aerial vehicle elastic computing method based on the fog and cloud environment as claimed in claim 1, wherein: in S05 and S06, the scheduling center generates a fog task allocation sub-scheme, an unmanned aerial vehicle task allocation sub-scheme and a cloud task allocation sub-scheme in sequence when generating an initial task allocation scheme for the compute-intensive task; when an initial task allocation scheme is generated for the data intensive tasks, an unmanned aerial vehicle task allocation sub-scheme, a fog task allocation sub-scheme and a cloud task allocation sub-scheme are sequentially generated; the generated initial task allocation scheme needs to include information such as a request task ID, a first processing device ID, a percentage of tasks undertaken by the first processing device, a second processing device ID, a percentage of tasks undertaken by the second processing device, a third processing device ID, a percentage of tasks undertaken by the third processing device, a predicted task completion duration, and the like.
7. The unmanned aerial vehicle elastic computing method based on the fog and cloud environment as claimed in claim 1, wherein: in S09, the scheduling center carries out overdue risk assessment on all executing tasks, and calculates the processing speed according to the data volume and the processing time processed by the tasks, so as to judge whether the residual data volume can be completed in time; the calculation method of the overdue risk value is the ratio of the overdue time length to the expected task completion time length, and when the overdue risk value is larger than a certain threshold value, the task is considered to have higher overdue risk.
8. The unmanned aerial vehicle elastic computing method based on the fog and cloud environment as claimed in claim 1, wherein: in S12, if the scheduling center finds that the data freshness threshold in the to-be-secondarily-optimized set D2 is lower than 5%, the task is discarded and all occupied resources of the task are released, and for the to-be-secondarily-optimized tasks whose data freshness thresholds are higher than 5%, the to-be-secondarily-optimized tasks are sorted from low to high according to the data freshness, and the order is the priority of the to-be-optimized scheme, so as to generate a to-be-secondarily-optimized task priority table, where the to-be-secondarily-optimized task priority table needs to include information such as priority, request task ID, and updated data freshness.
9. The unmanned aerial vehicle elastic computing method based on the fog and cloud environment as claimed in claim 1, wherein: the unmanned aerial vehicle executes an elastic calculation process in the operation process, and the elastic calculation process comprises the following steps:
the unmanned aerial vehicle receives the task allocation scheme and the unmanned aerial vehicle task scheme, goes to a first appointed place according to an appointed route in the scheme, and sends an arrival instruction to the dispatching center after arriving at the appointed place;
if the unmanned aerial vehicle arrives at the designated place on time, an arrival instruction is sent to the dispatching center, the corresponding data source or the fog equipment is connected to receive the calculation data, and a task allocation scheme is executed;
if the unmanned aerial vehicle does not arrive at the designated place on time or does not complete the calculation task on schedule, sending a message of not executing the task allocation scheme on schedule to the dispatching center, and waiting for the dispatching center to send the updated scheme;
if the unmanned aerial vehicle finishes the task allocation scheme, judging whether all the allocated tasks are finished according to a plan according to the unmanned aerial vehicle task scheme, if all the allocated tasks are not finished, sending a task completion response to the dispatching center by the unmanned aerial vehicle, and flying to the next designated place according to a designated route in the unmanned aerial vehicle task scheme updated in real time;
if the unmanned aerial vehicle finishes all the distributed tasks in the unmanned aerial vehicle task scheme, detecting the electric quantity of the unmanned aerial vehicle again, if the unmanned aerial vehicle detects that the electric quantity is lower than a threshold value, sending a charging instruction to a dispatching center, flying to a charging pile closest to the unmanned aerial vehicle for charging, and waiting for the dispatching center to send a new task distribution scheme and an unmanned aerial vehicle task scheme after charging is finished; otherwise, the scheduling center is directly waited to send a new task allocation scheme and an unmanned aerial vehicle task scheme.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010554363.5A CN111723985B (en) | 2020-06-17 | 2020-06-17 | Unmanned aerial vehicle elastic computing method based on fog and cloud environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010554363.5A CN111723985B (en) | 2020-06-17 | 2020-06-17 | Unmanned aerial vehicle elastic computing method based on fog and cloud environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111723985A true CN111723985A (en) | 2020-09-29 |
| CN111723985B CN111723985B (en) | 2022-08-09 |
Family
ID=72567116
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010554363.5A Active CN111723985B (en) | 2020-06-17 | 2020-06-17 | Unmanned aerial vehicle elastic computing method based on fog and cloud environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111723985B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103324525A (en) * | 2013-07-03 | 2013-09-25 | 东南大学 | Task scheduling method in cloud computing environment |
| CN107274053A (en) * | 2017-05-03 | 2017-10-20 | 浙江工商大学 | The wisdom logistics data method for digging dispatched based on mixed cloud |
| CN108737560A (en) * | 2018-05-31 | 2018-11-02 | 南京邮电大学 | Cloud computing task intelligent dispatching method and system, readable storage medium storing program for executing, terminal |
| CN110598951A (en) * | 2019-09-23 | 2019-12-20 | 南京邮电大学 | Mobile charging method for distribution unmanned aerial vehicle |
-
2020
- 2020-06-17 CN CN202010554363.5A patent/CN111723985B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103324525A (en) * | 2013-07-03 | 2013-09-25 | 东南大学 | Task scheduling method in cloud computing environment |
| CN107274053A (en) * | 2017-05-03 | 2017-10-20 | 浙江工商大学 | The wisdom logistics data method for digging dispatched based on mixed cloud |
| CN108737560A (en) * | 2018-05-31 | 2018-11-02 | 南京邮电大学 | Cloud computing task intelligent dispatching method and system, readable storage medium storing program for executing, terminal |
| CN110598951A (en) * | 2019-09-23 | 2019-12-20 | 南京邮电大学 | Mobile charging method for distribution unmanned aerial vehicle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111723985B (en) | 2022-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10474504B2 (en) | Distributed node intra-group task scheduling method and system | |
| CN110297699B (en) | Scheduling method, scheduler, storage medium and system | |
| CN113222305B (en) | Order scheduling method, order scheduling device, storage medium and electronic equipment | |
| CN102724103B (en) | Proxy server, hierarchical network system and distributed workload management method | |
| CN111950870B (en) | Method and system for scheduling data transmission resources of day foundation measurement and control in an integrated manner according to needs | |
| Khochare et al. | Heuristic algorithms for co-scheduling of edge analytics and routes for UAV fleet missions | |
| CN105373426B (en) | A kind of car networking memory aware real time job dispatching method based on Hadoop | |
| Wei et al. | Agent-based simulation for uav swarm mission planning and execution | |
| US11734623B2 (en) | Fleet scheduler | |
| Sorkhoh et al. | Optimizing information freshness for MEC-enabled cooperative autonomous driving | |
| CN112230677A (en) | Unmanned aerial vehicle group task planning method and terminal equipment | |
| Bednowitz et al. | Dispatching and loitering policies for unmanned aerial vehicles under dynamically arriving multiple priority targets | |
| CN111552558A (en) | Scheduling method and device of heterogeneous cloud resources | |
| Raj et al. | Effective cost mechanism for cloudlet retransmission and prioritized VM scheduling mechanism over broker virtual machine communication framework | |
| CN108415760B (en) | Crowd sourcing calculation online task allocation method based on mobile opportunity network | |
| CN111776896B (en) | Elevator dispatching method and device | |
| CN104917839A (en) | Load balancing method for use in cloud computing environment | |
| Alkouz et al. | Provider-centric allocation of drone swarm services | |
| Ullah et al. | LSTPD: least slack time-based preemptive deadline constraint scheduler for Hadoop clusters | |
| WO2018163174A1 (en) | Market equilibrium mechanism for task allocation | |
| CN106407007B (en) | Cloud resource configuration optimization method for elastic analysis process | |
| CN111723985B (en) | Unmanned aerial vehicle elastic computing method based on fog and cloud environment | |
| CN113190342A (en) | Method and system architecture for multi-application fine-grained unloading of cloud-edge cooperative network | |
| CN106534312B (en) | A kind of service request selection of facing mobile apparatus and dispatching method | |
| Xiaohuan et al. | An aggregate flow based scheduler in multi-task cooperated UAVs network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |