An Optimization Method for Resource Allocation Scheme of Amphibious Unmanned Vehicles for Mission Orientation

Abstract

This invention discloses a resource allocation optimization method for amphibious unmanned vehicles (UAVs) oriented towards specific tasks, belonging to the technical field of amphibious UAV resource allocation optimization. This invention comprehensively considers usage scenarios such as water areas, land areas, and water-land areas, as well as tasks such as rapid patrol and efficient cargo carrying. It selects appropriate amphibious UAV locomotives and their control strategies, and chooses to install suitable payloads, configure reasonable energy and supplies, thereby determining the optimal resource allocation scheme for the amphibious UAV oriented towards specific tasks. Ultimately, this guides users in rapid configuration. Based on task requirements, this invention utilizes a pre-simulated cost mapping function and multi-objective optimization techniques to obtain the optimal resource allocation scheme for amphibious UAVs in typical scenarios. This shortens the calculation time and cost of optimization objectives for amphibious UAVs under given tasks and scenarios, improves users' rapid decision-making ability during application, and significantly reduces reliance on human experience.

Description

Technical Field

[0001] This invention belongs to the technical field of amphibious unmanned vehicle resource allocation optimization, specifically relating to a method for optimizing resource allocation schemes for amphibious unmanned vehicles in a mission-oriented manner. Background Technology

[0002] Amphibious unmanned vehicles (UAVs) possess the capability to navigate on land, water, and transition between land and water, and are widely used in civilian sectors for flood control and disaster relief, coastal patrol, material transport in water-prone areas, and beach recreation. The use of amphibious vehicles is subject to changing mission priorities and scenario requirements, making resource allocation increasingly complex. Determining the specific type of locomotive and its control strategy, the payload configuration, and the amount of energy carried to better adapt to application scenarios and maximize task completion presents a significant challenge to the rapid decision-making capabilities of users. For example, when used for emergency rescue and resupply, amphibious vehicles require not only a large payload but also sufficient land-water mobility; when used for border and coastal patrol, they require high speed, flexible land-water transition, long-range detection equipment, and strong detection capabilities. Meanwhile, the environment in which amphibious vehicles are used will also affect the configuration of their running gear and control strategies. For example, when amphibious vehicles are mainly used in hard road environments, their running gear tends to be wheeled, while when amphibious vehicles are mainly used in sandy or muddy environments, their running gear tends to be triangular tracks or strip tracks.

[0003] Different applications and usage scenarios require platforms to have different functions, and the needs vary in different complex scenarios. There are one or more requirements, such as minimizing energy consumption, fastest completion time, and highest task completion rate. Traditionally, the selection of the locomotive mechanism, payload selection, and energy configuration of amphibious unmanned vehicles often rely on human experience. This problem itself is a multi-objective optimization problem, and these requirements are often coupled and mutually exclusive. The NSGA-III algorithm is a multi-objective optimization algorithm that supports mixed inputs of continuous and discrete variables and quickly solves multi-objective optimization problems with Pareto dominance. Summary of the Invention

[0004] (I) Purpose of the Invention

[0005] The purpose of this invention is to provide a resource allocation optimization method for amphibious unmanned vehicles (UAVs) oriented towards missions. The technical problem this invention aims to solve is that existing amphibious UAVs, when performing missions across land and water, rely on human experience to determine the selection of their locomotion mechanism and the configuration of their mission payloads. This invention provides a resource allocation optimization method for amphibious UAVs oriented towards missions. This optimization method comprehensively considers the usage scenario (water, land, and water-land areas, etc.) and the mission (rapid patrol, efficient cargo carrying, etc.), selects a suitable amphibious UAV locomotion mechanism and its control strategy, selects appropriate payloads (electro-optical reconnaissance payloads, robotic arm payloads, etc.), and configures reasonable energy sources (fuel, batteries, etc.) and supplies to determine the optimal resource allocation scheme for amphibious UAVs oriented towards missions, ultimately guiding users to carry out rapid configuration.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, this invention comprehensively considers the mission and usage environment of amphibious unmanned vehicles, and uses a pre-simulation estimation of the travel cost function and an improved over-target optimization algorithm to optimize the resource (walking mechanism and control strategy, load selection, energy supply, etc.) configuration scheme of the unmanned vehicle, proposing a mission-oriented resource configuration optimization method for amphibious unmanned vehicles.

[0008] The present invention adopts the following technical solution:

[0009] An optimization method for resource allocation schemes of amphibious unmanned vehicles oriented towards tasks includes the following steps:

[0010] Step 1: Pre-construct a traffic mapping map for the unit distance travel cost of amphibious unmanned vehicles within three domains (water area, land area, and water-land area). r (m,v,α,β);

[0011] In the formula, m is the amphibious unmanned vehicle's load mass (including energy, payload, etc.), v is the amphibious unmanned vehicle's speed, α is the walking mechanism and its control strategy, β is the traffic condition coefficient, representing the difficulty of passing through the road section, representing wind and wave conditions in water, representing slope conditions in the water-land section, and representing road conditions in the land area, and r represents one of the three passage sections of land area, water area, and water-land section respectively;

[0012] Step 2: Define the parameters to be configured for the amphibious unmanned vehicle resource configuration scheme;

[0013] The variables that need to be configured include: 1. Amphibious unmanned vehicle payload or resource load. jThe first variable determines the payload mass of the amphibious unmanned vehicle. Different resource configurations result in different payloads; for example, a lidar payload is approximately 10 kg, while a refueling payload might be 500 kg, causing drastic changes in the payload and time-sensitive targets. The second variable is the amphibious unmanned vehicle's walking mechanism and control strategy α, which determines how the vehicle traverses special terrains. The third variable is the amphibious unmanned vehicle's speed v on various road sections. i The speed of the amphibious unmanned vehicle depends not only on the configuration of the walking mechanism and the mass of the payload, but also on the environmental road conditions in which the mission is performed; the subscript i indicates that the amphibious unmanned vehicle is traveling at different stages, and the subscript j indicates the j-th resource;

[0014] Step 3: Based on specific task requirements and the amphibious unmanned vehicle's usage scenario, define a multi-objective optimization problem for resource allocation, with the following optimization performance indicators:

[0015]

[0016] st0 <v i <V r

[0017]

[0018] In the formula: J1, J2, and J3 are three optimization objectives in the multi-objective optimization problem of amphibious unmanned vehicle resource allocation; J1 is the loss objective: that is, minimizing the overall travel cost of the task. Assume the task can be divided into N segments based on the travel interval, and the travel cost of the i-th segment is traffic. r i k t i J1 represents the mileage of each segment in the task; J2 represents the time objective: the time cost to complete the entire task, and the travel time k for the i-th segment. t i / v i Various tasks typically have the goal of completing them as quickly as possible, i.e., minimizing time costs; J3 represents the task objective: that is, resource allocation should be sufficient to support the amphibious unmanned vehicle in achieving its mission, load j Indicates mass as For the j-th type, select a suitable payload or supply resource, where M is the upper limit of amphibious unmanned vehicle resource configuration. When selecting a payload j that matches the mission, load... j =1, otherwise load j =0;

[0019] By comprehensively optimizing the above objective function, the task-oriented resource allocation optimization problem is transformed into a multi-objective optimization problem;

[0020] Step 4: Optimize the crossover, mutation, perturbation methods and probabilities of the NSGA-III algorithm;

[0021] Step 5: Solve the multi-objective resource allocation problem for amphibious unmanned vehicles based on the improved NSGA-III algorithm;

[0022] The process of solving the multi-objective optimization problem based on the above multi-objective function is as follows:

[0023] S51. Initialize the total number of iterations N gen Population size N grp Random initialization includes N grp The initial population P of individuals t ;

[0024] S52. Begin the i-th iteration: Starting from the parent population P t Crossover, mutation, and other operations are performed to generate offspring populations.

[0025] Q t Merge the parent and offspring populations into a new population of size 2N. grp population R t =P t +Q t ;

[0026] S53. Based on the multi-objective function, apply R... t Perform a non-dominated sort to obtain non-dominated layer sorts F1, F2, ..., and then place the members F1, ..., F1 in S sequentially. t Go to the middle;

[0027] S54. If |S t |=N grp Then proceed to the next iteration, and the next generation population P... t+1 =S t Otherwise, F l+1 Individuals at the front of the layer are added to P in sequence. t+1 In, until the number of its individuals reaches N. grp ;

[0028] S55. Repeat S2 until i ≥ N gen

[0029] S56. Obtain the Pareto optimal solution set in the population. n is the number of non-dominated solutions;

[0030] Step 6: Obtain the optimal solution for amphibious unmanned vehicle resource allocation;

[0031] Using a weight-based method to obtain the optimal solution from non-dominated solutions, the following evaluation is performed:

[0032]

[0033] In the formula, σ1, σ2, and σ3 are weighting factors, and the weight values ​​of each are determined according to the task requirements. For example, in disaster relief, the primary consideration is the requirements of speed and resupply, so σ2 and σ3 have relatively larger weights. However, in routine patrols, the requirement is to complete the task while controlling costs, so σ1 and σ3 have relatively larger weights.

[0034] Regarding the unit distance travel cost mapping for amphibious unmanned vehicles within the three domains in step 1, traffic... r (m i ,v i The calculation of α,β) involves varying the load, speed, locomotive mechanism and control strategy, and road conditions in water, land, and waterway environments. Nested loops are used to obtain the mapping relationship of the unit distance travel cost of the unmanned vehicle under different road conditions. r In this case, when the load is too heavy to pass through a certain section of the road, the cost is set to infinity.

[0035] (III) Effective Returns

[0036] Compared with the prior art, the advantages of the present invention are:

[0037] 1. By utilizing the present invention, the optimal configuration scheme of amphibious unmanned vehicle resources in typical scenarios can be obtained based on task requirements, using a pre-simulated estimated cost mapping function and multi-objective optimization technology. This reduces the calculation time and cost of optimization objectives for amphibious unmanned vehicles under given tasks and scenarios, improves the user's ability to make quick decisions during application, and also significantly reduces the reliance on human experience.

[0038] 2. This invention comprehensively considers the influence of water, land and water-land environments on amphibious unmanned vehicles, as well as the influence of resource allocation such as different walking mechanisms and strategies of amphibious unmanned vehicles, to obtain the optimal resource allocation scheme and maximize the mission efficiency of amphibious unmanned vehicles. Attached Figure Description

[0039] Figure 1 It is a resource allocation optimization framework based on unit distance cost mapping;

[0040] Figure 2 This is a flowchart illustrating the present invention; Detailed Implementation

[0041] The following is in conjunction with the appendix Figure 1 , Figure 2 The invention will be explained and illustrated in more detail through case studies of island rescue and resupply operations.

[0042] This invention proposes a method to estimate the unit distance travel cost of amphibious unmanned vehicles (UAVs) in three domains through pre-simulation, combine it with a multi-objective function constructed for the task, and use an improved NSGA-III algorithm to quickly obtain an optimized resource allocation scheme for amphibious UAVs oriented towards the task. Unlike the traditional method of determining the type of walking mechanism, load, energy, and speed based on usage experience, this patent fully considers the usage environment and task characteristics, and evaluates the optimal resource allocation scheme for the UAV through accurate simulation, ensuring that the configuration scheme can maximize the utilization efficiency of the amphibious UAV.

[0043] In one ideal implementation of the present invention, a resource allocation optimization method for amphibious unmanned vehicles oriented towards tasks includes the following steps:

[0044] Step 1: Pre-construct a traffic mapping for the unit distance travel cost of amphibious unmanned vehicles within three domains: water, land, and the water-land area. r (m i ,v i ,α,β);

[0045] In the simulation software, load simulations of amphibious unmanned vehicles are set up in three domains, and nested loop tests are performed. i ,v i The motion of the amphibious unmanned vehicle under different values ​​of four variables, α, β, was recorded, including its motion energy consumption, motion bumpiness, and vehicle control accuracy; where m i The payload mass of the amphibious unmanned vehicle (including energy, load, etc.) is incremented in 50kg increments until the maximum payload of the amphibious unmanned vehicle is reached; v i The travel speed of the amphibious unmanned vehicle is incremented by 1 m / s to the upper limit of the vehicle's speed in each of the three domains. α represents the walking mechanism and its control strategy, including wheeled mechanism, triangular track mechanism, long track mechanism and its control strategy, etc. β is the traffic condition coefficient, representing the difficulty of passing through the road section. In water, it represents the headwind and tailwind values ​​on the water, incremented by 1 m / s to the upper limit of the amphibious unmanned vehicle's design. In the water-land section, it represents the slope degree, incremented by 1° to the upper limit of the amphibious unmanned vehicle's design.

[0046] Step 2: Define the parameters to be configured for the amphibious unmanned vehicle resource configuration scheme;

[0047] In this case study, we consider island rescue and resupply missions and usage scenarios. Variable 1: Load or resource load. j For resupplying materials, the algorithm is not concerned with the type of resupply, but with the final resupply mass; Variable 2: the walking mechanism control strategy α, the mechanism configuration scheme for the amphibious unmanned vehicle to land and leave the island, which can be wheeled, triangular tracked, or long tracked; Variable 3: the speed v of the amphibious unmanned vehicle on each road segment. i ;

[0048] Step 3: Based on specific task requirements and the amphibious unmanned vehicle usage scenario, define a multi-objective optimization problem for resource allocation, with the following optimization performance indicators:

[0049]

[0050] st0 < v i <V r

[0051]

[0052] In the formula: J1, J2, and J3 are three optimization objectives in the multi-objective optimization problem of amphibious unmanned vehicle resource allocation; J1 is the loss objective: that is, minimizing the overall travel cost of the task. Assume the task can be divided into N segments based on the travel interval, and the travel cost of the i-th segment is traffic. r i k t i J1 represents the mileage of each segment in the task; J2 represents the time objective: the time cost to complete the entire task, and the travel time k for the i-th segment. t i / v i Various tasks typically have the goal of completing them as quickly as possible, i.e., minimizing time costs; J3 represents the task objective: that is, resource allocation should be sufficient to support the amphibious unmanned vehicle in achieving its mission, load j Indicates mass as For the j-th type, select a suitable payload or supply resource, where M is the upper limit of amphibious unmanned vehicle resource configuration. When selecting a payload j that matches the mission, load... j =1, otherwise load j =0;

[0053] In an embodiment of the present invention, the above task flow is divided into eight segments: land load, land-water inter-situation load, water load, land-water inter-situation load (landing), land-water inter-situation empty load (leaving the island), water empty load, land-water inter-situation empty load, and land empty load, i.e., N=8; where when N≥4, the load mass is reduced by the supply mass;

[0054] By comprehensively optimizing the above objective function, the task-oriented resource allocation optimization problem is transformed into a multi-objective optimization problem;

[0055] Step 4: Optimize the crossover, mutation, perturbation methods and probabilities of the NSGA-III algorithm;

[0056] Step 5: Solve the multi-objective resource allocation problem for amphibious unmanned vehicles based on the improved NSGA-III algorithm;

[0057] The process of solving the multi-objective optimization problem based on the above multi-objective function is as follows:

[0058] S51. Initialize the total number of iterations N gen =1000, population size N grp =80, random initialization includes N grp The initial population P of individuals t ;

[0059] S52. Begin the i-th iteration: Starting from the parent population P t Crossover, mutation, and other operations are performed to generate a progeny population Q. t Merge the parent and offspring populations into a new population of size 2N. grp population R t =P t +Q t ;

[0060] S53. Based on the multi-objective function, apply R... t Perform a non-dominated sort to obtain non-dominated layer sorts F1, F2, ..., and then place the members F1, ..., F1 in S sequentially. t Go to the middle;

[0061] S54. If |S t |=N grp Then proceed to the next iteration, and the next generation population P... t+1 =S t Otherwise, F l+1 Individuals at the front of the layer are added to P in sequence. t+1 In, until the number of its individuals reaches N. grp ;

[0062] S55. Repeat S2 until i ≥ N gen

[0063] S56. Obtain the Pareto optimal solution set in the population. n is the number of non-dominated solutions;

[0064] Step 6: Obtain the optimal solution for amphibious unmanned vehicle resource allocation;

[0065] Using a weight-based method to obtain the optimal solution from non-dominated solutions, the following evaluation is performed:

[0066]

[0067] In the formula, σ1, σ2, and σ3 are weighting factors, and the weight values ​​of each are determined according to the task requirements. For example, in disaster relief, the primary consideration is the requirements of speed and resupply, so σ2 and σ3 have relatively large weights. However, in routine patrols, the requirement is to complete the task while controlling costs, so σ1 and σ3 have relatively large weights.

[0068] The above description, in conjunction with specific embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A method for optimizing task-oriented resource allocation scheme of amphibious unmanned vehicle, characterized in that, Comprising the following steps: Step 1, pre-constructing the amphibious unmanned vehicle unit distance passing cost mapping in the water area, land area and water-land area ; wherein, is the load mass of the amphibious unmanned vehicle in each road section, is the travel speed of the amphibious unmanned vehicle in each road section, is the traveling mechanism of the amphibious unmanned vehicle and its control strategy, is the passing condition coefficient of the amphibious unmanned vehicle, which represents the difficulty of passing the road section, and represents the wave condition in water area, the road condition on land, and the slope condition in the water-land section; respectively represents one of the three sections of the passing section, i.e. the land section, the water section, and the water-land section. Step 2, for amphibious unmanned vehicle resource configuration scheme, define the parameters to be configured; The variables that need to be configured include 1: amphibious unmanned vehicle load or resource for determining the load quality of the amphibious unmanned vehicle, different resources have different qualities; 2: Amphibious unmanned vehicle walking mechanism and control strategy , the variable determines the walking mode of the amphibious unmanned vehicle through special terrain; 3: The speed of the amphibious unmanned vehicle on each road section , the speed of the amphibious unmanned vehicle, depends on the configuration of the walking mechanism and the mass of the load, also depends on the environment road of the task execution; subscript indicates that the amphibious unmanned vehicle is running in different stages, subscript indicates the th resource; Step 3, according to the task demand and amphibious unmanned vehicle use scene, define the resource configuration multi-objective optimization problem, and the optimization performance index is: In the formula: Three optimization objectives in the multi-objective optimization problem of amphibious unmanned vehicle resource allocation, Loss objective: that is, minimizing the whole-process traffic cost of the task, assuming that the task can be divided into N segments according to the traffic interval, and the traffic cost of the i-th segment is traffic r i , k t i The mileage of each segment in the task; Time efficiency objective: that is, the time cost of completing the whole task, and the traffic time of the i-th segment is k t i / v i That is, minimizing the time cost; Task objective: that is, the resource allocation can support the amphibious unmanned vehicle to achieve the task, Indicates that the j-th type of suitable load or supply resource with quality Amphibious unmanned vehicle resource allocation upper limit, selecting a load matching the task When , otherwise ;​ Comprehensive optimization of the above objective function, the task-oriented resource configuration optimization problem is converted into a multi-objective optimization problem; Step 4, the cross, mutation, disturbance mode and probability of NSGA-III algorithm are optimized; Step 5, the amphibious unmanned vehicle resource configuration multi-objective based on the improved NSGA-III algorithm is solved; Based on the above multi-objective function, the multi-objective optimization solution is as follows: S51. Initialize the total number of iterations , population size , randomly initialize an initial population containing individuals ; S52. Start the first iteration Next iteration: Generate a child population from the parent population by performing crossover and mutation operations Merge the parent and child populations into a new population of size ;​ S53. performing non-dominated sorting on the multi-objective function to obtain non-dominated layer sorting S54. putting members in the non-dominated layer into in turn;​​ S54. If then go to the next iteration, and the next generation population , otherwise, add individuals in front of the layer one by one into until the number of individuals reaches ; S55. Repeat S52 until S56. Obtain the Pareto optimal solution set in the population n is the number of non-dominated solutions. Step 6, obtain the optimal solution of amphibious unmanned vehicle resource configuration scheme; The method based on weight is used as the optimal solution from the non-dominated solution, and the evaluation is: In the formula, are weight factors, and the weight values of each item are determined according to the task requirements.

2. The method of claim 1, wherein, The step 1 two-domain amphibious unmanned vehicle unit distance passing cost mapping The calculation changes the load, driving speed, walking mechanism and control strategy, road condition in the water area, land area and water-land area environment respectively, and adopts nested loop to obtain the unmanned vehicle unit distance passing cost mapping relationship under different road conditions ; wherein, when the load is too heavy to pass a certain road, the cost is set to infinity.

3. The method of claim 1, wherein, The load is an optical reconnaissance load or a mechanical arm load.

4. The method of claim 1, wherein, The task demand is patrol or load.

Images