Touch screen for unmanned aerial vehicle and instruction decomposition method of touch screen

Abstract

The invention discloses a touch screen for an unmanned aerial vehicle and an instruction decomposition method of the touch screen. The method comprises the following steps: S1, performing standardization processing on task semantic data; s2, generating a complete unmanned aerial vehicle task instruction set; s3, generating an initial population solution according to the complete unmanned aerial vehicle task instruction set; s4, optimizing the initial population solution based on an elk optimization algorithm to generate an optimal task decomposition scheme; s5, generating an unmanned aerial vehicle executable subtask set; s6, collecting feedback data in the task execution process; and S7, dynamically adjusting the task decomposition scheme by using the collected task execution feedback data, and when the feedback data indicates that an obstacle or deviation exists in the execution of the sub-tasks, re-calling the elk optimization algorithm and the fuzzy logic reasoning module to optimize the optimal task decomposition scheme to generate a new unmanned aerial vehicle executable sub-task set. According to the method, the change of the task scene can be dynamically responded, and higher flexibility and execution efficiency are shown in the execution of the multi-target task.

Description

Technical Field

[0001] The present invention relates to the technical field of touch screens for unmanned aerial vehicles, and in particular to a touch screen for unmanned aerial vehicles and a method for decomposing instructions of the touch screen. Background Art

[0002] As an important interface for human-computer interaction, the quality of parsing and decomposing input commands of drone touch screen directly determines the performance of drone mission execution. In the existing technology, most drone touch screen command parsing technologies rely on traditional grammar parsing and rule matching methods. Traditional methods have certain applicability when parsing simple and clear commands, but they are insufficient when faced with the uncertainty and ambiguity of user input. For example, users may have difficulty in accurately understanding their intentions due to operational errors, vague command expressions, or incomplete input, thereby causing task parsing errors. In addition, existing command parsing technologies usually lack the ability to complete the semantics of the context, resulting in some inputs that cannot be effectively used by the system.

[0003] In terms of task decomposition, most existing technologies adopt static task decomposition strategies, which decompose complex tasks into several subtasks through pre-defined rules. Existing methods are relatively effective in dealing with low-complexity tasks, but they are prone to fall into local optimality for complex mission scenarios that require dynamic adaptation of drones, and lack a global perspective. Traditional task decomposition methods also often ignore the dynamics of task execution priority and resource allocation, resulting in a lack of flexibility and execution efficiency in task decomposition schemes. In addition, some optimization algorithms have problems of slow convergence and poor adaptability when solving task decomposition problems, making it difficult to meet real-time requirements.

[0004] In summary, the existing technologies have the following significant defects in the analysis of drone touch screen commands and task decomposition: the analysis technology cannot effectively handle uncertain input and contextual semantic completion, the task decomposition method lacks global optimization capabilities and cannot dynamically adapt to complex scenarios, and the traditional optimization algorithm has deficiencies in task decomposition efficiency and result quality. The above problems directly limit the application effect of drones in multi-target and dynamic mission scenarios. There is an urgent need for a new method that combines intelligent optimization algorithms with fuzzy logic reasoning to overcome the limitations of the existing technology. Summary of the invention

[0005] An object of the present invention is to provide a touch screen for an unmanned aerial vehicle and a command decomposition method for the touch screen. The present invention can dynamically respond to changes in mission scenarios and show higher flexibility and execution efficiency in the execution of multi-target tasks.

[0006] A method for decomposing a touch screen command for an unmanned aerial vehicle according to an embodiment of the present invention comprises the following steps:

[0007] S1. Receive the task instruction input by the user through the touch screen of the drone, convert the input task instruction into task semantic data for the drone, and standardize the task semantic data;

[0008] S2. Use the fuzzy logic reasoning module to perform fuzzy analysis on the standardized task semantic data, correct and supplement the uncertainty information in the task semantic data according to the preset fuzzy rules, and generate a complete set of UAV task instructions;

[0009] S3. Initialize the population of the elk optimization algorithm, set the initial elk population number, population structure and fitness function, and generate the initial population solution based on the complete set of UAV mission instructions;

[0010] S4. Optimize the initial population solution based on the elk optimization algorithm to generate the optimal task decomposition solution;

[0011] S5. Decompose the complete UAV task instruction set into multiple executable subtasks according to the optimal task decomposition scheme, and prioritize and allocate resources for the executable subtasks to generate a set of executable subtasks for the UAV;

[0012] S6. The drone executable subtask set is transmitted to the drone execution module, and the execution status of the subtask is monitored in real time, and feedback data during the task execution is collected;

[0013] S7. Use the collected task execution feedback data to dynamically adjust the task decomposition plan. When the feedback data indicates that there are obstacles or deviations in the execution of subtasks, re-call the ELK optimization algorithm and fuzzy logic reasoning module to optimize the optimal task decomposition plan and generate a new set of subtasks that can be executed by the drone.

[0014] Optionally, the S1 step includes:

[0015] S11. Receive the task instruction I input by the user through the touch screen of the drone input ,The input content is either a clear instruction or a vague description;

[0016] S12. The task instruction I input Converted into mission semantic data D for drones semantic , the task semantics is generated based on the contextual content of the input task instruction and includes the following parameters:

[0017] UAV mission target T goal , indicating the specific operational objectives that the drone needs to complete;

[0018] UAV Mission Area A region , indicating the geographical range in which the drone needs to perform the mission;

[0019] UAV mission constraints C constraint , including task execution time limit, energy consumption limit and path obstacle limit;

[0020] S13. According to the preset mission parameter range, the UAV mission target T goal Format the expression for the drone mission area A region Perform geographic coordinate system conversion to generate standardized geographic location information A region,norm , the UAV mission constraints C constraint Perform constraint parameter analysis, convert fuzzy constraints into clear numerical ranges, and generate standardized UAV mission constraints C constraint,norm , and combine them to generate a complete UAV mission semantic dataset D semantic,norm .

[0021] Optionally, the S2 step includes:

[0022] S21. Perform semantic fuzzy analysis on the fuzzy or incomplete parts of the UAV mission semantic data, using the fuzzy membership function μ T (x) represents the adaptability of the UAV mission target, using the fuzzy membership function μ A (x, y) performs fuzzy analysis on the UAV mission area to indicate whether a certain coordinate point (x, y) in the area belongs to the mission area, using the fuzzy membership function μ C (z) Fuzzy representation of the UAV mission constraints, where z is the constraint value, reflecting the tightness of the mission constraints;

[0023] S22. Targeting UAV missions T goal,norm , based on the input semantics and context information, complete the task goal:

[0024]

[0025] Among them, T i is the task target to complete the candidate items, i∈N, N is the candidate task set, μ T (T i ) is the membership value of the task target, indicating that T i The adaptability of i is the task priority weight, d i is the correlation distance of context information. The smaller the value, the better the match between the task goal and the context. α is an adjustment parameter used to balance the weights of membership and context matching.

[0026] Targeting UAV Mission Area A region,norm , based on fuzzy reasoning to complete the range of the region:

[0027] Areg i on,comp l e t e =∪ j∈M [μ A (x j ,y j )·ψ(x j ,y j )];

[0028] Among them, (x j ,y j ) is the coordinate point in the region, j∈M, M is the set of candidate region points, μ A (x j ,y j ) is the membership value of the coordinate point, indicating whether the point belongs to the target area, ψ(x j ,y j ) is the completion factor, which is defined based on context information as:

[0029]

[0030] Among them, r represents the distance threshold between the reference point and the current point, and β is the expansion speed of the control area;

[0031] For UAV mission constraints C constraint,norm , using fuzzy reasoning to generate new constraint completion values:

[0032] C constraint,complete =∫ z∈L μ C (z)·g(z)dz;

[0033] Where z is the value of the constraint, L is the domain of the constraint, and μ C (z) is the membership value of the constraint condition, indicating the adaptability of z, and g(z) is the completion function, which is defined as follows in combination with the dynamic adjustment of the environment parameters of task execution:

[0034]

[0035] Among them, z0 is the current optimal value of the constraint condition, k is the function steepness, and γ is the complement strength coefficient;

[0036] S23. UAV mission target T after semantic completion goal,complete 、UAV mission area A region,complete and the UAV mission constraints C constraint,complete Generate a complete set of UAV mission instructions D instruction ;

[0037] S24. Defuzzify the UAV mission instruction set and output the complete UAV mission instruction set after defuzzification.

[0038] Optionally, the step S24 includes:

[0039] S241. Defuzzify the semantic description of the UAV mission target and use the maximum value method of the membership function to select the best semantic description of the mission target:

[0040]

[0041] Among them, T goal,final is the final task goal after defuzzification;

[0042] S242. Defuzzify the UAV mission area and determine the center range of the mission area based on the optimal path planning algorithm within the fuzzy area:

[0043]

[0044] Among them, (x k ,y k ) is the candidate coordinate of the center point of the region, k∈K, K is the candidate set of the center point, μ A (x j ,y j ) is the membership value of the coordinate point in the region, indicating its adaptability to the task area, P optimal is the optimal path center point set of the defuzzified area;

[0045] S243. Defuzzify the UAV mission constraints and use the constraint membership maximum method to clarify the resource and time allocation plan:

[0046]

[0047] Among them, z is the candidate value of the constraint, L is the domain of the constraint, C constraint,final The final constraint condition value after defuzzification, including time allocation and resource allocation;

[0048] S244. According to the defuzzified UAV mission target T goal,final 、Mission area center range P optimal and task constraints C constraint,final Generate a complete set of drone mission instructions:

[0049] D instruction,complete ={T goal,final ,P optimal ,C constraint,final}.

[0050] Optionally, the S4 step includes:

[0051] S41. Based on the complete UAV mission instruction set D instruction,complete Initialize the population, where each individual in the population represents a UAV task decomposition plan;

[0052] Define the population size N population , which is the total number of task decomposition solutions that can be generated, and divide the population into several families, each with N individuals family , where the number of leaders is N leader ,The leader represents the high-priority task decomposition plan;

[0053] Define the fitness function f fitness , evaluate the execution effect of the task decomposition scheme in the UAV mission:

[0054] f fitness (x) = w1·f priority (x)+w2·f resource (x)+w3·f time (x);

[0055] Among them, x represents the task decomposition plan, including task objectives, task areas and task constraints, and f priority (x) represents the rationality of task priority allocation in the task decomposition scheme, f resource (x) represents the efficiency of resource allocation, f time (x) represents the rationality of task time allocation, w1, w2, w3 are weight parameters, which are dynamically adjusted according to task requirements;

[0056] S42. Generate an optimized UAV task decomposition plan based on the collaborative behavior of the leader and followers of the elk group, and perform the task decomposition on the leader L of each family. i Using the fitness value f of the task decomposition scheme fitness (L i ) determines its influence. The leader solution represents the current optimal task decomposition strategy. j Update task breakdown:

[0057]

[0058] in, is the task decomposition scheme for the follower in the tth generation, is the leader’s task decomposition plan in the tth generation, reflecting the current optimal instruction decomposition method. λ is the step size parameter, which controls the adjustment range of the task decomposition plan. A new set of candidate UAV task decomposition plans S is generated by updating. candidate ;

[0059] S43. Generate a new task decomposition scheme through the interaction between the leader and the follower, and optimize the decomposition of the UAV's task instructions. In each family, based on the leader L i and follower F j The interaction generates a new task decomposition scheme x new :

[0060]

[0061] Among them, x new represents the newly generated task decomposition plan, η1 is the interaction factor, which indicates the degree of innovation of the task plan;

[0062] The newly generated task decomposition scheme is optimized according to the actual operation requirements of the UAV and added to the candidate task decomposition scheme set S candidate ;

[0063] S44. Filter the candidate task decomposition solution set S according to the specific requirements of the UAV task candidate , select the best solution with the highest fitness:

[0064]

[0065] Among them, x optimal The optimal task decomposition scheme includes task target priority, resource allocation and path planning. The candidate schemes are sorted according to the fitness value and the selected scheme is suitable for the current mission scenario of the UAV;

[0066] S45. Decompose the optimal task into x optimal The output is sent to the UAV mission execution module to instruct the UAV to perform operations according to the decomposed mission plan.

[0067] Optionally, the step S5 includes:

[0068] S51. According to the optimal task decomposition plan x optimal For the complete UAV mission instruction set D instruction,complete Decompose the task, extract the task objectives, task areas and task constraints in the optimal task decomposition plan, decompose the task objectives into several independent sub-objectives, divide them according to the priority and area range of the sub-objectives, and Contains specific task operations, execution areas, and constraints, generating multiple sets of executable subtasks:

[0069]

[0070] S52. Prioritize the executable subtask set and process high-priority subtasks first:

[0071]

[0072] in, is the i-th subtask, represents the adaptability membership of the subtask goal, Indicates the urgency of the subtask. Indicates the degree of dependence of the subtask on other tasks. w4, w5, and w6 are weight parameters for priority sorting, which are dynamically adjusted according to task requirements.

[0073] S53. Allocate resources to the executable subtask set, calculate the resource allocation plan according to the current resource status of the drone, and allocate specific resources to each subtask according to the calculation results:

[0074]

[0075] in, Assign to subtask Resources, R available is the total amount of resources currently available to the drone, ρ i For subtask The resource demand weight is determined by the task complexity and area scope, n is the total number of subtasks, and the current resource status includes power, flight time, and load capacity;

[0076] S54. Integrate the priority sorting and resource allocation results into a set of subtasks that can be executed by the drone D subtasks :

[0077]

[0078] Each subtask contains task objectives, priorities, and resource allocation plans.

[0079] A touch screen for an unmanned aerial vehicle comprises an internally embedded memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements an instruction decomposition method for a touch screen for an unmanned aerial vehicle when executing the computer program.

[0080] The beneficial effects of the present invention are:

[0081] (1) The present invention introduces the elk optimization algorithm and applies the estrus, calving and selection period mechanisms of elk group behavior to the optimization of UAV task decomposition. Compared with the traditional optimization algorithm, the elk optimization algorithm can better balance the capabilities of global search and local search in the process of population initialization, individual update and solution selection, and avoid falling into local optimality. In the estrus stage, the leader and follower behaviors are used to generate high fitness solutions. In the calving period, innovative solutions are generated through interaction. In the selection period, the optimal solution is selected through global evaluation, so that the task decomposition can achieve global optimal effects in terms of priority, resource allocation and path planning.

[0082] (2) The present invention effectively handles the ambiguity and uncertainty problems in the drone touch screen input through the fuzzy logic reasoning module, and dynamically adjusts and completes the task objectives, task areas and constraints by constructing fuzzy membership functions, fuzzy reasoning rules and semantic completion mechanisms. In the task area fuzzification, the precise task scope is calculated through fuzzy membership functions and area completion factors to ensure the integrity and accuracy of task analysis.

[0083] (3) The present invention prioritizes and allocates resources to the executable subtask set generated by the optimal task decomposition scheme in combination with the dynamic characteristics of the UAV task, thereby forming an executable subtask set for the UAV. The priority sorting method comprehensively considers the adaptability of the task goal, the degree of urgency and the task dependency, and realizes real-time optimization of the task priority by dynamically adjusting the weight. The resource allocation method calculates the resource allocation scheme in combination with the real-time status of the UAV to ensure the high efficiency of task execution. Compared with the traditional static task decomposition method, the present invention can dynamically respond to changes in the task scenario and show higher flexibility and execution efficiency in the execution of multi-objective tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0085] Figure 1 The present invention provides a flow chart of a touch screen for an unmanned aerial vehicle and a method for decomposing instructions of the touch screen. DETAILED DESCRIPTION

[0086] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0087] refer to Figure 1 , a method for decomposing a command of a touch screen for a drone, comprising the following steps:

[0088] S1. Receive the task instruction input by the user through the touch screen of the drone, convert the input task instruction into task semantic data for the drone, and standardize the task semantic data;

[0089] S2. Use the fuzzy logic reasoning module to perform fuzzy analysis on the standardized task semantic data, correct and supplement the uncertainty information in the task semantic data according to the preset fuzzy rules, and generate a complete set of UAV task instructions;

[0090] S3. Initialize the population of the elk optimization algorithm, set the initial elk population number, population structure and fitness function, and generate the initial population solution based on the complete set of UAV mission instructions;

[0091] S4. Optimize the initial population solution based on the elk optimization algorithm to generate the optimal task decomposition solution;

[0092] S5. Decompose the complete UAV task instruction set into multiple executable subtasks according to the optimal task decomposition scheme, and prioritize and allocate resources for the executable subtasks to generate a set of executable subtasks for the UAV;

[0093] S6. The drone executable subtask set is transmitted to the drone execution module, and the execution status of the subtask is monitored in real time, and feedback data during the task execution is collected;

[0094] S7. Use the collected task execution feedback data to dynamically adjust the task decomposition plan. When the feedback data indicates that there are obstacles or deviations in the execution of subtasks, re-call the ELK optimization algorithm and fuzzy logic reasoning module to optimize the optimal task decomposition plan and generate a new set of subtasks that can be executed by the drone.

[0095] In this implementation, step S1 includes:

[0096] S11. Receive the task instruction I input by the user through the touch screen of the drone input ,The input content is either a clear instruction or a vague description;

[0097] S12. Task instruction I input Converted into mission semantic data D for drones semantic , the task semantics is generated based on the contextual content of the input task instruction and includes the following parameters:

[0098] UAV mission target T goal , indicating the specific operational objectives that the drone needs to complete;

[0099] UAV Mission Area A region , indicating the geographical range in which the drone needs to perform the mission;

[0100] UAV mission constraints C constraint , including task execution time limit, energy consumption limit and path obstacle limit;

[0101] S13. According to the preset mission parameter range, the UAV mission target T goal Format the expression for the drone mission area A region Perform geographic coordinate system conversion to generate standardized geographic location information A region,norm , the UAV mission constraints C constraint Perform constraint parameter analysis, convert fuzzy constraints into clear numerical ranges, and generate standardized UAV mission constraints C constraint,norm , and combine them to generate a complete UAV mission semantic dataset D semantic,norm .

[0102] In this embodiment, step S2 includes:

[0103] S21. Perform semantic fuzzy analysis on the fuzzy or incomplete parts of the UAV mission semantic data, using the fuzzy membership function μ T (x) represents the adaptability of the UAV mission target, using the fuzzy membership function μ A (x, y) performs fuzzy analysis on the UAV mission area to indicate whether a certain coordinate point (x, y) in the area belongs to the mission area, using the fuzzy membership function μ C (z) Fuzzy representation of the UAV mission constraints, where z is the constraint value, reflecting the tightness of the mission constraints;

[0104] S22. Targeting UAV missions T goal,norm , based on the input semantics and context information, complete the task goal:

[0105]

[0106] Among them, T i is the task target to complete the candidate items, i∈N, N is the candidate task set, μ T (T i ) is the membership value of the task target, indicating that T i The adaptability of i is the task priority weight, d i is the correlation distance of context information. The smaller the value, the better the match between the task goal and the context. α is an adjustment parameter used to balance the weights of membership and context matching.

[0107] Targeting UAV Mission Area A region,norm , based on fuzzy reasoning to complete the range of the region:

[0108] Aregion,complete =∪ j∈M [μ A (x j ,y j )·ψ(x j ,y j )];

[0109] Among them, (x j ,y j ) is the coordinate point in the region, j∈M, M is the set of candidate region points, μ A (x j ,y j ) is the membership value of the coordinate point, indicating whether the point belongs to the target area, ψ(x j ,y j ) is the completion factor, which is defined based on context information as:

[0110]

[0111] Among them, r represents the distance threshold between the reference point and the current point, and β is the expansion speed of the control area;

[0112] For UAV mission constraints C constraint,norm , using fuzzy reasoning to generate new constraint completion values:

[0113] C constraint,complete =∫ z∈L μ C (z)·g(z)dz;

[0114] Where z is the value of the constraint, L is the domain of the constraint, and μ C (z) is the membership value of the constraint condition, indicating the adaptability of z, and g(z) is the completion function, which is defined as follows in combination with the dynamic adjustment of the environment parameters of task execution:

[0115]

[0116] Among them, z0 is the current optimal value of the constraint condition, k is the function steepness, and γ is the complement strength coefficient;

[0117] S23. UAV mission target T after semantic completion goal,complete 、UAV mission area A region,complete and the UAV mission constraints C constraint,complete Generate a complete set of UAV mission instructions D instruction ;

[0118] S24. Defuzzify the UAV mission instruction set and output the complete UAV mission instruction set after defuzzification.

[0119] In this implementation, step S24 includes:

[0120] S241. Defuzzify the semantic description of the UAV mission target and use the maximum value method of the membership function to select the best semantic description of the mission target:

[0121]

[0122] Among them, T goal,final is the final task goal after defuzzification;

[0123] S242. Defuzzify the UAV mission area and determine the center range of the mission area based on the optimal path planning algorithm within the fuzzy area:

[0124]

[0125] Among them, (x k ,y k ) is the candidate coordinate of the center point of the region, k∈K, K is the candidate set of the center point, μ A (x j ,y j ) is the membership value of the coordinate point in the region, indicating its adaptability to the task area, P optimal is the optimal path center point set of the defuzzified area;

[0126] S243. Defuzzify the UAV mission constraints and use the constraint membership maximum method to clarify the resource and time allocation plan:

[0127]

[0128] Among them, z is the candidate value of the constraint, L is the domain of the constraint, C constraint,final The final constraint condition value after defuzzification, including time allocation and resource allocation;

[0129] S244. According to the defuzzified UAV mission target T goal,final 、Mission area center range P optimal and task constraints C constraint,final Generate a complete set of drone mission instructions:

[0130] D instruction,complete ={T goal,final ,P optimal ,C constraint,final}.

[0131] In this embodiment, step S4 includes:

[0132] S41. Based on the complete UAV mission instruction set Dinstruction,complete Initialize the population, where each individual in the population represents a UAV task decomposition plan;

[0133] Define the population size N population , which is the total number of task decomposition solutions that can be generated, and divide the population into several families, each with N individuals family , where the number of leaders is N leader ,The leader represents the high-priority task decomposition plan;

[0134] Define the fitness function f fitness , evaluate the execution effect of the task decomposition scheme in the UAV mission:

[0135] f fitness (x) = w1·f priority (x)+w2·f resource (x)+w3·f time (x);

[0136] Among them, x represents the task decomposition plan, including task objectives, task areas and task constraints, and f priority (x) represents the rationality of task priority allocation in the task decomposition scheme, f resource (x) represents the efficiency of resource allocation, f time (x) represents the rationality of task time allocation, w1, w2, w3 are weight parameters, which are dynamically adjusted according to task requirements;

[0137] S42. Generate an optimized UAV task decomposition plan based on the collaborative behavior of the leader and followers of the elk group, and perform the task decomposition on the leader L of each family. i Using the fitness value f of the task decomposition scheme fitness (L i ) determines its influence. The leader solution represents the current optimal task decomposition strategy. j Update task breakdown:

[0138]

[0139] in, is the task decomposition scheme for the follower in the tth generation, is the leader’s task decomposition plan in the tth generation, reflecting the current optimal instruction decomposition method. λ is the step size parameter, which controls the adjustment range of the task decomposition plan. A new set of candidate UAV task decomposition plans S is generated by updating. candidate ;

[0140] S43. Generate a new task decomposition scheme through the interaction between the leader and the follower, and optimize the decomposition of the UAV's task instructions. In each family, based on the leader Li and follower F j The interaction generates a new task decomposition scheme x new :

[0141]

[0142] Among them, x new represents the newly generated task decomposition plan, η1 is the interaction factor, which indicates the degree of innovation of the task plan;

[0143] The newly generated task decomposition scheme is optimized according to the actual operation requirements of the UAV and added to the candidate task decomposition scheme set S candidate ;

[0144] S44. Filter the candidate task decomposition solution set S according to the specific requirements of the UAV task candidate , select the best solution with the highest fitness:

[0145]

[0146] Among them, x optimal The optimal task decomposition scheme includes task target priority, resource allocation and path planning. The candidate schemes are sorted according to the fitness value and the selected scheme is suitable for the current mission scenario of the UAV;

[0147] S45. Decompose the optimal task into x optimal The output is sent to the UAV mission execution module to instruct the UAV to perform operations according to the decomposed mission plan.

[0148] In this implementation, step S5 includes:

[0149] S51. According to the optimal task decomposition plan x optimal For the complete UAV mission instruction set D instruction,complete Decompose the task, extract the task objectives, task areas and task constraints in the optimal task decomposition plan, decompose the task objectives into several independent sub-objectives, divide them according to the priority and area range of the sub-objectives, and Contains specific task operations, execution areas, and constraints, generating multiple sets of executable subtasks:

[0150]

[0151] S52. Prioritize the executable subtask set and process high-priority subtasks first:

[0152]

[0153] in, is the i-th subtask, represents the adaptability membership of the subtask goal, Indicates the urgency of the subtask. Indicates the degree of dependence of the subtask on other tasks. w4, w5, and w6 are weight parameters for priority sorting, which are dynamically adjusted according to task requirements.

[0154] S53. Allocate resources to the executable subtask set, calculate the resource allocation plan according to the current resource status of the drone, and allocate specific resources to each subtask according to the calculation results:

[0155]

[0156] in, Assign to subtask Resources, R available is the total amount of resources currently available to the drone, ρ i For subtask The resource demand weight is determined by the task complexity and area scope, n is the total number of subtasks, and the current resource status includes power, flight time, and load capacity;

[0157] S54. Integrate the priority sorting and resource allocation results into a set of subtasks that can be executed by the drone D subtasks :

[0158]

[0159] Each subtask contains task objectives, priorities, and resource allocation plans.

[0160] A touch screen for an unmanned aerial vehicle comprises an internally embedded memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements an instruction decomposition method for the touch screen for an unmanned aerial vehicle when executing the computer program.

[0161] Embodiment 1:

[0162] Example At 3 pm on August 15, 2024, an earthquake caused the collapse of multiple buildings in the central area of ​​​​H city. The rescue team quickly deployed drones to perform post-disaster search and rescue missions. Due to the complex on-site environment, the command center operator quickly input instructions through the touch screen: "Search areas A and B disaster-stricken areas, give priority to searching crowded places, and upload real-time photos as soon as possible." The input instructions contain vague descriptions (such as "priority" and "as soon as possible") and incomplete area definitions (such as the specific boundaries of areas A and B are not clear). The command center needs to parse the vague touch screen input into complete mission instructions and decompose and optimize the drone mission.

[0163] At 3:05 p.m., the drone used the computer program built into the touch screen to receive the mission instructions input by the command center and began to parse them. Since the words "priority" and "as soon as possible" in the instructions were fuzzy descriptions, the system performed semantic analysis through the fuzzy logic reasoning module:

[0164] The real-time disaster score of area A is 87, and that of area B is 75. The system determines to search area A first.

[0165] “As soon as possible” is converted into actual time requirement through fuzzy time membership function, which is defined as completing the search task of crowd gathering places in area A within 5 minutes.

[0166] The boundaries between Area A and Area B are completed with specific coordinate points using historical drone images and real-time data of the disaster area.

[0167] At 3:08 pm, after the analysis was completed, the system generated a complete set of drone mission instructions:

[0168] 1. Drone 1: Search the crowd gathering area in the north of Area A, take photos and upload them;

[0169] 2. Drone 2: Search the crowd gathering area in the southern part of Area A and drop emergency supplies;

[0170] 3. Drone 3: Real-time monitoring of the southeast area of ​​Zone B, transmitting video images.

[0171] At 3:10 p.m., the system decomposed the task instruction set based on the ELK optimization algorithm.

[0172] Initialize the population: Generate 30 task decomposition plans, each of which contains the UAV mission objectives, mission areas, and resource allocation.

[0173] Estrus stage: The leader (current optimal solution) and the follower jointly update the task objectives. Solutions with no obvious fitness differences are eliminated, and solutions with high fitness enter the next stage.

[0174] Calving stage: Generate innovative solutions through interaction. In the original solution of the embodiment, UAV 2 was responsible for the northern part of area A, but due to unreasonable resource allocation (insufficient power to complete), the system updated the solution to transfer the task to UAV 1.

[0175] Selection phase: The system evaluates candidate solutions and ultimately selects the optimal task decomposition solution, increasing the task area coverage by 15% and the power allocation utilization rate by 92%.

[0176] At 3:15 pm, the drone began its mission:

[0177] Drone 1 entered the northern part of Area A and scanned heat sources in multiple ruins, located the possible positions of trapped persons, took three high-definition photos (file names photo1.jpg, photo2.jpg, photo3.jpg) and uploaded them to the command center.

[0178] Drone 2 completed the delivery of 5 kg of emergency supplies in the southern part of Area A. At the same time, it detected thick smoke in the area and fed back the location information to the command center in real time, suggesting adjustments to the search strategy.

[0179] Drone 3 was shooting real-time video in the southeast of area B, but due to signal interference, part of the picture was blurred. The system switched the task to Drone 2 and completed the picture transmission through the backup channel.

[0180] After the task is completed, the command center compares the data of the method of the present invention with the data of the traditional task decomposition method, and the results are as follows Table 1:

[0181] Table 1 Performance of the traditional method and the method of the present invention in task execution

[0182] index Traditional methods Method of the present invention Average task resolution time (seconds) 15 5 Task decomposition success rate 80％ 97％ Average task completion time (minutes) 35 18 Regional coverage 70％ 85％ Resource utilization (power allocation efficiency) 75％ 92％ Fuzzy instruction parsing success rate 60％ 94％

[0183] In the specific case, UAV 1 completed the search mission in the northern part of Area A at 3:20 and accurately located the positions of three possible trapped people. UAV 2 completed the material delivery mission in the southern part of Area A. Due to the interference of thick smoke, the system quickly adjusted the mission plan. The real-time image transmission of UAV 3 successfully helped the command center to understand the dynamic situation in the southeastern area of ​​the disaster area in real time.

[0184] This example demonstrates the advantages of the method of the present invention in complex task decomposition through the post-earthquake rescue scenario. The ability to parse fuzzy instructions increases the efficiency of drone task execution by nearly 50%, significantly improves the mission area coverage and resource utilization, and provides an efficient solution for drone scheduling in complex scenarios.

[0185] The present invention introduces the elk optimization algorithm and applies the estrus, calving and selection period mechanisms of elk group behavior to the optimization of UAV task decomposition. Compared with traditional optimization algorithms, the elk optimization algorithm can better balance the capabilities of global search and local search in the processes of population initialization, individual update and solution selection, and avoid falling into local optimality. In the estrus stage, the leader and follower behaviors are used to generate high fitness solutions, in the calving period, innovative solutions are generated through interaction, and in the selection period, the optimal solution is selected through global evaluation to achieve global optimal effects in task decomposition in terms of priority, resource allocation and path planning.

[0186] The present invention effectively handles the fuzziness and uncertainty problems in the touch screen input of the UAV through the fuzzy logic reasoning module, and dynamically adjusts and completes the task objectives, task areas and constraints by constructing fuzzy membership functions, fuzzy reasoning rules and semantic completion mechanisms. In the task area fuzzification, the precise task scope is obtained by calculating the fuzzy membership function and the area completion factor, thereby ensuring the integrity and accuracy of the task analysis.

[0187] The present invention prioritizes and allocates resources to the executable subtask set generated by the optimal task decomposition scheme in combination with the dynamic characteristics of the UAV task, so as to form an executable subtask set of the UAV. The priority sorting method comprehensively considers the adaptability of the task target, the urgency and the task dependency, and realizes the real-time optimization of the task priority by dynamically adjusting the weight. The resource allocation method calculates the resource allocation scheme in combination with the real-time status of the UAV to ensure the high efficiency of task execution. Compared with the traditional static task decomposition method, the present invention can dynamically respond to the changes of the task scenario, and shows higher flexibility and execution efficiency in the execution of multi-objective tasks.

[0188] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for decomposing commands of a touch screen for an unmanned aerial vehicle, characterized in that: The steps include: S1. Receive the task instruction input by the user through the touch screen of the drone, convert the input task instruction into task semantic data for the drone, and standardize the task semantic data; S2. Use the fuzzy logic reasoning module to perform fuzzy analysis on the standardized task semantic data, correct and supplement the uncertainty information in the task semantic data according to the preset fuzzy rules, and generate a complete set of UAV task instructions; S3. Initialize the population of the elk optimization algorithm, set the initial elk population number, population structure and fitness function, and generate the initial population solution based on the complete set of UAV mission instructions; S4. Optimize the initial population solution based on the elk optimization algorithm to generate the optimal task decomposition solution; S5. Decompose the complete UAV task instruction set into multiple executable subtasks according to the optimal task decomposition scheme, and prioritize and allocate resources for the executable subtasks to generate a set of executable subtasks for the UAV; S6. The drone executable subtask set is transmitted to the drone execution module, and the execution status of the subtask is monitored in real time, and feedback data during the task execution is collected; S7. Use the collected task execution feedback data to dynamically adjust the task decomposition plan. When the feedback data indicates that there are obstacles or deviations in the execution of subtasks, re-call the ELK optimization algorithm and fuzzy logic reasoning module to optimize the optimal task decomposition plan and generate a new set of subtasks that can be executed by the drone.

2. The command decomposition method for a touch screen for an unmanned aerial vehicle according to claim 1, characterized in that: The S1 step includes: S11. Receive the task instruction I input by the user through the touch screen of the drone input ,The input content is either a clear instruction or a vague description; S12. The task instruction I input Converted into mission semantic data D for drones semantic , the task semantics is generated based on the contextual content of the input task instruction and includes the following parameters: UAV mission target T goal , indicating the specific operational objectives that the drone needs to complete; UAV Mission Area A region , indicating the geographical range in which the drone needs to perform the mission; UAV mission constraints C constraint , including task execution time limit, energy consumption limit and path obstacle limit; S13. According to the preset mission parameter range, the UAV mission target T goal Format the expression for the drone mission area A region Perform geographic coordinate system conversion to generate standardized geographic location information A region,norm , the UAV mission constraints C constraint Perform constraint parameter analysis, convert fuzzy constraints into clear numerical ranges, and generate standardized UAV mission constraints C constraint,norm , and combine them to generate a complete UAV mission semantic dataset D semantic,norm .

3. The command decomposition method for a touch screen for an unmanned aerial vehicle according to claim 1, characterized in that: The S2 step includes: S21. Perform semantic fuzzification analysis on the fuzzy or incomplete parts of the UAV mission semantic data, using the fuzzy membership function μ T (x) represents the adaptability of the UAV mission target, using the fuzzy membership function μ A (x, y) performs fuzzy analysis on the UAV mission area to indicate whether a certain coordinate point (x, y) in the area belongs to the mission area, using the fuzzy membership function μ C (z) Fuzzy representation of the UAV mission constraints, where z is the constraint value, reflecting the tightness of the mission constraints; S22. Targeting UAV missions T goal,norm , based on the input semantics and context information, complete the task goal: Among them, T i is the task target to complete the candidate items, i∈N, N is the candidate task set, μ T (T i ) is the membership value of the task target, indicating that T i The adaptability of i is the task priority weight, d i is the correlation distance of context information. The smaller the value, the better the match between the task goal and the context. α is an adjustment parameter used to balance the weights of membership and context matching. This formula is used to select the most adaptable task target from the candidate task target set N, and dynamically select the optimal target from multiple candidate task targets by combining fuzzy membership and context relevance. Targeting UAV Mission Area A region,norm , based on fuzzy reasoning to complete the range of the region: A region,complete =∪ j∈M [m A (x j ,y j )·ψ(x j ,y j )]; Among them, (x j ,y j ) is the coordinate point in the region, j∈M, M is the set of candidate region points, μ A (x j ,y j ) is the membership value of the coordinate point, indicating whether the point belongs to the target area, ψ(x j ,y j ) is the completion factor, which is defined based on context information as: Among them, r represents the distance threshold between the reference point and the current point, and β is the expansion speed of the control area; This formula combines the membership degree of fuzzy region points with context information to dynamically complete the region boundary and ensure that the generated task area has a clear geometric definition. For UAV mission constraints C constraint,norm , using fuzzy reasoning to generate new constraint completion values: C constraint,complete =∫ z∈L μ C (z)·g(z)dz; Where z is the value of the constraint, L is the domain of the constraint, and μ C (z) is the membership value of the constraint condition, indicating the adaptability of z, and g(z) is the completion function, which is defined as follows in combination with the dynamic adjustment of the environment parameters of task execution: Among them, z0 is the current optimal value of the constraint condition, k is the function steepness, and γ is the complement strength coefficient; This formula generates the most suitable constraint conditions for the current task environment by combining membership integral with the completion function, ensuring the operability and constraint compliance during task execution. S23. UAV mission target T after semantic completion goal,complete 、UAV mission area A region,complete and the UAV mission constraints C constraint,complete Generate a complete set of UAV mission instructions D instruction ; S24. Defuzzify the UAV mission instruction set and output the complete UAV mission instruction set after defuzzification.

4. The command decomposition method for a touch screen for an unmanned aerial vehicle according to claim 1, characterized in that: The step S24 includes: S241. Defuzzify the semantic description of the UAV mission target and use the maximum value method of the membership function to select the best semantic description of the mission target: Among them, T goal,final is the final task goal after defuzzification; S242. Defuzzify the UAV mission area and determine the center range of the mission area based on the optimal path planning algorithm within the fuzzy area: Among them, (x k ,y k ) is the candidate coordinate of the center point of the region, k∈K, K is the candidate set of the center point, μ A (x j ,y j ) is the membership value of the coordinate point in the region, indicating its adaptability to the task area, P optimal is the optimal path center point set of the defuzzified area; S243. Defuzzify the UAV mission constraints and use the constraint membership maximum method to clarify the resource and time allocation plan: Among them, z is the candidate value of the constraint, L is the domain of the constraint, C constraint,final The final constraint condition value after defuzzification, including time allocation and resource allocation; S244. According to the defuzzified UAV mission target T goal,final 、Mission area center range P optimal and task constraints C constraint,final Generate a complete set of drone mission instructions: D instruction,complete ={T goal,final ,P optimal ,C constraint,final }。 5. The method for decomposing a command of a touch screen for an unmanned aerial vehicle according to claim 1, characterized in that: The S4 step includes: S41. Based on the complete UAV mission instruction set D instruction,complete Initialize the population, where each individual in the population represents a UAV task decomposition plan; Define the population size N population , which is the total number of task decomposition solutions that can be generated, and divide the population into several families, each with N individuals family , where the number of leaders is N leader ,The leader represents the high-priority task decomposition plan; Define the fitness function f fitness , evaluate the execution effect of the task decomposition scheme in the UAV mission: f fitness (x)=w1·f priority (x)+w2·f resource (x)+w3·f time (x); Among them, x represents the task decomposition plan, including task objectives, task areas and task constraints, and f priority (x) indicates the rationality of the task priority allocation in the task decomposition scheme, f resource (x) represents the efficiency of resource allocation, f time (x) represents the rationality of task time allocation, w1, w2, w3 are weight parameters, which are dynamically adjusted according to task requirements; S42. Generate an optimized UAV task decomposition plan based on the collaborative behavior of the leader and followers of the elk group, and perform the task decomposition on the leader L of each family. i Using the fitness value f of the task decomposition scheme fitness (L i ) determines its influence. The leader solution represents the current optimal task decomposition strategy. j Update task breakdown: in, is the task decomposition scheme for the follower in the tth generation, is the leader’s task decomposition scheme in the tth generation, reflecting the current optimal instruction decomposition method. λ is the step size parameter, which controls the adjustment range of the task decomposition scheme. A new set of candidate UAV task decomposition schemes S is generated by updating. candidate ; S43. Generate a new task decomposition scheme through the interaction between the leader and the follower, optimize the decomposition of the UAV's task instructions, and in each family, based on the leader L i and follower F j The interaction generates a new task decomposition scheme x new : Among them, x new represents the newly generated task decomposition plan, η1 is the interaction factor, which indicates the degree of innovation of the task plan; The newly generated task decomposition scheme is optimized according to the actual operation requirements of the UAV and added to the candidate task decomposition scheme set S candidate ; S44. Filter the candidate task decomposition solution set S according to the specific requirements of the UAV task candidate , select the best solution with the highest fitness: Among them, x optimal The optimal task decomposition scheme includes task target priority, resource allocation and path planning. The candidate schemes are sorted according to the fitness value and the selected scheme is suitable for the current mission scenario of the UAV; S45. Decompose the optimal task into x optimal The output is sent to the UAV mission execution module to instruct the UAV to perform operations according to the decomposed mission plan.

6. The command decomposition method for a touch screen for an unmanned aerial vehicle according to claim 1, characterized in that: The S5 step includes: S51. According to the optimal task decomposition plan x optimal For the complete UAV mission instruction set D instruction,complete Decompose the task, extract the task objectives, task areas and task constraints in the optimal task decomposition plan, decompose the task objectives into several independent sub-objectives, divide them according to the priority and area range of the sub-objectives, and Contains specific task operations, execution areas, and constraints, generating multiple sets of executable subtasks: S52. Prioritize the executable subtask set and process high-priority subtasks first: in, is the i-th subtask, represents the adaptability membership of the subtask goal, Indicates the urgency of the subtask. Indicates the degree of dependence of the subtask on other tasks. w4, w5, and w6 are weight parameters for priority sorting, which are dynamically adjusted according to task requirements. S53. Allocate resources to the executable subtask set, calculate the resource allocation plan according to the current resource status of the drone, and allocate specific resources to each subtask according to the calculation results: in, Assign to subtask Resources, R available is the total amount of resources currently available to the drone, ρ i For subtask The resource demand weight is determined by the task complexity and area scope, n is the total number of subtasks, and the current resource status includes power, flight time, and load capacity; S54. Integrate the priority sorting and resource allocation results into a set of executable subtasks for the drone D subtasks : Each subtask contains task objectives, priorities, and resource allocation plans.

7. A touch screen for an unmanned aerial vehicle, comprising an internally embedded memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for decomposing instructions for a touch screen for an unmanned aerial vehicle according to any one of claims 1 to 6 is implemented.

Images