A method, device, system and storage medium for selecting heterogeneous networks of unmanned aerial vehicles

Abstract

The present invention discloses a method and device, system and storage medium for selecting heterogeneous networks of unmanned aerial vehicles, including: step S1, constructing a heterogeneous network selection model for multiple unmanned aerial vehicles in a confined space; step S2, optimizing the heterogeneous network selection model for multiple unmanned aerial vehicles in a confined space by multiple communication indicators; wherein, weighting multiple indicators is performed to convert the multi-objective optimization problem into a single objective, and the network satisfaction is obtained to evaluate the access network quality, and the optimal network selection decision for the unmanned aerial vehicle is obtained. The technical solution of the present invention can achieve balanced optimization of the transmission rate, delay and switching cost of unmanned aerial vehicles in a confined space while meeting the needs of multiple types of services.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicles, and in particular, relates to a method and device, a system, and a storage medium for selecting a heterogeneous network of an unmanned aerial vehicle. Background Art

[0002] In recent years, disasters have frequently occurred in confined spaces such as subways, garages, and mines due to insufficient supervision capabilities and unstable geological structures. Faced with the risk of accidents in confined spaces such as fires, collapses, and mine rock bursts, it is urgent to improve daily monitoring levels and emergency response capabilities. By establishing efficient and stable data collection and transmission communication links, real-time and accurate data support is provided for disaster warning and emergency rescue. Due to objective constraints such as complex geological conditions and poor communication conditions in confined spaces, fixed sensors have a small monitoring range and limited communication capabilities, making it difficult to achieve comprehensive perception of the scene and rapid return of information. The rapid development of technologies such as simultaneous localization and mapping (SLAM) and visual inertial navigation has enabled drones to achieve accurate flight positioning, navigation, and obstacle avoidance in confined spaces. Lightweight, high-load, and high-endurance drones are equipped with intelligent sensing equipment and communication equipment to collect and return on-site information in real time, effectively improving the breadth and accuracy of spatial perception and providing a new solution for environmental monitoring and safety inspections.

[0003] At present, underground spaces, confined structures and other restricted scenarios are covered by 4G and 5G mobile communication systems and IEEE802.11a / b / g / n / ac / ax standard Wi-Fi wireless networks to meet the communication needs of users in diverse application scenarios. When drones transmit different types of business data such as audio and video, the network status they are in evolves rapidly with their own movement, and they need to select an access network that meets their business needs based on the surrounding network environment. Traditional network selection mechanisms usually use the strength of the network signal received by the user to make judgments, and cannot adaptively switch networks in real time according to business changes. Therefore, it is necessary to design a business-driven heterogeneous network selection method based on different business types, balance and optimize communication indicators such as transmission rate, latency, and switching cost, and ensure the stability of drone communication links.

[0004] At present, the network selection of UAVs is mainly evaluated by the received signal strength, and the continuity and stability of user communication are ensured by reducing unnecessary network switching. Reference [1] (J. Xie, W. Gao and C. Li, "Heterogeneous network selection optimization algorithm based on a Markov decision model," in China Communications, vol. 17, no. 2, pp. 40-53, Feb. 2020) proposed a network selection optimization algorithm based on the Markov decision process, which avoids the ping-pong effect of network switching by predicting the network status. Reference [2] (H. Yang, B. Hu and L. Wang, "A deep learning based handover mechanism for UAV networks," 2017 20th International Symposium on Wireless Personal Multimedia Communications (WPMC), Bali, Indonesia, 2017, pp. 380-384) uses deep learning to predict the movement trajectory of the UAV, thereby improving the success rate of network switching. For heterogeneous network scenarios, the paper [3] (D. Guo, L. Tang, X. Zhang and Y.-C. Liang, "Joint Optimization of Handover Control and Power Allocation Based on Multi-Agent Deep Reinforcement Learning," in IEEE Transactions on Vehicular Technology, vol. 69, no. 11, pp. 13124-13138, Nov. 2020) designed a deep reinforcement learning (DRL) framework based on the proximal policy optimization (PPO) algorithm, which balances the user's received signal strength and network switching frequency.Reference [4] (L. Qiang, J. Li and E. Altman, "A Novel Distributed Network Selection Scheme for Heterogeneous Wireless Network Environments," in IEEE Transactions on Control of Network Systems, vol. 4, no. 3, pp. 575-586, Sept. 2017) combines energy efficiency, signal strength, network cost, delay and bandwidth to establish a utility function, characterizes user preferences and network performance indicators, and uses a heuristic multi-criteria algorithm to solve the network selection problem from vehicle to basic network facilities. The above algorithms focus on the network selection problem of a single drone. However, for multi-drone application scenarios, it is necessary to design a more efficient and stable network selection algorithm to ensure the network quality of all drones.

[0005] Due to the complexity and particularity of confined spaces, the selection of heterogeneous networks for drones also needs to consider the following issues: (1) The impact of the complex propagation environment in confined spaces on signal propagation. (2) The degree of adaptability of multi-access heterogeneous networks to the differentiated needs of multiple types of services. (3) The high real-time requirements for decision-making due to the mobility of drones and the time-varying nature of services. Summary of the invention

[0006] The technical problem that Benjamin wants to solve is to provide a method and device, system, and storage medium for selecting heterogeneous networks for drones, which can achieve balanced optimization of the transmission rate, delay, and switching cost of drones in confined spaces while meeting the needs of multiple types of services.

[0007] To achieve the above object, the present invention adopts the following technical solution:

[0008] A method for selecting a heterogeneous network of a UAV, comprising:

[0009] Step S1, constructing a constrained space multi-UAV heterogeneous network selection model;

[0010] Step S2, optimizing the multi-communication indexes of the heterogeneous network selection model for multiple UAVs in constrained space; wherein, weighting multiple indexes is performed to transform the multi-objective optimization problem into a single objective, and the network satisfaction is obtained to evaluate the access network quality, thereby obtaining the optimal network selection decision for the UAV.

[0011] As a preferred option, the heterogeneous network of multiple drones in a confined space selects multiple communication indicators to be optimized as follows:

[0012]

[0013]

[0014]

[0015] in, is the minimum transmission rate of service s, is the maximum delay, v is the UAV, and d is the accessible network.

[0016] Preferably, in step S2, by combining IFAHP with the CRITIC weighted algorithm and assigning combined weights to each indicator, the multi-objective optimization problem is converted into a single-objective binary integer linear programming problem, and a single-objective optimization function is constructed.

[0017] The present invention also provides a UAV heterogeneous network selection device, comprising:

[0018] A building block for constructing a heterogeneous network selection model for multiple UAVs in constrained spaces;

[0019] The selection module is used to optimize the multi-communication indicators of the heterogeneous network selection model of multiple UAVs in constrained space; among them, the multi-objective optimization problem is transformed into a single objective by empowering multiple indicators, and the network satisfaction is obtained to evaluate the access network quality, and the optimal network selection decision of the UAV is obtained.

[0020] As a preferred option, the heterogeneous network of multiple drones in a confined space selects multiple communication indicators to be optimized as follows:

[0021]

[0022]

[0023]

[0024] in, is the minimum transmission rate of service s, is the maximum delay, v is the UAV, and d is the accessible network.

[0025] As a preference, the selection module transforms the multi-objective optimization problem into a single-objective binary integer linear programming problem by combining IFAHP with the CRITIC weighted algorithm and assigning combined weights to each indicator, thereby constructing a single-objective optimization function.

[0026] The present invention also provides a UAV heterogeneous network selection system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a UAV heterogeneous network selection method when executed by the processor.

[0027] The present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the UAV heterogeneous network selection method when running.

[0028] Taking into account the mobility of drones and the complex propagation characteristics of confined spaces, the requirements of multiple types of services for network performance are used as constraints to construct a multi-objective combinatorial optimization problem for heterogeneous network selection of drones. The present invention converts it into a single-objective binary integer linear programming problem based on the combined weighted algorithm of IFAHP-CRITIC. By calculating and comparing the network quality satisfaction of the accessible networks around the drones, the network selection decision of multiple drones is obtained to improve the stability of the drone communication link. The technical solution of the present invention can achieve a balanced optimization of the transmission rate, delay and switching cost of drones in confined spaces while meeting the needs of multiple types of services. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0030] Figure 1 This is a flow chart of a method for selecting a heterogeneous network for a drone according to an embodiment of the present invention;

[0031] Figure 2 It is a schematic diagram of a heterogeneous network selection model for multiple UAVs in a constrained space according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0033] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] Embodiment 1:

[0035] like Figure 1 As shown, an embodiment of the present invention provides a method for selecting a heterogeneous network of a drone, comprising:

[0036] Step S1, constructing a constrained space multi-UAV heterogeneous network selection model;

[0037] Step S2, optimizing the multi-communication indexes of the heterogeneous network selection model for multiple UAVs in constrained space; wherein, weighting multiple indexes is performed to transform the multi-objective optimization problem into a single objective, and the network satisfaction is obtained to evaluate the access network quality, thereby obtaining the optimal network selection decision for the UAV.

[0038] As an implementation method of the present invention, in step S1, the constrained space multi-UAV heterogeneous network selection model is as follows: Figure 2 As shown. Define the drone collection as V is the total number of drone nodes, and network nodes are Indicates that the network includes two types: 5G and Wi-Fi6. and Respectively represent 5G base stations (Basedstations, BSs) and WiFi6 access points (Access points, APs), and satisfy In addition, different types of business data sets are represented as

[0039] The embodiment of the present invention is based on a multimode channel model and a geometrical optics (GO) model, and comprehensively considers the influence of factors such as the ceiling, walls, and ground on signal propagation, and derives the electromagnetic field distribution of the excitation plane of the confined space communication cross-section. By weighted summing up the field strengths of all modes, an estimated value of the confined space channel gain is obtained.

[0040] Assume that the transmitter power at (x0,y0,0) is Pt, the electric field strength is E0, and the received signal power at the coordinate (x,y,z) is estimated to be for:

[0041]

[0042] Among them, GU and GN represent the antenna gains of the drone and access device respectively. is the electric field strength of mode (m,n), C mn , α mn and β mn Respectively represent the modal intensity, attenuation coefficient and phase shift coefficient of each mode in the excitation plane, then the channel gain is:

[0043] With reference to 5G NR and IEEE 802.11ax communication protocols, and taking into account the communication network resources and business requirements in restricted scenarios, network transmission rate, latency, switching latency and other indicators are comprehensively adopted to perform network selection and balancing optimization.

[0044] According to the 3GPP Release 16 standard, the approximate theoretical data transmission rate of a drone accessing a 5G base station is as follows (2):

[0045]

[0046] Among them, v Layers is the maximum number of data streams (layers), Q m is the maximum modulation order, f is the proportional factor, R max is the maximum target code rate, μ is the subcarrier spacing parameter set, is the average duration of an OFDM symbol in a subframe of μ, is the maximum number of RWs allocated for a bandwidth BW corresponding to μ, where BW is the maximum bandwidth supported by the UE in the frequency band or frequency band combination, and OH refers to the spectrum overhead.

[0047] When the drone is connected to a WiFi6 access point, according to the 802.11ax standard, the ideal data transmission rate is expressed as:

[0048] dr wifi =(N×M×R×S) / T s (3)

[0049] Among them, N represents the number of subcarriers, M represents the number of subcarrier codes, R represents the coding rate, S represents the spatial flow rate, and Ts represents the system time.

[0050] Due to the existence of large and small scale fading in the wireless channel, the approximate actual data rate is:

[0051] dr=dr′·ξ (4)

[0052] The delay of drone v includes transmission delay Processing delay Queuing delay and propagation delay

[0053]

[0054] in, is the transmission delay of UAV v, which refers to the time required for data to reach the receiver from the transmitter through the transmission medium, that is, L represents the size of the data packet, d r v is the data transmission rate of drone v accessing the network. The transmission delay is related to the transmission rate and has been reflected in indicator 1. Therefore, the transmission delay is no longer considered when calculating the delay of drone v:

[0055]

[0056] For the same type of network, the data processing process is the same, so the processing delay is a constant. In addition, the propagation delay of mode (m,n) is:

[0057]

[0058] Where f represents the frequency of the transmitted signal. The total propagation delay of drone v is expressed as:

[0059]

[0060] When a drone chooses to access a new network, it needs to be authenticated and authorized, and establish a connection with the new network by obtaining the corresponding network configuration information. The network switching delay generated by this process can be expressed as:

[0061]

[0062] Where k is the number of switching attempts. is the single network switching delay of UAV v. In addition, if the UAV does not switch networks, k is 0.

[0063] As an implementation method of an embodiment of the present invention, in step S2, taking into account the above communication indicators, the embodiment of the present invention constructs a heterogeneous network selection multi-objective combined optimization problem to find the optimal network selection decision for drones. To simplify the problem, a combined weighted algorithm is first designed to weight multiple indicators, convert the multi-objective optimization problem into a single objective, and obtain a unified utility function (i.e., network satisfaction) to evaluate the access network quality.

[0064] In heterogeneous network scenarios, drones need to meet different business requirements and jointly optimize the three indicators of transmission rate, transmission delay and switching delay to improve the overall network satisfaction. In addition, considering the limited hardware configuration of drones and the limited network resources in confined space, this patent gives the following settings: (1) Drones can only access one network at the same time. (2) Drones accessing the same network share the network bandwidth equally. (3) Each network has a strict limit on the number of access users to ensure that drones can maintain communication through the allocated minimum bandwidth. Based on the above conditions, the heterogeneous network selection problem of multiple communication indicators for multiple drones in confined space is constructed, which can be expressed as:

[0065]

[0066]

[0067]

[0068] Among them, the minimum transmission rate of service s is The maximum delay is The three indicators in the multi-objective optimization problems (10-a), (10-b), and (10-c) conflict with each other, and a single indicator cannot be optimized, and it is difficult to find the optimal solution that weighs the indicators. In order to simplify the optimization function, the embodiment of the present invention combines the IFAHP (intuitionistic fuzzy analytic hierarchy process) and the CRITIC (the criteria importance through inter-crieria correlation) weighting algorithm, and converts the multi-objective optimization problem into a single-objective binary integer linear programming problem by assigning a combined weight to each indicator, thereby constructing a single-objective optimization function.

[0069] IFAHP effectively reflects the different preferences of services for various communication indicators; CRITIC comprehensively considers the differences and connections between communication indicators and can reflect the impact of time-varying channel states on communication performance.

[0070] Further, the IFAHP-based weighting method in step S2 includes:

[0071] Step 211: construct an intuitive fuzzy preference matrix for service s based on the network performance preferences obtained from the priorities of different indicators. Table 1 shows the preference level for indicator i in business s. and unprefer level In addition, it is stipulated that i,j=1,...,S, Where S is the number of indicators.

[0072] Table 1

[0073]

[0074] Step 212: Construct a perfect multiplicatively consistent intuitive preference relationship and obtain an intuitive fuzzy preference consistency matrix As shown in equation (11):

[0075]

[0076]

[0077]

[0078] Step 213: By transforming the intuitionistic fuzzy preference matrix H s Compared with the intuitionistic fuzzy preference consistency matrix, the distance metric is obtained Only when When , the intuitive fuzzy preference matrix given is consistent. τ is the consistency threshold. For the HNS-USU model, the embodiment of the present invention directly provides the preference relationship between various indicators, so that the intuitive fuzzy preference consistency matrix obtained in step 2 is The consistency condition must be met.

[0079] Step 214: Calculate the consistent IFV weight of service s The specific expressions are as follows:

[0080]

[0081] Step 215: Calculate the hesitation γ of the HNS-USU model using the weight obtained in step 214, as shown in formula (13):

[0082]

[0083] By combining uncertainty and IFV weights, the subjective preference vector is calculated Expressed as:

[0084]

[0085] in, The larger the value of , the more the decision maker prefers indicator i. Therefore, the preference vector Can be quantified as subjective weights:

[0086]

[0087] Further, the CRITIC-based weighting method in step S2 includes:

[0088] Step 221: In order to eliminate the influence of different dimensions on the evaluation results, each indicator needs to be dimensionless. Define the perception matrix in It represents the estimated value of communication index i when UAV v accesses network d.

[0089] The CRITIC method generally uses forward or reverse processing rather than standardization. For benefit indicators such as communication rate, MinMax normalization (MMN) is used, while for cost indicators such as delay and switching cost, reverse minimum-maximum normalization (RMMN) is used. The specific processing method is as follows:

[0090]

[0091] Here, i represents the type of indicator (i=1,...,K), and d represents the accessible network (d=1,...,D). It reflects the normalized values ​​of each indicator and forms a normalized matrix It is particularly important to note that if but

[0092] Step 222: Calculate the index difference. The standard deviation of the ith index can be expressed as:

[0093]

[0094] in, is the normalized mean of index i, given by express.

[0095] Step 223: Calculate the index conflict. First, use formula (29) to solve the correlation coefficient between index i and i'.

[0096]

[0097] The conflict formula is further derived:

[0098]

[0099] Step 224: The amount of information for each indicator can be expressed as:

[0100]

[0101] The larger the value of Cj is, the greater the role of the jth evaluation index in the entire evaluation index system. Therefore, the objective weight of drone v is expressed as:

[0102]

[0103] Further, the combined weighted method in step S2 is as follows:

[0104] In summary, the combined weight of the communication index i is expressed as:

[0105]

[0106] in, represents the weight of the communication index i based on the IFAHP algorithm, Represents the weight of the communication index i based on the CRITIC algorithm.

[0107] By adding the combined weights to the network performance perception standardized matrix and weighting each indicator, the performance evaluation matrix of the drone v executing service s is obtained:

[0108]

[0109] Therefore, formula (10-a) can be transformed into:

[0110]

[0111] For multiple indicator constraints (10-b), define the communication indicator requirement satisfaction flag:

[0112]

[0113] in, Indicates whether the drone v executing service s meets the communication constraints when accessing network d. If not, the utility function value of network d is 0.

[0114] By joint And formula (13) to get the network performance satisfaction of drone v executing service s to access network d:

[0115]

[0116] For the multi-UAV network selection problem, the overall network performance can be expressed by the sum of the performance satisfaction of the UAV access networks: Therefore, the multi-objective optimization problem (10) for UAV network selection can be transformed into:

[0117]

[0118] In the embodiment of the present invention, after each timestamp, the drone performs a network perception and switching decision. By using this algorithm, the drone can switch to the most suitable access network during movement or service switching. The specific implementation is as follows: First, the location of the network access device in the underground heterogeneous scene and the number of drone service data S are initialized. Then, each drone v calculates the combined weight of the drone v for the service s according to the current situation, such as its location coordinates, the currently connected network device, and the ongoing service type, and evaluates the network satisfaction of the surrounding accessible networks. After that, it is easy to find the network that can provide the greatest utility and determine whether the restrictions of the network access device are met after switching. In other words, this step ensures that the drone can obtain the minimum communication bandwidth from the access network. Finally, the optimal matching network of the drone v that meets the conditions is obtained. Based on the above analysis, the algorithm is summarized in Algorithm 1.

[0119]

[0120]

[0121] In summary, the embodiments of the present invention achieve a balance among transmission rate, delay and switching cost by comprehensively considering the service preferences of the drone while meeting the normal communication needs of the drone, thereby improving the average access network performance satisfaction.

[0122] The embodiment of the present invention studies the heterogeneous network selection problem of multiple UAVs in constrained scenarios, and constructs a multi-objective combined optimization model for network selection in combination with the complex propagation characteristics of confined spaces. For the optimization problem of multiple communication indicators, taking into account the requirements of different types of services for communication indicators, a combined weighted method based on IFAHP and CRITIC is designed to convert the multi-objective optimization equation into a single objective to more intuitively evaluate the access network quality, thereby obtaining real-time and high-quality network selection decisions. While ensuring the basic communication requirements of UAVs, the embodiment of the present invention can better achieve balanced optimization among the three performances of transmission rate, transmission delay and switching cost. Compared with network selection algorithms based on signal strength, the embodiment of the present invention has stronger stability and robustness, and can better solve the network selection problem of UAVs in confined spaces.

[0123] Embodiment 2:

[0124] The embodiment of the present invention further provides a UAV heterogeneous network selection device, comprising:

[0125] A building block for constructing a heterogeneous network selection model for multiple UAVs in constrained spaces;

[0126] The selection module is used to optimize the multi-communication indicators of the heterogeneous network selection model of multiple UAVs in constrained space; among them, the multi-objective optimization problem is transformed into a single objective by empowering multiple indicators, and the network satisfaction is obtained to evaluate the access network quality, and the optimal network selection decision of the UAV is obtained.

[0127] As an implementation method of an embodiment of the present invention, the heterogeneous network of multiple drones in a confined space selects multiple communication indicators for optimization as follows:

[0128]

[0129]

[0130]

[0131] in, is the minimum transmission rate of service s, is the maximum delay.

[0132] As an implementation method of an embodiment of the present invention, the selection module converts the multi-objective optimization problem into a single-objective binary integer linear programming problem by combining the IFAHP and CRITIC weighted algorithms and assigning combined weights to each indicator, thereby constructing a single-objective optimization function.

[0133] Embodiment 3:

[0134] An embodiment of the present invention further provides a UAV heterogeneous network selection system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a UAV heterogeneous network selection method when executed by the processor.

[0135] Embodiment 4:

[0136] An embodiment of the present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the UAV heterogeneous network selection method when running.

[0137] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for selecting heterogeneous networks of unmanned aerial vehicles, characterized in that: include: Step S1, constructing a constrained space multi-UAV heterogeneous network selection model; Step S2, optimizing the heterogeneous network selection model of multiple UAVs in constrained space for multiple communication indicators of heterogeneous network selection of multiple UAVs in constrained space; wherein, weighting multiple indicators is performed to transform the multi-objective optimization problem into a single objective, and the network satisfaction is obtained to evaluate the access network quality, and the optimal network selection decision of the UAV is obtained; The problem of selecting multiple communication indexes to build a heterogeneous network of multiple UAVs in a constrained space is expressed as: In the multi-objective optimization problem (10-a), the three optimization equations correspond to maximizing data rate, minimizing delay, and minimizing switching cost respectively; (10-b) is the constraint of the optimization problem. For the first and second, the minimum transmission rate of service s is The maximum delay is The normal transmission conditions of the service are given; the third one is to ensure that each drone only accesses one network, while the fourth one is to limit the number of drones accessing each network; (10-c) is a flag used to mark whether drone v is connected to network d; Combining IFAHP with CRITIC weighted algorithm, the multi-objective optimization problem is transformed into a single-objective binary integer linear programming problem by assigning combined weights to each indicator, and a single-objective optimization function is constructed; The weighting method based on IFAHP includes: Step 211: construct an intuitive fuzzy preference matrix for service s based on the network performance preferences obtained from the priorities of different indicators. is the preference level and non-preference level of indicator i in business s; Step 212: Construct a perfect multiplication-consistent intuitive preference relationship and obtain an intuitive fuzzy preference consistency matrix As shown in equation (11): in in, and are the preference and non-preference levels after consistency processing, is an auxiliary variable; Step 213: By transforming the intuitionistic fuzzy preference matrix H s Compared with the intuitionistic fuzzy preference consistency matrix, the distance metric is obtained Step 214: Calculate the consistent IFV weight of service s Step 215: Calculate the hesitation γ of the HNS-USU model using the weights obtained in step 214. By combining uncertainty and IFV weights, the subjective preference vector is calculated in, The larger the value of, the more the decision maker prefers indicator i, and the preference vector Can be quantified as subjective weights: The CRITIC-based weighting method in step S2 includes: Step 221: Perform dimensionless processing on each indicator and define the perception matrix in, It represents the estimated value of the communication index i when the drone v accesses the network d. The MinMax normalization (MMN) is used for the communication rate index, and the reverse minimum-maximum normalization (RMMN) is used for the delay and switching indicators. The specific processing method is as follows: Where i represents the type of indicator (i=1,...,K), d represents the accessible network (d=1,...,D), It reflects the normalized values ​​of each indicator and forms a normalized matrix Step 222: Calculate the index difference, the standard deviation of the ith index It can be expressed as: in, is the normalized mean of index i, given by express; Step 223: Calculate the index conflict. First, use formula (18) to solve the correlation coefficient r between index i and i'. i,i′ ; Derivation of conflict Step 224: Information volume for each indicator The product of standard deviation and conflict can be expressed as: The larger the value of Cj is, the greater the role of the jth evaluation index in the entire evaluation index system. The objective weight of UAV v is expressed as: The combined weight of communication index i is expressed as: in, represents the weight of the communication index i based on the IFAHP algorithm, represents the weight of the communication index i based on the CRITIC algorithm; By adding the combined weights to the network performance perception standardized matrix and weighting each indicator, the performance evaluation matrix of the drone v executing service s is obtained: in, represents the estimated value of the communication index i when the UAV v is connected to the network d after dimensionless processing; Formula (10-a) can be transformed into: For multiple indicator constraints (10-b), define the communication indicator requirement satisfaction flag: in, Indicates whether the drone v executing service s meets the communication restrictions when accessing network d; By joint And formula (13) to get the network performance satisfaction of drone v executing service s to access network d: For the multi-UAV network selection problem, the overall network performance is represented by the sum of the performance satisfaction of the UAVs accessing the network: The multi-objective optimization problem of drone network selection ((10-a), (10-b), (10-c)) is transformed into:

2. A drone heterogeneous network selection device for implementing the drone heterogeneous network selection method according to claim 1, characterized in that: include: A building block for constructing a heterogeneous network selection model for multiple UAVs in constrained spaces; The selection module is used to optimize the multi-communication indicators of the heterogeneous network selection model of multiple UAVs in constrained space; among them, the multi-objective optimization problem is transformed into a single objective by empowering multiple indicators, and the network satisfaction is obtained to evaluate the access network quality, and the optimal network selection decision of the UAV is obtained.

3. A UAV heterogeneous network selection system, characterized in that: include: A memory and a processor, wherein the memory stores a computer program executed by the processor, and when the computer program is executed by the processor, the method for selecting a heterogeneous network of a drone as claimed in claim 1 is executed.

4. A storage medium, characterized in that: The storage medium stores a computer program, which executes the UAV heterogeneous network selection method as claimed in claim 1 when running.

Images