Mission planning method for shipborne UAV swarm based on time series mining algorithm

Abstract

A method for planning a shipborne UAV swarm task based on a time series mining algorithm collects multiple simulation operation data of a shipborne UAV swarm, discretizes the position data according to the time data, calculates the profit parameter, obtains the utility data set of the shipborne UAV mounting, and then stores it in an improved cross-linked list data structure, generates a shipborne UAV swarm task mode through high-utility time series data mining, and finally adjusts the simulation parameters according to the shipborne UAV swarm task mode to realize the shipborne UAV swarm task planning assistance. The present invention realizes the optimization of task planning by improving the cross-linked list data structure in conjunction with data mining.

Description

Technical Field

[0001] The present invention relates to a technology in the field of UAV swarm control, specifically a shipborne UAV swarm task planning method based on a time series mining algorithm. Background Art

[0002] Modern information warfare is a confrontation between systems, and it is necessary to build a corresponding combat system through equipment coordination. In modern information warfare, ship-borne drones can detect long-range aerial targets and play an important role in the detection phase of system operations. However, in the context of multi-domain operations, the battlefield environment is more complex, and the detection coverage of a single ship-borne drone is limited, which often cannot meet mission requirements. Summary of the invention

[0003] Aiming at the variable scenarios in which the existing technology is difficult to apply to the coordinated combat missions of shipborne UAV swarms, the present invention proposes a shipborne UAV swarm task planning method based on a time series mining algorithm, and realizes the optimization of task planning by improving the cross-linked list data structure and combining it with data mining.

[0004] The present invention is achieved through the following technical solutions:

[0005] The present invention relates to a method for planning a group of shipborne unmanned aerial vehicles based on a time series mining algorithm, comprising:

[0006] Step 1: Obtain the running data of multiple shipborne drone swarm simulations;

[0007] Step 2: After discretizing the position data in the operation data according to the time data in the operation data, the utility data set of the shipborne UAV mounting is obtained by calculating the profit parameter;

[0008] Step 3: After storing the utility data set of the shipborne UAVs in an improved cross-linked list data structure, the shipborne UAV swarm mission mode is generated through high-utility time series data mining;

[0009] Step 4: According to the mission mode of the shipborne UAV swarm, adjust the simulation parameters to realize the mission planning assistance of the shipborne UAV swarm.

[0010] Technical Effects

[0011] The present invention mines the location information, time information, and utility information in the multi-UAV simulation data, and provides auxiliary decision-making for the shipborne UAV group mission planning through three-dimensional data, and provides guiding suggestions for improving the efficiency of the shipborne UAV group mission planning. Compared with the prior art, the present invention realizes the extraction of hidden characteristics behind the simulation data, and obtains an efficient mode related to the shipborne UAV group mission planning, thereby providing suggestions for the shipborne UAV group mission planning and meeting the shipborne UAV group mission planning needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a flow chart of the present invention;

[0013] Figure 2 This is a schematic diagram of the initial position simulation of the present invention;

[0014] Figure 3 This is a schematic diagram of the mid-level anti-missile simulation of the present invention;

[0015] Figure 4 It is a relationship diagram between the projects of the present invention;

[0016] Figure 5 The head node data structure of the present invention;

[0017] Figure 6 The linked list node data structure of the present invention;

[0018] Figure 7 The index node data structure of the present invention;

[0019] Figure 8 The data structure of the present invention;

[0020] Fig. 9 It is the simulation data mining result;

[0021] Fig.10 To compare the performance before and after optimization. DETAILED DESCRIPTION

[0022] like Figure 1 As shown, this embodiment relates to a method for planning a group of shipborne UAVs based on a time series mining algorithm, including:

[0023] Step 1: Obtain the position data, time data, carrying information of each drone and task completion result judgment index of each drone during multiple shipborne drone group simulation phases, including:

[0024] Step 1.1: Use the Yuanqi combat scenario simulation software (www.shareetech.com) to obtain the position data, time data, and original data of each drone's carrying information during the simulation phase of the carrier-based drone group. Before the simulation begins, preset the forces and their positions, including: friendly fighters, friendly carrier-based drones, enemy fighters and their initial positions, such as Figure 2 shown.

[0025] Step 1.2: Select Figure 3The middle-layer missile defense in the multi-layer missile defense in the Yuanqi combat scenario simulation software shown in the figure is simulated: when the initial troop configuration of both sides remains unchanged, only the mission planning method of the ship-borne UAV of one's own side is updated and simulated each time. After 100 simulations, the ship health rate, combat time, and combat unit position information are collected, and the ship health rate is used as the task completion result.

[0026] Step 2: Construct a shipborne UAV swarm mission planning dataset containing the time, location, and utility data of the shipborne UAV swarm, including:

[0027] Step 2.1: Based on the combat time in step 1.2, construct a time series database D = {T1...T n} and group all items with the same hypothetical sequence number id into one group and sort them in ascending order by the start time of the items, where: each T contains item i, hypothetical start time start-time, hypothetical end time end-time, start-time <end-time。

[0028] As shown in Table 1, the rows represent tasks and the columns represent the projects executed to perform the tasks, where each project has its own start and end time. For example, (A,0,2) means that the start execution time of project A is 0 and the end time is 2.

[0029] Table 1

[0030] 1(A,1,3),(B,1,3),(A,6,12),(B,8,11),(B,14,16),(C,17,20)

[0031] 2(A,1,2),(B,2,6),(A,10,12),(B,11,15),(C,14,17)

[0032] 3(A,4,7),(B,11,12),(C,12,14),(B,17,17),(C,18,19)

[0033] 4(B,1,5),(A,6,14),(B,7,10),(C,8,9),(B,16,18)

[0034] Step 2.2: Based on the time series database described in step 2.1, the position data is discretized to obtain a discrete time library as shown in Table 2.

[0035] The discretization process is as follows: let the position of the item at the start time be (x1, y1) and the position at the end time be (x2, y2), and discretize the item position, as follows:

[0036] Table 2

[0037]

[0038] Step 2.3: Calculate the profit parameters of the discrete time library and obtain the utility data set of shipborne UAV mounting, as shown in Table 3.

[0039] The profit parameter calculation is specifically as follows: Where: pr(i j ) is i in database D j Unit profit, st(i j ) is the selected organizational structure coefficient, ab(i j ) is the basic capability coefficient of the equipment corresponding to the function (such as sensor detection accuracy), opco(i j ) is the cost coefficient required for operation, maco(i j ) is the cost coefficient required for manufacturing, ne(i j ) is the data transmission networking coefficient.

[0040] Table 3

[0041]

[0042] Step 3: Perform time series mining on the shipborne UAV swarm mission planning dataset to generate the shipborne UAV swarm mission mode, including:

[0043] Step 3.1: Define the execution order relationship between items, specifically: follows, meets, overlaps, contains, and matches, such as Figure 4 shown.

[0044] Step 3.2: Construct an improved cross linked list data structure and an indirect adjacent node linked list to store data, such as Figure 5 and Figure 6 As shown, it specifically includes: the start and end time of each task element, the relationship between two connected task times represented by ①, ②, ③, ④, and ⑤ respectively, the first interval pointer pointing to the directly adjacent node, and the second interval pointer pointing to the indirectly adjacent node.

[0045] like Figure 8As shown, it is a data structure based on an improved cross linked list data structure and an indirect adjacent node linked list for storing utility data of each item and connecting each item and a discrete time library, which is suitable for storing data sets containing multiple information such as utility, time, and location, wherein: the improved cross linked list data structure is used to store a linked list of head node information, including: the start and end time of the current item, and the execution sequence relationship data, while linking the improved cross linked list data structure and the indirect adjacent node linked list of the next item; the indirect adjacent node linked list is used to store a linked list of node information connected to the current item, including: the start and end time of the linked item, the execution sequence relationship, utility, and transaction number data.

[0046] The utility u(i j ,T q )=q(i j ,T q )×pr(i j ), where: q(i j ,T q ) is transaction T q in i j The number of occurrences, pr(i j ) is i in database D j The unit profit corresponding to the item, T q The utility of item X in is defined as The total utility of X in database D is

[0047] Step 3.3: Prune the data structure obtained in step 3.2 according to the set minimum utility threshold to obtain a high-utility time series pattern. Specifically, compare the utility value of each node with the set minimum utility threshold. When the utility value of the node is less than the minimum utility threshold, prune the current node. If it is greater than the minimum utility threshold, judge the next node until all nodes are judged. Fig. 9 As shown in the figure, A, B, C, and D represent different tasks of different shipborne drones, and f, m, o, c, and a represent the execution relationships between items, namely follows, meets, overlaps, contains, and matches. For example, "Am B fD" means that executing task B immediately after executing task A and executing task D a little after the completion of task B can achieve better anti-missile task execution efficiency.

[0048] The high-utility time series pattern refers to a sequence pattern that meets the minimum utility threshold and contains time position information.

[0049] Step 4: Based on the high-utility time series pattern obtained in step 3, adjust the task sequence of the shipborne UAVs to implement the task planning assistance of the shipborne UAV group, including:

[0050] Step 4.1: Adjust the task configuration in the simulation scenario according to the mined mission mode of the shipborne drone group to obtain the optimized simulation model. The specific adjustment operations include: (1) According to the mined mission mode, filter the sensor mount types that appear repeatedly and adjust the drone mounts to the mounts that appear frequently in the mode; (2) According to the mined mission mode, filter the task sequences that appear repeatedly and select the task sequences that appear frequently for execution during tactical design.

[0051] Step 4.2: Run the optimized simulation model 100 times. After obtaining the optimized data, take the health rate (0-100%) of our ship after anti-missile combat as the performance judgment index, and arrange the index in ascending order to observe the performance comparison before and after optimization, such as Fig.10 As shown in the figure, it can be seen that using high-utility time series mining algorithm to optimize the mission planning of ship-borne drone swarms can effectively improve the effectiveness of anti-missile operations.

[0052] After specific practical experiments, the above method was run in a mid-level anti-missile scenario under the environment setting of the Yuanqi combat scenario simulation software. The average health rate before optimization was 26.2%, and the average health rate after optimization was 30.28%. After optimization, the health rate of the ship increased by 4.08%.

[0053] Compared with the existing technology, this method conducts anti-missile scenario simulation of ship-borne UAV swarm based on mid-level anti-missile scenario, and mines the simulation data through the proposed algorithm to obtain the optimal task planning method of ship-borne UAV swarm under different threat scenarios. Finally, the simulation results are optimized according to the mined patterns, and the efficiency before and after optimization is compared, and it is concluded that the proposed method can effectively improve the task planning efficiency of ship-borne UAV swarm.

[0054] The above-mentioned specific implementation can be partially adjusted in different ways by those skilled in the art without departing from the principle and purpose of the present invention. The protection scope of the present invention shall be based on the claims and shall not be limited by the above-mentioned specific implementation. Each implementation scheme within its scope shall be subject to the constraints of the present invention.

Claims

1. A method for shipborne UAV swarm mission planning based on time series mining algorithm, characterized in that: include: Step 1: Obtain the running data of multiple shipborne drone swarm simulations; Step 2: After discretizing the position data in the operation data according to the time data in the operation data, the utility data set of the shipborne UAV mounting is obtained by calculating the profit parameter; Step 3: After storing the utility data set of the shipborne UAVs in an improved cross-linked list data structure, the shipborne UAV swarm mission mode is generated through high-utility time series data mining; Step 4: According to the mission mode of the shipborne UAV swarm, adjust the simulation parameters to realize the mission planning assistance of the shipborne UAV swarm; The improved cross linked list data structure is used to store the linked list of head node information, including: the start and end time of the current item, the execution order relationship data, and the improved cross linked list data structure and the indirect adjacent node linked list of the next item; The indirect adjacent node linked list is a linked list for storing node information connected to the current item, including: start and end time, execution order relationship, utility, and transaction number data of the linked item; The profit parameters are specifically: Where: pr(i j ) is i in database D j Unit profit, st(i j ) is the selected organizational structure coefficient, ab(i j ) is the basic capability coefficient of the equipment corresponding to the function, opco(i j ) is the cost coefficient required for operation, maco(i j ) is the cost coefficient required for manufacturing, ne(i j ) is the data transmission networking mode coefficient; The above-mentioned shipborne UAV swarm mission mode is generated in the following way: Step 3.1: Define the execution order relationship between items, specifically: follows, meets, overlaps, contains, and matches; Step 3.2: Construct an improved cross linked list data structure and an indirect adjacent node linked list to store data, including: the start and end time of each task element, the relationship between two connected task times represented by ①, ②, ③, ④, and ⑤ respectively, the firstinterval pointer pointing to the directly adjacent node, and the secondinterval pointer pointing to the indirectly adjacent node; Step 3.3: Prune the data structure obtained in step 3.2 according to the set minimum utility threshold to obtain a high-utility time series pattern. Specifically, compare the utility value of each node with the set minimum utility threshold. When the utility value of the node is less than the minimum utility threshold, prune the current node. If it is greater than the minimum utility threshold, judge the next node until all nodes are judged.

2. The method for planning a group of shipborne UAVs according to claim 1, characterized in that: The discretization process is as follows: let the position of the item at the start time be (x1, y1) and the position at the end time be (x2, y2), and discretize the item position, as follows:

3. The method for planning a group of shipborne UAVs according to claim 1, characterized in that: The operating data is obtained in the following manner: Step 1.1: Use the Yuanqi combat scenario simulation software to obtain the position data, time data, and original data of the carrying information of each drone during the simulation phase of the carrier-based drone group. Before the simulation starts, preset the forces and their positions, including: own fighter jets, own carrier-based drones, enemy fighter jets and their initial positions; Step 1.2: Select the middle-layer anti-missile system in the multi-layer anti-missile system in the Yuanqi combat scenario simulation software for simulation: When the initial force configuration of both sides remains unchanged, only the mission planning method of the ship-borne UAV of one's own side is updated and simulated each time. After 100 simulations, the ship health rate, combat time, and combat unit position information are collected, and the ship health rate is used as the task completion result.

4. The method for planning a group of shipborne UAVs according to claim 1, characterized in that: The utility data set of the shipborne UAV mount is obtained by the following method: Step 2.1: Based on the combat time in step 1.2, construct a time series database D = {T1...T n } and group all items with the same hypothetical sequence number id into one group and sort them in ascending order by the start time of the items, where: each T contains item i, hypothetical start time start-time, hypothetical end time end-time, start-time <end-time; Step 2.2: Based on the time series database described in step 2.1, the position data is discretized to obtain a discrete time database; Step 2.3: Calculate the profit parameters for the discrete time library and obtain the utility data set of shipborne UAV mounting.

5. The method for planning a group of shipborne UAVs according to claim 1, characterized in that: The above-mentioned shipborne UAV swarm mission planning assistance is achieved by the following methods: Step 4.1: According to the mined mission mode of the shipborne UAV group, the mission configuration in the simulation scenario is adjusted to obtain the optimized simulation model. The specific adjustment operations include: 1) According to the mined mission mode, the sensor mount types that appear repeatedly are screened, and the UAV mounts are adjusted to the mounts that appear frequently in the mode; 2) According to the mined mission mode, the mission sequences that appear repeatedly are screened, and the mission sequences that appear frequently are selected for execution during tactical design; Step 4.2: Run the optimized simulation model 100 times. After obtaining the optimized data, use the health rate of our ship after anti-missile combat as the performance judgment index, and arrange the indicators in ascending order to observe the performance comparison before and after optimization.

Images