Heterogeneous ground support network deployment method for unmanned aerial vehicle autonomous inspection

Abstract

The invention provides a heterogeneous ground support network deployment method for unmanned aerial vehicle autonomous inspection, which comprises the following steps: constructing a construction cost estimation model of an unmanned aerial vehicle ground support network based on three parts of ground nodes, antennas and transformation cost; based on the construction cost estimation model, according to two constraint conditions of unmanned aerial vehicle air-ground communication transmission time delay and ground network node deployment positions, a problem model is constructed for an unmanned aerial vehicle ground support network construction cost minimization target; and for the constructed problem model, a node search algorithm with the maximum cluster coverage rate is adopted to solve the problem, and the deployment design of the unmanned aerial vehicle autonomous inspection ground support network under the condition of a given unmanned aerial vehicle inspection task coverage area is realized. According to the method, the problems of construction and deployment of the unmanned aerial vehicle ground support network under the condition of a given inspection task coverage area can be solved, and automatic optimization of the deployment position of the network node is realized.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle inspection and wireless network communication, and in particular to a heterogeneous ground support network deployment method for unmanned aerial vehicle autonomous inspection. Background Art

[0002] Drone inspections have become an important scientific and technological support force for machines to replace people in various fields such as transportation, electricity, agriculture, and forestry. In recent years, with the development and upgrading of various types of machine nest products, the construction of ground-based support facilities for drones has been accelerated, and the full realization of autonomous drone flight has become an important implementation path for the development of the low-altitude economy. However, carrying out drone inspections in remote areas with high mountains and dense forests will face the problem of severe attenuation of communication signals and limited autonomous flight distance of drones. For example, although the drone equipment itself can support an inspection range of up to 10 kilometers, in mountainous environments such as Hubei, Hunan, and Sichuan, when drones are conducting forestry and electricity inspections, the autonomous flight distance in severe cases does not even exceed 1 kilometer.

[0003] Using the 4G / 5G communication services provided by mobile operators to provide communication services for drones and ground nests is an important method to expand the autonomous flight distance. However, since drone inspections are often used in special fields, data such as drone location and mission content need to consider transmission security, and mobile public network transmission methods are difficult to meet privacy protection requirements. At the same time, the design of operator base stations is mainly to provide services to people on the ground. The coverage area and orientation of base station antennas are different from the flight characteristics of drones. Therefore, the present invention considers using wireless self-organizing network communication technology to provide a ground support network for drone inspection services, and to expand the flight coverage of autonomous drone inspections by building a local communication network.

[0004] Taking into account issues such as network construction, operating costs and drone flight safety, how to reduce costs as much as possible by optimizing the number and location of network node deployment while meeting the constraints of ground support network coverage and transmission rate has become an important issue that needs to be urgently considered in engineering construction. Summary of the invention

[0005] The present invention provides a method for deploying a heterogeneous ground support network for autonomous inspection of unmanned aerial vehicles. The technical problem it solves is the difficulty of building and deploying a ground support network for unmanned aerial vehicles under the condition of a given inspection task coverage area. The present invention constructs a ground construction cost estimation model for the ground support network for autonomous inspection of unmanned aerial vehicles, considers the differentiated inspection support capabilities of two types of network nodes, namely, machine nests and relay communications, and designs a node search algorithm that maximizes group coverage, thereby realizing automatic optimization of the deployment position of network nodes.

[0006] The specific technical solutions of the present invention are as follows:

[0007] A method for deploying a heterogeneous ground support network for autonomous inspection by unmanned aerial vehicles includes the following steps:

[0008] Step 1: Based on the three parts of ground nodes, antennas and transformation costs, a construction cost estimation model for the UAV ground support network is constructed;

[0009] Step 2: Based on the construction cost estimation model designed in step 1, according to the two constraints of the UAV air-to-ground communication transmission delay and the ground network node deployment location, a problem model is constructed to minimize the construction cost of the UAV ground support network;

[0010] Step 3: For the problem model constructed in step 2, a node search algorithm with maximum cluster coverage is used to solve the problem, and the deployment design of the UAV autonomous inspection ground support network is realized under the conditions of a given UAV inspection mission coverage area.

[0011] Furthermore, in step 1, a construction cost estimation model for the UAV ground support network is constructed, which specifically includes:

[0012] For a network with a given topology design, the construction cost estimation model is broken down into the following parts:

[0013]

[0014] in, represents the network topology, and Respectively The node set and link set of An estimation function representing the overall construction cost of the ground support network, They are used for cost assessment of relay node construction, machine nest construction, directional antenna construction and reconstruction mechanism implementation within the network; the construction cost of machine nests and relay nodes is calculated based on the following formula:

[0015]

[0016] c n and c m Represent the deployment unit price of the machine nest and relay node respectively; and Used to indicate the selection of candidate locations, as long as The i-th position in The ground nodes selected for deploying the UAV ground support network are otherwise If the location is used to deploy machine nest equipment, there is otherwise If used to deploy relay nodes, then otherwise In addition to these two cases, there are

[0017] remember The unit price of antenna construction is The antenna construction cost expression is as follows:

[0018]

[0019] The relay node candidate position is upgraded to the machine nest candidate position to improve the scalability of the network, and the minimum reconstruction cost of a single node is recorded as c r (·), the implementation cost of the reconstruction strategy is expressed as:

[0020]

[0021] Furthermore, the problem model constructed in step 2 specifically includes: introducing R(v r ):v1→v2→…→v r To describe the information transmission link between the UAV and the ground support network, v1 and v r are the machine nest and inspection task point on the corresponding transmission path respectively, and there are represents the set of task points inspected by drones; assuming that the antenna transmission power of each node is given, the data transmission rate between each link is calculated based on the number of neighboring nodes and the link length. In total, Available transmission paths, the path with the shortest transmission delay is By definition, this path is expressed as:

[0022]

[0023] In engineering construction, the network construction problem will provide an optimization solution for minimizing the estimated construction cost. Combined with the above definition, the problem model for minimizing the construction cost of the UAV ground support network is expressed as follows:

[0024]

[0025] Among them, constraint (6a) is the transmission delay constraint of the UAV air-to-ground communication, which means that when the UAV hovers at each inspection task point, the communication transmission delay of its control signaling must meet the threshold requirement for system security considerations; constraint (6b) is the deployment position constraint of the ground network node, which means that the drone nest and the network relay node cannot be deployed at the same candidate position.

[0026] Furthermore, step 3 specifically includes:

[0027] The network performance evaluation function is introduced and expressed as follows:

[0028]

[0029] Represents the number of inspection points in the UAV inspection task set that can be covered by the ground support network;

[0030] A new data structure named cluster is defined. A cluster refers to a ground network structure consisting of a machine nest and multiple relay nodes, where each relay node can access the machine nest within a limited number of hops and may contain multiple access paths.

[0031] All candidate deployment locations are combined with visible links to form a candidate ground network, and multiple candidate clusters are formed from each machine nest node. Each candidate cluster is a subset of the candidate ground network. When a machine nest node is given, since a given machine nest node can support access to multiple surrounding relay nodes, the candidate cluster containing the machine nest may still have multiple topological connection forms. The candidate location is selected. The mass formed as the machine nest is recorded as Since the data transmission delay increases with the increase of the number of ground hops, but the number of inspection task points that can be covered by the group structure of a given machine nest node does not continue to increase with the increase of the number of relay nodes, it is better to choose a group with lower construction cost to cover a single task point; since the drone can only be controlled by one nearby machine nest in principle when hovering at each task point, the single-hop link of "machine nest-machine nest" cannot exist in the transmission path R(v r ) in which the node location selection and topology design of the network construction problem are transformed into the optimal clique search problem;

[0032] The cluster coverage maximization node search algorithm implements the selection of ground node deployment locations. Specifically, a cluster structure search that can meet the maximum network efficiency is implemented for each machine nest candidate location in the machine nest candidate set. In the initial stage, each machine nest candidate cluster only contains the candidate node itself. Then, the visible neighbor relay nodes that can make the network efficiency growth rate of the machine nest candidate cluster the highest are selected in turn to join the candidate cluster. Finally, the nodes with the highest network efficiency among all candidate clusters are selected to be retained, and all nodes in the candidate cluster are deleted from the ground network node candidate location set. The process is iterated until all candidate location sets are empty, or the drone inspection task points have been fully covered by the ground support network.

[0033] The advantages of the present invention are:

[0034] (1) A heterogeneous ground support network deployment method for UAV autonomous inspection is proposed, which can solve the problem of construction and deployment of UAV ground support network under given inspection coverage area conditions.

[0035] (2) A ground support network construction method considering two types of network nodes, drone nests and relay communications, was proposed. The modification costs of antenna and nest deployment locations were taken into account in the network construction cost evaluation model. The node deployment location selection and drone support plan were in line with the actual engineering construction, which can provide good technical support for application implementation.

[0036] (3) A node search algorithm that maximizes cluster coverage is designed. Through two rounds of iterations, the automatic calculation of the network node deployment location can be realized. While meeting the coverage target of the drone inspection mission points, the network construction cost is reduced. At the same time, the network connection links can be automatically generated according to the connectivity between nodes, and the node deployment location selection and topology structure design are realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a schematic diagram of the group search process of the present invention;

[0038] Figure 2 It is a flow chart of the node search algorithm for maximizing cluster coverage of the present invention.

[0039] Figure 3 It is a schematic diagram of the application effect of the network construction and deployment method proposed in the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] The embodiment of the present invention provides a method for deploying a heterogeneous ground support network for autonomous inspection by unmanned aerial vehicles, comprising the following steps:

[0042] Step 1: Based on the three parts of ground nodes, antennas and transformation costs, a construction cost estimation model for the UAV ground support network is constructed;

[0043] Step 2: Based on the construction cost estimation model constructed in step 1, according to the two constraints of the UAV air-to-ground communication transmission delay and the ground network node deployment location, a problem model is constructed for the goal of minimizing the construction cost of the UAV ground support network;

[0044] Step 3: For the problem model constructed in step 2, a node search algorithm with maximum cluster coverage is used to solve the problem, and the deployment design of the UAV autonomous inspection ground support network is realized under the conditions of a given UAV inspection mission coverage area.

[0045] Furthermore, the construction cost estimation model of the UAV ground support network is constructed in step 1, which specifically includes:

[0046] The nodes of the drone ground support network need to realize the connection between the drone and the ground nodes at the same time. Considering the mobility of the drone and the static characteristics of the ground nodes, two types of antennas, omnidirectional antennas and directional antennas, are used in the ground support network to realize air-to-ground and ground-to-ground communication service support. Considering the actual construction feasibility of the project, in the present invention, the nodes of the drone ground support network are intended to be constructed in specific locations that can meet the construction conditions, such as village committees, power poles, etc. Therefore, the optional deployment locations of all ground support network nodes are intended to be recorded as The set of nodes selected to deploy the relay network is denoted as The ground support network includes two types of nodes: drone nests and relay communication. Since the nests for drones to start and stop often have high requirements for power supply and network supply, the candidate locations for the nest deployment are: A subset of The set of all actual machine nest deployment positions is recorded as Considering that in actual projects, the candidate machine nest deployment points cannot fully meet the needs of drone inspection tasks, other points need to be modified, such as by laying additional light.

[0047] Based on the above, for a network with a given topology design, the construction cost estimation model can be broken down into the following parts:

[0048]

[0049] in, represents the network topology, and Respectively The node set and link set of . According to the previous design content, there are is an estimation function representing the overall construction cost of the ground support network, They are used for cost assessment of relay node construction, machine nest construction, directional antenna construction and reconstruction mechanism implementation within the network. Specifically, the construction cost of machine nests and relay nodes can be calculated based on the following formula:

[0050]

[0051] cn and c m Represent the deployment unit price of the machine nest and relay node respectively. and Used to indicate the selection of candidate locations, as long as The i-th position in The ground nodes selected for deploying the UAV ground support network are otherwise If the location is used to deploy machine nest equipment, then otherwise If used to deploy relay nodes, then otherwise Apart from these two cases, we have

[0052] Therefore, remember The unit price of antenna construction is We can give the following expression for the antenna construction cost:

[0053]

[0054] According to the design described above, the candidate locations of relay nodes can be upgraded to candidate locations of machine nests to improve the scalability of the network. The minimum reconstruction cost of a single node is recorded as c r (·), then the implementation cost of the reconstruction strategy can be expressed as:

[0055]

[0056] The specific process of problem modeling in step 2 includes:

[0057] Introducing R(v r ):v1→v2→…→v r To describe the information transmission link between the UAV and the ground support network, v1 and v r are the machine nest and inspection task point on the corresponding transmission path respectively, and there are Represents the set of task points inspected by drones. Assuming that the antenna transmission power of each node is given, the data transmission rate between each link can be calculated based on the number of neighboring nodes and the link length. In total, Available transmission paths, the path with the shortest transmission delay is By definition, this path can be expressed as

[0058] In engineering construction, the network construction problem will provide an optimization solution for minimizing the estimated construction cost. Combined with the above definition, the problem of ground support network construction for drone inspection can be expressed as follows:

[0059]

[0060]

[0061] Among them, constraint (6a) is a communication constraint, which means that when the drone hovers at each inspection task point, the communication transmission delay of its control signaling must meet the threshold requirement for system security considerations; constraint (6b) is a location constraint, which means that the drone nest and the network relay node cannot be deployed at the same candidate location.

[0062] The solution method described in step 3 is to solve the problem of building the UAV ground support network constructed in step 2. The specific process is to introduce a newly constructed network performance evaluation function, which is expressed as follows:

[0063]

[0064] Represents the number of inspection points in the drone inspection task set that can be covered by the ground support network. Next, a new data structure is defined, named group. Specifically, a group refers to a ground network structure that includes a machine nest and multiple relay nodes, where each relay node can access the machine nest within a limited number of hops and may contain multiple access paths.

[0065] like Figure 1 As shown in the figure, all candidate deployment locations combined with visible links can form a candidate ground network, and multiple candidate cluster structures can be formed from each machine nest node, and each candidate cluster is a subset of the candidate ground network. When a machine nest node is given, since it can support the access of multiple surrounding relay nodes, the candidate cluster containing the machine nest may still have multiple topological connection forms. For ease of expression, the candidate location is selected. The mass formed as the machine nest is recorded as Since the data transmission delay increases with the increase of the number of ground hops, the number of inspection task points that can be covered by the cluster structure of a given nest node will not continue to increase with the increase of the number of relay nodes. Therefore, choosing a cluster with lower construction cost to cover a single task point will become a better option. Since the drone can only be controlled by a nearby nest in principle when hovering at each task point, the "nest-nest" single-hop link cannot exist in the transmission path R(v r ). In this way, the node location selection and topology design of the network construction problem can be transformed into the optimal clique search problem.

[0066] Based on this, we can give Figure 2The cluster coverage maximization node search algorithm shown in the figure is used to implement the selection of the deployment location of the ground node. Specifically, the algorithm needs to implement a cluster structure search that can meet the maximum network efficiency for each machine nest candidate location in the machine nest candidate set. In the initial stage, each machine nest candidate cluster only contains the candidate node itself. Then, the visible neighbor relay nodes that can make the network efficiency growth rate of the machine nest candidate cluster the highest are selected in turn to join the candidate cluster. Finally, the nodes with the highest network efficiency among all candidate clusters are selected to be retained, and all nodes in the candidate cluster are deleted from the set of candidate locations for ground network nodes. This process is iterated until all candidate location sets are empty, or the drone inspection task points have been fully covered by the ground support network.

[0067] Figure 3 The application effect analysis of the present invention is given through experimental simulation. Among them, three types of deployment location candidate sets are set, namely All, Quarter, and Half, which contain 440, 200, and 100 candidate locations in a given space respectively. The method of the present invention is denoted as GMS, and compared with the genetic algorithm, denoted as GA-2. In addition, the number of nodes of the candidate machine nest is selected respectively. Is 10 and 20. Figure 3 (a) shows the comparison of network performance evaluation results. The results show that in all scenarios, the network performance of the method proposed in the present invention is better than that of the comparison algorithm. Figure 3 (b) shows a comparative analysis of the coverage of drone inspection points by the constructed network after different methods for network deployment location selection. From the results, except when the number of machine nest nodes is 10, the coverage rate of the inspection points by the genetic algorithm is slightly greater than that of the method described in the present invention. In all other scenarios, the coverage support capability of the present invention for drone inspections is better.

[0068] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for deploying a heterogeneous ground support network for autonomous inspection by unmanned aerial vehicles, characterized in that: The following steps are involved: Step 1: Based on the three parts of ground nodes, antennas and transformation costs, a construction cost estimation model for the UAV ground support network is constructed; Step 2: Based on the construction cost estimation model designed in step 1, according to the two constraints of the UAV air-to-ground communication transmission delay and the ground network node deployment location, a problem model is constructed to minimize the construction cost of the UAV ground support network; Step 3: For the problem model constructed in step 2, a node search algorithm with maximum cluster coverage is used to solve the problem, and the deployment design of the UAV autonomous inspection ground support network is realized under the conditions of a given UAV inspection mission coverage area.

2. The method for deploying a heterogeneous ground support network for autonomous inspection by unmanned aerial vehicles according to claim 1, characterized in that: Furthermore, in step 1, a construction cost estimation model for the UAV ground support network is constructed, which specifically includes: For a network with a given topology design, the construction cost estimation model is broken down into the following parts: in, represents the network topology, and ε G Respectively The node set and link set of An estimation function representing the overall construction cost of the ground support network, They are used for cost assessment of relay node construction, machine nest construction, directional antenna construction and reconstruction mechanism implementation within the network; the construction cost of machine nests and relay nodes is calculated based on the following formula: c n and c m Represent the deployment unit price of the machine nest and relay node respectively; and Used to indicate the selection of candidate locations, as long as The i-th position in The ground nodes selected for deploying the UAV ground support network are otherwise If the location is used to deploy machine nest equipment, there is otherwise If used to deploy relay nodes, then otherwise In addition to these two cases, there are remember The unit price of antenna construction is The antenna construction cost expression is as follows: The relay node candidate position is upgraded to the machine nest candidate position to improve the scalability of the network, and the minimum reconstruction cost of a single node is recorded as c r (·), the implementation cost of the reconstruction strategy is expressed as:

3. The method for deploying a heterogeneous ground support network for autonomous inspection by unmanned aerial vehicles according to claim 2, characterized in that: The step 2 of constructing the problem model specifically includes: Introducing R(v r ):v1→v2→…→v r To describe the information transmission link between the UAV and the ground support network, v1 and v r are the machine nests and inspection task points on the corresponding transmission paths, and represents the set of task points inspected by drones; assuming that the antenna transmission power of each node is given, the data transmission rate between each link is calculated based on the number of neighboring nodes and the link length. In total, Available transmission paths, the path with the shortest transmission delay is By definition, this path is expressed as: In engineering construction, the network construction problem will provide an optimization solution for minimizing the estimated construction cost. Combined with the above definition, the problem model for minimizing the construction cost of the UAV ground support network is expressed as follows: Among them, constraint (6a) is the transmission delay constraint of the UAV air-to-ground communication, which means that when the UAV hovers at each inspection task point, the communication transmission delay of its control signaling must meet the threshold requirement for system security considerations; constraint (6b) is the deployment position constraint of the ground network node, which means that the drone nest and the network relay node cannot be deployed at the same candidate position.

4. The method for deploying a heterogeneous ground support network for autonomous inspection by unmanned aerial vehicles according to claim 3, characterized in that: Step 3 specifically includes: The network performance evaluation function is introduced and expressed as follows: Represents the number of inspection points in the UAV inspection task set that can be covered by the ground support network; A new data structure named cluster is defined. A cluster refers to a ground network structure consisting of a machine nest and multiple relay nodes, where each relay node can access the machine nest within a limited number of hops and may contain multiple access paths. All candidate deployment locations are combined with visible links to form a candidate ground network, and multiple candidate cluster structures are formed from each machine nest node. Each candidate cluster is a subset of the candidate ground network. The mass formed as the machine nest is recorded as Choosing a group with lower construction cost to cover a single task point becomes a better option; the node location selection and topology design of the network construction problem is transformed into the search problem of the optimal group; The cluster coverage maximization node search algorithm implements the selection of ground node deployment locations. Specifically, a cluster structure search that can meet the maximum network efficiency is implemented for each machine nest candidate location in the machine nest candidate set. In the initial stage, each machine nest candidate cluster only contains the candidate node itself. Then, the visible neighbor relay nodes that can make the network efficiency growth rate of the machine nest candidate cluster the highest are selected in turn to join the candidate cluster. Finally, the nodes with the highest network efficiency among all candidate clusters are selected to be retained, and all nodes in the candidate cluster are deleted from the ground network node candidate location set. The process is iterated until all candidate location sets are empty, or the drone inspection task points have been fully covered by the ground support network.

Images