An air-ground channel modeling and simulation method integrating UAV characteristics

Abstract

The present invention discloses a method for simulating an air-ground channel integrating the characteristics of unmanned aerial vehicles, including: establishing an air-ground channel model for unmanned aerial vehicles; inputting the types of air-ground communication scenarios for unmanned aerial vehicles, carrier frequencies, position vectors and velocity vectors of unmanned aerial vehicles / receiving ends / scatterers at the initial moment, flight attitudes, antenna half-power beam widths, and fuselage structure data sets; calculating the Rice factor; calculating the power coefficients of the line-of-sight path and the n th non-line-of-sight path; obtaining the path loss of the ground segment through neural network training; calculating the shadow fading between the receiving end and the transmitting end, and calculating the body shadow coefficients of the line-of-sight path and the non-line-of-sight path by using the fuselage structure, flight attitude and half-power beam width of the antenna; inputting the Rice factor, time delay, phase, power coefficient, path loss and shadow fading into the air-ground channel model for unmanned aerial vehicles to output channel fading. The present invention can achieve efficient and accurate simulation and reproduction of the air-ground channel, and support the evaluation and optimization of unmanned aerial vehicle communication systems.

Description

Technical Field

[0001] The present invention belongs to the technical field of modeling and generating of wireless fading channels, and particularly relates to a method for modeling and simulating an air-ground channel integrating the characteristics of unmanned aerial vehicles (UAVs). Background Art

[0002] UAVs have the advantages of convenient operation, low cost, strong flexibility, etc., and have been widely used in various communication scenarios such as aerial base stations, relay communication, emergency communication, and battlefield communication. In the above communication scenarios, as the medium for signal transmission, the channel directly affects the transmission quality and performance of the communication system, and is of great significance for designing, optimizing, and evaluating UAV communication systems. Different from traditional terrestrial mobile communications, UAVs have characteristics such as fast flight speed and rotation of the fuselage attitude, and traditional mobile channel models are difficult to comprehensively describe the new characteristics of UAV channels. Therefore, constructing an air-ground channel model considering the unique attributes of UAVs has important engineering application value.

[0003] Currently, most channel modeling methods mainly focus on the speed change of the mobile terminal. However, the channel characteristics of UAVs are closely related to the flight altitude and flight attitude. In addition, most air-ground channel models for UAVs regard UAVs as point masses and ignore the influence of the fuselage structure on the channel.

[0004] In view of this, the present invention proposes a more realistic method for modeling and simulating the UAV air-ground channel. This method comprehensively considers factors such as the fuselage structure, flight altitude, and flight attitude of UAVs at the modeling level, and specifically gives the generation and calculation methods of channel parameters at the simulation level, realizing the efficient and accurate reproduction of the UAV air-ground channel. Summary of the Invention

[0005] Aiming at the deficiencies of the above-mentioned existing technologies, the purpose of the present invention is to provide a method for modeling and simulating an air-ground channel integrating the characteristics of UAVs, so as to solve the problem of inaccurate calculation of channel parameters such as power caused by ignoring the characteristics of UAVs such as self-occlusion and fuselage rotation in the existing technologies.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0007] A method for modeling and simulating an air-ground channel integrating the characteristics of UAVs of the present invention comprises the following steps:

[0008] 1) Considering the fuselage structure, flight altitude, and flight attitude of the UAV, establish an air-ground channel model of the UAV;

[0009] 2) Input the type of UAV air-ground communication scenario, carrier frequency, position vectors and velocity vectors of the UAV / receiver / scatterer at the initial moment, flight attitude, antenna half-power beam width, and fuselage structure data set;

[0010] 3) Calculate the Rice factor using the communication scenario type and carrier frequency; calculate the time delay and phase of the line-of-sight path and the nth non-line-of-sight path using the initial position vector and velocity vector; further calculate the power coefficients of the line-of-sight path and the nth non-line-of-sight path.

[0011] 4) Divide the path propagation loss between the receiver and the transmitter into an air segment and a ground segment; calculate the path loss of the air segment based on the flight altitude of the UAV, and use neural network training to obtain the path loss of the ground segment.

[0012] 5) Calculate the shadow fading between the receiver and the transmitter using the flight altitude and communication scenario type, and calculate the body shadow coefficients of the line-of-sight path and the non-line-of-sight path using the body structure, flight attitude, and half-power beam width of the antenna.

[0013] 6) Input the Rice factor, time delay, phase, power coefficient, path loss, and shadow fading calculated above into the UAV air-to-ground channel model to output the channel fading.

[0014] Further, the UAV air-to-ground channel model in step 1) is as follows:

[0015]

[0016] Among them, H represents the channel fading, PL represents the path propagation loss, SF represents the shadow fading, and h(t, τ) represents the channel impulse response between the transceiver, and the specific expression is as follows:

[0017]

[0018] Among them, K(t) is the Rice factor, C LoS (t) and C n (t) are the body shadow coefficients of the line-of-sight path and the nth non-line-of-sight path respectively, P LoS (t) and P n (t) are the power coefficients of the line-of-sight path and the nth non-line-of-sight path respectively, Φ LoS (t) and Φ n (t) are the phases of the line-of-sight path and the nth non-line-of-sight path respectively, τ LoS (t) and τ n (t) are the time delays of the line-of-sight path and the nth non-line-of-sight path respectively, N is the number of non-line-of-sight paths, and j is the imaginary unit.

[0019] In particular, the influence of the body structure on the channel is mainly reflected in the calculation process of C LoS (t), the influence of the flight attitude on the channel is mainly reflected in the calculation process of C n (t), and the influence of the flight altitude on the channel is mainly reflected in the calculation processes of PL and SF.

[0020] Further, the method for calculating the Rice factor in the UAV air-ground communication channel in step 3) is as follows:

[0021] K(t) = ρ K χ K (t) + κ K log 10 f c + μ K (3)

[0022] Wherein, the scene parameters μ K , ρ K , κ K and χ K (t) are closely related to the type of communication scenario; μ K and ρ K respectively represent the mean and variance of the Rice factor, χ K (t) represents a random variable obeying a Gaussian distribution with a mean of 0 and a variance of 3, κ K represents the frequency-dependent factor, and f c represents the carrier frequency.

[0023] Further, the method for calculating the time delay and phase of the line-of-sight path and the nth non-line-of-sight path in step 3) is as follows:

[0024] 31) Calculate the position vectors of the receiver, transmitter, and scatterer at time t, and the method is as follows:

[0025]

[0026] Wherein,

[0027]

[0028] Wherein, v Tx / Rx / S represents the modulus value, and φ Tx / Rx / S and ψ Tx / Rx / S respectively represent the azimuth angle and elevation angle of the moving speed;

[0029] 32) Calculate the distance vectors of the line-of-sight path and the nth non-line-of-sight path at time t, and the method is as follows:

[0030] 321) Use the position vectors of the receiver and transmitter obtained in step 31) to calculate the distance vector of the line-of-sight path at time t, and the method is as follows:

[0031]

[0032] Wherein, · represents the Euclidean distance of a vector, e p and e qrespectively represent the vectors from the positions of the receiving end and the transmitting end at the initial moment to the positions of the receiving end and the transmitting end at time t, d LoS (t 0 ) represents the propagation distance of the line-of-sight path at the initial moment, and the calculation method is as follows:

[0033] d LoS (t 0 ) = L Tx (t 0 ) - L Rx (t 0 ) (6)

[0034] 322) Use the position vectors of the receiving end and the transmitting end obtained in step 31) to calculate the distance vector of the nth non-line-of-sight path at time t, and the method is as follows:

[0035] d n (t) = a n (t) - b n (t) (7)

[0036] where, a n (t) represents the vector from the UAV transmitting end to the scatterer, and is specifically expressed as a n (t) = L Tx (t) - L S (t), b n (t) represents the vector from the ground receiving end to the scatterer, and is specifically expressed as b n (t) = L Rx (t) - L S (t);

[0037] 33) According to the distance vector of the line-of-sight path and the distance vector of the nth non-line-of-sight path in step 32), calculate the propagation delay and phase of the line-of-sight path and the nth non-line-of-sight path at time t, and the method is as follows:

[0038]

[0039]

[0040] where, c represents the speed of light, and λ represents the signal wavelength.

[0041] Furthermore, in step 3), calculate the power coefficients of the line-of-sight path and the nth non-line-of-sight path, and the method is as follows:

[0042] By normalizing all power coefficients, the final value of the power coefficient of the line-of-sight path is 1. At the same time, the power calculation method of the nth non-line-of-sight path is as follows:

[0043]

[0044] Among them, represents the power of the nth non-line-of-sight path before normalization, and the calculation method is as follows:

[0045]

[0046] Among them, r τ represents the time delay scaling coefficient, G n follows a Gaussian distribution with a mean of 0 and a variance of 3, and σ τ represents the random time delay spread.

[0047] Furthermore, the specific steps of step 4) include:

[0048] 41) Calculate the propagation loss PL of the line-of-sight path LoS , and the method is as follows:

[0049] PL LoS = 20log 10 ||d LoS || + 20log 10 f c + c 1 (12)

[0050] Among them, c 1 represents the scene-related correction coefficient;

[0051] 42) Calculate the propagation loss of the non-line-of-sight path, and divide the propagation loss into the path loss PL AIR of the air segment and the path loss PL GRD of the ground segment, and the method is as follows:

[0052]

[0053] Among them, h UAV represents the flight altitude of the UAV, h GRD represents the average building height, and PL n (h) represents the propagation loss of the nth non-line-of-sight path, and h represents the observation height;

[0054] 421) Calculate the path loss of the air segment, and the method is as follows:

[0055]

[0056] Among them, represents the pitch angle of the departure angle of the nth non-line-of-sight path, and is specifically expressed as Among them, e z represents the base vector in the direction of the z-axis of the coordinate axis;

[0057] 422) Calculate the path loss of the ground segment, and the method is as follows:

[0058] 4221) Based on ray tracing and field measurements, obtain the scenario dataset U = {u n}, where u n = {P n , τ n , α n , β n}, and P n , τ n , α n , β n are respectively the power, delay, azimuth angle of arrival, and elevation angle of arrival of the nth path; at the same time, divide the scenario dataset into a training set and a validation set

[0059] 4222) Determine the neural network structure, where the neural network includes an input layer, a hidden layer, and an output layer; the input layer of the neural network inputs the delay, azimuth angle of arrival, and elevation angle of arrival parameters, the hidden layer of the neural network has X neurons, and the output layer of the neural network is expressed as:

[0060]

[0061] where w represents the weight matrix, b represents the bias matrix, σ represents the activation matrix, represents the weight value between the jth neuron in the kth layer and the ith neuron in the previous layer, represents the bias of the jth neuron in the kth layer, the activation function of the jth neuron in the kth layer;

[0062] 4223) Determine the error function of the ground segment path loss training network, and the error function:

[0063] E(τ, α, β; w, b, σ) = (PL GRD (τ, α, β) - P Tr ) 2 (16);

[0064] 4224) Through the gradient descent method, backpropagate to adjust the weight matrix and the bias matrix. When the error function value tends to be stable and reaches the minimum, the network training is stable.

[0065] Further, the specific steps of step 5) include:

[0066] 51) Calculate the shadow fading between the receiver and the transmitter, and the method is as follows:

[0067] Model the probability density function of the shadow fading as a lognormal distribution model related to the flight altitude, specifically as follows:

[0068]

[0069] Among them, p SF (SF) represents the probability density function of shadow fading. ρ(h) and μ(h) represent the standard deviation and mean of SF respectively, and the unit of both is dB. The specific methods for the values of ρ and μ are as follows:

[0070]

[0071]

[0072] Among them, the calculations of the scenario parameters g, l, and i are closely related to the communication scenario type. g ρ , l ρ and i ρ take the values of -88.95, 0.0927, and -94.2 respectively; g μ , l μ and i μ take the values of -94.2, -3.44, and 0.0318 respectively;

[0073] The shadow fading is obtained through the non-linear transformation of a random variable χ(h) ~ N(μ, ρ 2 ), as follows:

[0074] SF(h) = exp(ρχ(h) + μ) (20)

[0075] 52) Calculate the body shadow coefficients of the line-of-sight path and the non-line-of-sight path, and the method is as follows:

[0076] 521) Calculate the body shadow coefficient of the line-of-sight path as follows:

[0077] 5211) Define the body shadow radius R S , and calculate the body shadow range and the wing shadow range When the UAV has no fixed wing part, that is, the model is a rotor, only the body shadow range needs to be calculated, and the method is as follows:

[0078]

[0079]

[0080] Among them,

[0081]

[0082]

[0083] Among them, w b , l b and rb respectively represent the width, length and cross-sectional radius of the fuselage, w w represents the wing width, X L and X B respectively represent the leading-edge sweep angle and the trailing-edge sweep angle, and υ represents the dihedral angle;

[0084] 5212) Calculate the projection vector e of the line-of-sight path on the fuselage plane LoS , as follows:

[0085]

[0086] Among them, and are respectively the components of e LoS on the x and y axes, and the specific calculation method is as follows:

[0087]

[0088] Among them, ω(t), γ(t) and respectively represent the roll angle, pitch angle and yaw angle of the UAV attitude;

[0089] 5213) Based on the fuselage shadow range and the projection vector e LoS further determine whether the fuselage shadow occurs, and the specific method is as follows:

[0090] If e LoS belongs to range, calculate the line-of-sight path fuselage shadow coefficient C L o S (t), as follows:

[0091]

[0092] If e LoS belongs to range, calculate the line-of-sight path fuselage shadow coefficient, as follows:

[0093]

[0094] 522) Calculate the fuselage shadow coefficient C n (t) of the non-line-of-sight path, as follows:

[0095]

[0096] Among them,

[0097]

[0098] Among them, C θ (θ(t), ζ θ ) represents C n(t) Components on the three attitude angles, ζ represents the half-power beamwidth of the antenna,

[0099] Advantages of the present invention:

[0100] 1. On the premise of considering factors such as communication scenarios and three-dimensional motion of the transceiver, the present invention improves the path loss model affected by flight altitude, and introduces the body shadow coefficients of the line-of-sight path and non-line-of-sight path respectively to describe the influence brought by the UAV body structure and body attitude. The method of the present invention can support the software numerical reproduction and hardware real-time simulation of the UAV air-ground channel.

[0101] 2. The present invention uses geometric geographic information and topological relationships to generate channel parameters such as continuously evolving distance and position vectors, and combines neural networks to train and predict channel parameters such as path loss, which can achieve efficient and accurate simulation reproduction of the air-ground channel, thereby supporting the evaluation and optimization of the UAV communication system. Description of the drawings

[0102] Figure 1 It is a flowchart of the method of the present invention.

[0103] Figure 2 It is a schematic diagram of the air-ground communication scenario integrating UAV characteristics in the present invention.

[0104] Figure 3 It is a schematic diagram of the body structure data in the present invention.

[0105] Figure 4 It is a schematic diagram of the path propagation loss of the signal generated by the present invention.

[0106] Figure 5 It is a schematic diagram of the channel received power generated by the present invention. Detailed implementation manners

[0107] For the convenience of understanding by those skilled in the art, the present invention will be further described below in conjunction with the embodiments and the drawings. The content mentioned in the implementation manners does not limit the present invention.

[0108] The simulation duration of this example is 10 s, and the channel state update interval Δt = 0.05 s. The user input parameters are shown in Table 1, including communication scenarios, UAV body structure data sets, initial positions, speeds, three-dimensional motion trajectories, three-dimensional attitudes, pitch angles and azimuth angles of the motion directions, initial positions, speeds, three-dimensional motion trajectories, pitch angles and azimuth angles of the motion directions of the ground receiving end, carrier frequencies, and parameter information of the half-power beamwidth. The trajectories and attitude changes of the UAV transmitting end and the ground receiving end are shown in Table 2, and the scenario data sets are shown in Table 3.

[0109] Refer to Figure 1As shown in the figure, a method for simulating an air-ground channel integrating the characteristics of an unmanned aerial vehicle (UAV) of the present invention comprises the following steps:

[0110] 1) Considering the UAV body structure, flight altitude, and flight attitude, establish a UAV air-ground channel model. The schematic diagram of the communication scenario is as Figure 2 shown;

[0111] Among them, the UAV air-ground channel model is as follows:

[0112]

[0113] Among them, H represents channel fading, PL represents path propagation loss, SF represents shadow fading, and h(t, τ) represents the channel impulse response between the transmitter and receiver, and the specific expressions are as follows:

[0114]

[0115] Among them, K(t) is the Rice factor, C LoS (t) and C n (t) are respectively the body shadow coefficients of the line-of-sight path and the nth non-line-of-sight path, P LoS (t) and P n (t) are respectively the power coefficients of the line-of-sight path and the nth non-line-of-sight path, Φ LoS (t) and Φ n (t) are respectively the phases of the line-of-sight path and the nth non-line-of-sight path, τ LoS (t) and τ n (t) are respectively the time delays of the line-of-sight path and the nth non-line-of-sight path, N is the number of non-line-of-sight paths, and its value is 5, and j is the imaginary unit.

[0116] In particular, the influence of the body structure on the channel is mainly reflected in the calculation process of C LoS (t), the influence of the flight attitude on the channel is mainly reflected in the calculation process of C n (t), and the influence of the flight altitude on the channel is mainly reflected in the calculation processes of PL and SF.

[0117] 2) Input the UAV air-ground communication scenario type, carrier frequency, position vectors and velocity vectors of the UAV / receiver / scatterer at the initial moment, flight attitude, antenna half-power beam width, and body structure data set;

[0118] 3) Use the communication scenario type and carrier frequency to calculate the Rice factor; use the initial position vector and velocity vector to calculate the time delays and phases of the line-of-sight path and the nth non-line-of-sight path; further calculate the power coefficients of the line-of-sight path and the nth non-line-of-sight path;

[0119] Among them, the method for calculating the Rice factor in the UAV air-to-ground communication channel is as follows:

[0120] K(t) = ρ K χ K (t) + κ K log 10 f c + μ K (3)

[0121] Among them, the scenario parameters μ K , ρ K , κ K and χ K (t) are closely related to the type of communication scenario; μ K and ρ K represent the mean and variance of the Rice factor respectively, with values of 4.22 and 1.1, χ K (t) represents a random variable subject to a Gaussian distribution with a mean of 0 and a variance of 3, κ K represents the frequency-dependent factor, with a value of -11.7, f c represents the carrier frequency, with a value of 2.4 GHz.

[0122] The method for calculating the delay and phase of the line-of-sight path and the nth non-line-of-sight path is as follows:

[0123] 31) The method for calculating the position vectors of the receiving end, the transmitting end, and the scatterer at time t (the UAV is used as the signal transmitting end, and the ground vehicle or base station is used as the signal receiving end) is as follows:

[0124]

[0125] Among them,

[0126]

[0127] Among them, v Tx / Rx / S represents the modulus, φ Tx / Rx / S and ψ Tx / Rx / S represent the azimuth and elevation angles of the moving speed respectively, and the values are shown in Table 1;

[0128] Table 1

[0129] Variable Scenario Δt <![CDATA[f c > T <![CDATA[w b > <![CDATA[r b > <![CDATA[w w > υ Value Campus 0.1s 2.4GHz 10s 4m 6m 20m 7° Variable <![CDATA[X L > <![CDATA[X B > <![CDATA[L Tx (t 0 )]]> <![CDATA[L Rx (t 0 )]]> <![CDATA[φ Tx > <![CDATA[ψ Tx > <![CDATA[φ Rx > <![CDATA[ψ Rx <!-- 7 -->]]> Value 30° 18° [0,0,150] [0,0,0] 90° 0° -90° 0° Variable <![CDATA[h GRD ||]]> <![CDATA[v Tx (t)||]]> <![CDATA[||v Rx (t)||]]> <![CDATA[||v S (t)||]]> T θ <![CDATA[φ S > <![CDATA[ψ S > Value 50 m 20m / s 20m / s 1m / s 10s 180° 90° 0°

[0130] 32) The method for calculating the distance vectors of the line-of-sight path and the nth non-line-of-sight path at time t is as follows:

[0131] 321) Using the position vectors of the receiving end and the transmitting end obtained in step 31), the method for calculating the distance vector of the line-of-sight path at time t is as follows:

[0132]

[0133] Among them, · represents the Euclidean distance of a vector, e p and e q respectively represent the vectors pointing from the positions of the receiving end and the transmitting end at the initial moment to the positions of the receiving end and the transmitting end at time t. d LoS (t 0 ) represents the propagation distance of the line-of-sight path at the initial moment, and the calculation method is as follows:

[0134] d LoS (t 0 ) = L Tx (t 0 ) - L Rx (t 0 ) (6)

[0135] 322) Use the position vectors of the receiving end and the transmitting end obtained in step 31) to calculate the distance vector of the nth non-line-of-sight path at time t, and the method is as follows:

[0136] ||d n (t)|| = ||a n (t) - b n (t)|| (7)

[0137] Among them, a n (t) represents the vector from the UAV transmitting end to the scatterer, and specifically ||a n (t)|| = ||L Tx (t) - L S (t)||, b n (t) represents the vector from the ground receiving end to the scatterer, and specifically ||b n (t)|| = ||L Rx (t) - L S (t)||;

[0138] 33) According to the distance vector of the line-of-sight path and the distance vector of the nth non-line-of-sight path in step 32), calculate the propagation delay and phase of the line-of-sight path and the nth non-line-of-sight path at time t, and the method is as follows:

[0139]

[0140]

[0141] Among them, c represents the speed of light, and λ represents the signal wavelength.

[0142] Calculate the power coefficients of the line-of-sight path and the nth non-line-of-sight path, and the method is as follows:

[0143] By normalizing all power coefficients, the final value of the line-of-sight path power coefficient is set to 1. Meanwhile, the power calculation method for the nth non-line-of-sight path is as follows:

[0144]

[0145] Wherein, represents the power of the nth non-line-of-sight path before normalization, and the calculation method is as follows:

[0146]

[0147] Wherein, r τ represents the time-delay scaling coefficient, with a value of 2.5; G n follows a Gaussian distribution with a mean of 0 and a variance of 3, and σ τ represents the random time-delay spread, with a value of 2.34×10 -7 .

[0148] 4) Divide the path propagation loss between the receiver and the transmitter into an air segment and a ground segment; calculate the path loss of the air segment based on the flight altitude of the UAV, and use neural network training to obtain the path loss of the ground segment;

[0149] 41) Calculate the propagation loss PL LoS of the line-of-sight path, and the method is as follows:

[0150] PL LoS = 20log 10 ||d LoS || + 20log 10 f c + c 1 (12)

[0151] Wherein, c 1 represents the scenario-related correction coefficient, with a value of 32.4;

[0152] 42) Calculate the propagation loss of the non-line-of-sight path, and divide the propagation loss into the path loss PL AIR of the air segment and the path loss PL GRD of the ground segment, and the method is as follows:

[0153]

[0154] Wherein, h UAV represents the flight altitude of the UAV, h GRD represents the average building height, and PL n (h) represents the propagation loss of the nth non-line-of-sight path, and h represents the observation height;

[0155] 421) Calculate the path loss of the air segment, and the method is as follows:

[0156]

[0157] Among them, represents the elevation angle of the departure angle of the nth non-line-of-sight path, specifically expressed as where e z represents the base vector in the direction of the z-axis of the coordinate axis;

[0158] 422) Calculate the path loss of the ground segment as follows:

[0159] 4221) Based on ray tracing and actual measurements, obtain the scene dataset U = {u n} as shown in Table 3. Among them, u n = {P n , τ n , α n , β n}, where P n , τ n , α n , β n are respectively the power, delay, azimuth angle of arrival angle, and elevation angle of arrival angle of the nth path; at the same time, divide the scene dataset into a training set and a validation set

[0160] Table 3

[0161]

[0162] 4222) Determine the neural network structure. The neural network includes an input layer, a hidden layer, and an output layer; the input layer of the neural network inputs the delay, azimuth angle of arrival angle, and elevation angle of arrival angle parameters. The hidden layer of the neural network has X neurons, and the output layer of the neural network is expressed as:

[0163]

[0164] where w represents the weight matrix, b represents the bias matrix, σ represents the activation matrix, represents the weight value between the jth neuron in the kth layer and the ith neuron in the previous layer, represents the bias of the jth neuron in the kth layer, the activation function of the jth neuron in the kth layer;

[0165] 4223) Determine the error function of the ground segment path loss training network. The error function:

[0166] E(τ, α, β; w, b, σ) = (PL GRD (τ, α, β) - P Tr ) 2(16);

[0167] 4224) By using the gradient descent method, the weight matrix and bias matrix are adjusted through backpropagation. When the error function value tends to be stable and reaches the minimum, the network training is stable.

[0168] 5) Using the flight altitude and communication scenario type, calculate the shadow fading between the receiving end and the transmitting end. Using the fuselage structure, flight attitude, and half-power beam width of the antenna, calculate the body shadow coefficients of the line-of-sight path and non-line-of-sight path;

[0169] 51) Calculate the shadow fading between the receiving end and the transmitting end as follows:

[0170] Model the probability density function of shadow fading as a log-normal distribution model related to flight altitude, specifically as follows:

[0171]

[0172] where p SF (SF) represents the probability density function of shadow fading, ρ(h) and μ(h) represent the standard deviation and mean of SF respectively, and the units are both dB; the specific methods for the values of ρ and μ are as follows:

[0173]

[0174]

[0175] where the calculations of the scenario parameters g, l, and i are closely related to the communication scenario type, g ρ , l ρ and i ρ take the values of -88.95, 0.0927, -94.2 respectively; g μ , l μ and i μ take the values of -94.2, -3.44, 0.0318 respectively;

[0176] The shadow fading is obtained through the non-linear transformation of a random variable χ(h)~N(μ,ρ 2 ) that follows a Gaussian distribution, as follows:

[0177] SF(h) = exp(ρχ(h)+μ) (20)

[0178] 52) Calculate the body shadow coefficients of the line-of-sight path and non-line-of-sight path as follows:

[0179] 521) Calculate the body shadow coefficient of the line-of-sight path as follows:

[0180] 5211) Taking a fixed-wing UAV as an example, the schematic diagram of the fuselage structure data is as Figure 3As shown; define the radius R of the fuselage shadow S , calculate the fuselage shadow range based on the fuselage structure dataset and the wing shadow range The method is as follows:

[0181]

[0182]

[0183] Among them,

[0184]

[0185]

[0186] Among them, w b , l b and r b respectively represent the width, length and cross-sectional radius of the fuselage, w w represents the wing width, X L and X B respectively represent the leading-edge sweep angle and the trailing-edge sweep angle, υ represents the dihedral angle, and the values of the fuselage structure dataset are shown in Table 1;

[0187] 5212) Calculate the projection vector e of the line-of-sight path on the fuselage plane LoS , as follows:

[0188]

[0189] Among them, and are the components of e LoS on the x and y axes respectively, and the specific calculation method is as follows:

[0190]

[0191] Among them, ω(t), γ(t) and respectively represent the roll angle, pitch angle and yaw angle of the UAV attitude, and the values are shown in Table 2;

[0192] Table 2

[0193]

[0194] 5213) Further determine whether the fuselage shadow occurs based on the fuselage shadow range and the projection vector e LoS , the specific method is as follows:

[0195] If e LoS belongs to the range, calculate the line-of-sight path fuselage shadow coefficient C L oS (t) is as follows:

[0196]

[0197] If e LoS belongs to the range, calculate the line-of-sight path body shadow coefficient as follows:

[0198]

[0199] 522) Calculate the body shadow coefficient C n (t) of the non-line-of-sight path as follows:

[0200]

[0201] Among them,

[0202]

[0203] Among them, C θ (θ(t), ζ θ ) represents the components of C n (t) on the three attitude angles, ζ represents the half-power beam width of the antenna, and the value is 180°,

[0204] 6) Input the Rice factor, time delay, phase, power coefficient, path loss, and shadow fading calculated above into the UAV air-ground channel model to output the channel fading.

[0205] The effects obtained in this example can be further illustrated by the specific data obtained in the simulation experiment. The effects of this example are shown in two parts. The influence of flight altitude on the channel power is illustrated by Figure 4 , and the influence of the fuselage structure and flight attitude on the channel power is illustrated by Figure 5 . It can be observed from Figure 4 , Figure 5 that:

[0206] (1) Figure 4 gives the path loss generated by the present invention. It can be seen that the path loss generated by the model proposed by the present invention is divided into two obvious parts with the change of the UAV flight altitude, while the unsegmented model cannot reflect this characteristic. In addition, the path loss in the same scenario is obtained by using the ray tracing method, which has good consistency with the model proposed by the present invention.

[0207] (2) Figure 5The channel normalized received power generated by the present invention is given. It can be seen that there are two large fluctuations in the channel received power generated by the proposed model of the present invention. This is because the change in flight attitude causes the non-line-of-sight path to have a fuselage shadow phenomenon. In addition, when t = 2.2 s and t = 7.2 s, the received power drops to -55 dB. At this time, the fuselage shadow coefficient of the line-of-sight path is 0, and the channel fading coefficient without considering the fuselage shadow cannot reflect this characteristic.

[0208] There are many specific application ways of the present invention. The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art of this technology, several improvements can be made without departing from the principle of the present invention, and these improvements should also be regarded as the protection scope of the present invention.

Claims

1. An air - ground channel modeling and simulation method integrating UAV characteristics, characterized in that, the steps are as follows: 1) Considering the UAV fuselage structure, flight altitude and flight attitude, establish an air - ground channel model for the UAV; 2) Input the UAV air - ground communication scenario type, carrier frequency, initial position vectors and velocity vectors of the UAV / receiver / scatterer, flight attitude, antenna half - power beam width, and fuselage structure data set; 3) Using the communication scenario type and carrier frequency, calculate the Rice factor; using the initial position vectors and velocity vectors, calculate the delay and phase of the line - of - sight path and the nth non - line - of - sight path; further calculate the power coefficients of the line - of - sight path and the nth non - line - of - sight path; 4) Divide the path propagation loss between the receiver and the transmitter into an air segment and a ground segment; calculate the path loss of the air segment according to the flight altitude of the UAV, and use neural network training to obtain the path loss of the ground segment; 5) Using the flight altitude and communication scenario type, calculate the shadow fading between the receiver and the transmitter, and use the fuselage structure, flight attitude and antenna half - power beam width to calculate the body shadow coefficients of the line - of - sight path and the non - line - of - sight path; 6) Input the Rice factor, delay, phase, power coefficients, path loss and shadow fading calculated above into the UAV air - ground channel model, and output the channel fading.

2. The air - ground channel modeling and simulation method integrating UAV characteristics according to claim 1, characterized in that, the UAV air - ground channel model in step 1) is as follows: where, H represents the channel fading, PL represents the path propagation loss, SF represents the shadow fading, and h(t,τ) represents the channel impulse response between the transceiver, and the specific expression is as follows: where K(t) is the Rice factor, C LoS (t) and C n (t) are the body shadow coefficients of the LOS path and the nth NLOS path respectively, P LoS (t) and P n (t) are the power coefficients of the LOS path and the nth NLOS path respectively, Φ LoS (t) and Φ n (t) are the phases of the LOS path and the nth NLOS path respectively, τ LoS (t) and τ n (t) are the time delays of the LOS path and the nth NLOS path respectively, N is the number of NLOS paths, and j is the imaginary unit.

3. The air - ground channel modeling and simulation method integrating UAV characteristics according to claim 2, characterized in that, the method for calculating the Rice factor in the UAV air - ground communication channel in step 3) is as follows: K(t) = ρ K χ K (t) + κ K log 10 f c + μ K (3) Among them, the scenario parameter μ K , ρ K , κ K and χ K (t) are closely related to the type of communication scenario; μ K and ρ K respectively represent the mean and variance of the Rice factor, χ K (t) represents a random variable that follows a Gaussian distribution with a mean of 0 and a variance of 3, κ K represents the frequency-dependent factor, f c represents the carrier frequency.

4. The air - ground channel modeling and simulation method integrating UAV characteristics according to claim 3, characterized in that, the method for calculating the delay and phase of the line - of - sight path and the nth non - line - of - sight path in step 3) is as follows: 31) Calculate the position vectors of the receiver, the transmitter and the scatterer at time t, and the method is as follows: where, where ||v Tx / Rx / S || represents the modulus value, φ Tx / Rx / S and ψ Tx / Rx / S represent the azimuth angle and elevation angle of the moving speed respectively; 32) Calculate the distance vectors of the line - of - sight path and the nth non - line - of - sight path at time t, and the method is as follows: 321) Using the position vectors of the receiver and the transmitter obtained in step 31), calculate the distance vector of the line - of - sight path at time t, and the method is as follows: ||d LoS (t)|| = ||d LoS (t 0 ) - e p (t) + e q (t)|| (5) where, ||·|| represents the Euclidean distance of a vector, and e p (t) and e q (t) respectively represent the vectors pointing from the positions of the receiver and transmitter at the initial moment to the positions of the receiver and transmitter at time t, and d LoS (t 0 ) represents the propagation distance of the line-of-sight path at the initial moment, and the calculation method is as follows: ||d LoS (t 0 )|| = ||L Tx (t 0 ) - L Rx (t 0 )|| (6); 322) Using the position vectors of the receiver and the transmitter obtained in step 31), calculate the distance vector of the nth non - line - of - sight path at time t, and the method is as follows: ||d n (t) || = || a n (t) - b n (t) || (7) where a n (t) represents the vector from the UAV transmitting end to the scatterer, specifically expressed as ||a n (t)|| = ||L Tx (t) - L S (t)||, and b n (t) represents the vector from the ground receiving end to the scatterer, specifically expressed as ||b n (t)|| = ||L Rx (t) - L S (t)||; 33) According to the distance vectors of the line - of - sight path and the nth non - line - of - sight path in step 32), calculate the propagation delay and phase of the line - of - sight path and the nth non - line - of - sight path at time t, and the method is as follows: Φ LoS (t) = 2π(d LoS (t) mod λ) (9) where, c represents the speed of light and λ represents the signal wavelength.

5. The air - ground channel modeling and simulation method integrating UAV characteristics according to claim 4, characterized in that, the method for calculating the power coefficients of the line - of - sight path and the nth non - line - of - sight path in step 3) is as follows: By normalizing all power coefficients, the final value of the line-of-sight path power coefficient is set to 1. Meanwhile, the power calculation method for the nth non-line-of-sight path is as follows: Among them, represents the power of the nth non-line-of-sight path before normalization, and the calculation method is as follows: where r τ represents the time delay scaling factor, G n follows a Gaussian distribution with a mean of 0 and a variance of 3, and σ τ represents the random time delay spread.

6. The method for air-to-ground channel modeling and simulation that integrates UAV characteristics according to claim 5, characterized in that, the specific steps of step 4) include: 41) Calculate the propagation loss PL of the line-of-sight path LoS , the method is as follows: PL LoS = 20 log 10 ||d LoS (t)|| + 20 log 10 f c + c 1 (12) Among them, c 1 represents a scene-related correction coefficient; 42) Calculate the propagation loss of the non-line-of-sight path, and divide the propagation loss into the aerial segment path loss PL AIR and the ground segment path loss PL GRD , and the method is as follows: Among them, h UAV represents the flight altitude of the UAV, and h GRD represents the average height of the building. PL n (h) represents the propagation loss of the nth non-line-of-sight path, and h represents the observation height; 421) Calculate the path loss in the air segment, and the method is as follows: Among them, represents the elevation angle of the departure angle of the nth non-line-of-sight path, specifically expressed as where e z represents the basis vector in the direction of the z-axis of the coordinate axis; 422) Calculate the path loss in the ground segment, and the method is as follows: 4221) Based on ray tracing and field measurements, obtain the scene dataset U = {u n}, where u n = {P n , τ n , α n , β n}, and P n , τ n , α n , β n are respectively the power, delay, azimuth angle of arrival, and elevation angle of arrival of the nth path; at the same time, divide the scene dataset into a training set and a validation set 4222) Determine the neural network structure, and the neural network includes an input layer, a hidden layer, and an output layer; 4223) Determine the error function of the training network for the ground segment path loss, and the error function: E(τ,α,β;w,b,σ)=(PL GRD (τ,α,β;w,b,σ)-P n Tr ) 2 (16); 4224) Through the gradient descent method, adjust the weight matrix and bias matrix by backpropagation. When the error function value tends to be stable and reaches the minimum, the network training is stable.

Images