CN119094001A

CN119094001A - Trajectory optimization method for UAV relay in laser/RF hybrid aviation communication network

Info

Publication number: CN119094001A
Application number: CN202411288207.3A
Authority: CN
Inventors: 张曦文; 王渊; 王翔; 黄仰超; 韩海艳
Original assignee: Air Force Engineering University of PLA
Current assignee: Air Force Engineering University of PLA
Priority date: 2024-09-14
Filing date: 2024-09-14
Publication date: 2024-12-06

Abstract

The unmanned aerial vehicle relay track optimization method in the laser/radio frequency hybrid aviation communication network comprises the steps of system model construction, system transmission link construction, cloud model construction, unmanned aerial vehicle energy consumption model construction and unmanned aerial vehicle track optimization. The method comprises the steps of taking a fixed-wing unmanned aerial vehicle as a relay for linking a plane-mounted platform and a ground station laser/radio frequency hybrid communication link, comprehensively considering cloud cover, weather change and a data transmission rate threshold faced by an aviation laser/radio frequency communication network, modeling cloud obstacles to obtain an unmanned aerial vehicle deployment feasible region for guaranteeing line-of-sight communication, deriving a propulsion energy consumption formula of the fixed-wing unmanned aerial vehicle with variable height, providing a high-efficiency track planning algorithm to obtain tracks of maximum data accessibility, minimum energy consumption and optimal energy efficiency, and analyzing influences of weather change and cloud movement on the energy efficiency. The invention effectively improves the energy efficiency of the unmanned aerial vehicle auxiliary laser/radio frequency hybrid network and improves the usability of airborne laser communication.

Description

Unmanned aerial vehicle relay track optimization method in laser/radio frequency hybrid aviation communication network

Technical Field

The invention belongs to the technical field of airborne optical communication and communication energy efficiency track optimization, and particularly relates to an unmanned aerial vehicle relay track optimization method in a laser/radio frequency hybrid aviation communication network.

Background

With the development of information technology, the battlefield electromagnetic environment is increasingly complex. The continuous rise in battlefield intelligence, monitoring and scout task performance has placed high capacity, high reliability, high rate demands on aviation communication networks. At present, radio frequency links are mostly adopted for information transmission among the airborne platforms, and the obvious defects of single communication means, lower transmission rate, limited frequency band resources, easy interference and easy interception and the like exist. The airborne optical communication has the advantages of high speed, large bandwidth, strong interference resistance, good confidentiality, miniaturized terminal and the like. However, since the airborne platform is in a complex atmospheric environment, weather changes have a great influence on airborne optical communication. The advantages of the laser/radio frequency hybrid link and the advantages of the laser/radio frequency hybrid link are effectively combined, and the effectiveness and the reliability of the aviation communication link are greatly improved.

In an aeronautical communication network, the unmanned aerial vehicle is generally used as a mobile relay to provide a reliable wireless connection between two or more remote users to enhance the scalability of the aeronautical communication system. However, the endurance of the unmanned aerial vehicle is limited by the load, and the energy-saving design is a key of the unmanned aerial vehicle system. Unlike ground communication systems, unmanned aerial vehicles have additional flight energy consumption in addition to communication energy consumption, and flight energy consumption is typically much higher than communication energy consumption. The flight energy consumption of the unmanned aerial vehicle is influenced by the flight state of the unmanned aerial vehicle, so that the flight track of the unmanned aerial vehicle with reasonable design is required to effectively save the flight energy.

Disclosure of Invention

Aiming at the defects of the prior art, the invention provides an unmanned aerial vehicle relay track optimization method in a laser/radio frequency hybrid aviation communication network, which is used for an airborne laser/radio frequency communication system based on fixed wing unmanned aerial vehicle relay assistance, wherein an early warning machine-to-unmanned aerial vehicle relay and an unmanned aerial vehicle relay-to-ground station are both parallel laser/radio frequency hybrid transmission links, thick clouds with high liquid water content values are regarded as obstacles, and in order to keep a laser and radio frequency line-of-sight communication link, the unmanned aerial vehicle relay is deployed at a position without cloud coverage so as to bypass the shielding of the obstacles and keep the high-speed connection between the early warning machine and the ground station, and the method specifically comprises the following steps:

step one, constructing a system model;

Establishing a three-dimensional rectangular coordinate system, taking the gravity center of a ground station as an origin, taking the forward direction as an X-axis forward direction, taking the forward direction as a Y-axis forward direction, and taking the upward direction vertical to an X-Y plane as a Z-axis forward direction, assuming that three-dimensional position vectors of the ground station and an early warning machine are q _GS＝[0,0,0]^T and q _AWACS＝[0,0,H_AWACS]^T,H_AWACS respectively as Z-axis coordinates of the early warning machine, taking the system as a discrete time model, dividing the movement time of the unmanned aerial vehicle into N time slots with equal intervals, and setting the time slot intervals as delta _t, wherein the position coordinates of the unmanned aerial vehicle in the nth time slot are q [ N ] = [ X [ N ], Y [ N ], h [ N ] ] ^T, n=0, 1, N, wherein X [ N ], Y [ N ] respectively represent the projection coordinates of the unmanned aerial vehicle on the X axis, Y axis and the Z axis;

the distance d [ n ] between the drone and the ground station is expressed as:

The trajectory of the unmanned aerial vehicle is represented by a position vector q [ n ], a speed vector v [ n ] and an acceleration vector a [ n ], and the unmanned aerial vehicle is provided with the following discrete state space model:

v[n+1]=v[n]+a[n]δ_t,n=0,...N (3)

The initial position q _I and the final position q _F, and the initial velocity v _I and the final velocity v _F of the unmanned plane are determined in advance, and n=0 and n=n represent the initial time slot and the final time slot, respectively, and then there are:

q[0]=q_I,q[N]＝q_F (4)

v[0]=v_I,v[N]＝v_F (5)

mathematically considering the performance limitations of a drone:

V_min≤||v[n]||≤V_max,||a[n]||≤A_max (6)

Wherein V _max and V _min are respectively the maximum speed and the minimum speed of the unmanned aerial vehicle, a _max represents the maximum acceleration of the fixed-wing unmanned aerial vehicle, |·| represents the norm;

Step two, a system transmission link;

1. Laser link

The fading experienced in the laser link transmission between the drone and the ground station consists of atmospheric attenuation, turbulence and pointing error, the attenuation coefficient in the nth time slot of which is expressed as:

h_FSO[n]＝h_a[n]h_t[n]h_p[n] (7)

Wherein h _a[n]、h_t [ n ] and h _p [ n ] represent atmospheric attenuation, turbulence fading and pointing error, respectively;

the scattering and absorption of atmospheric particles results in an atmospheric attenuation of the beam during propagation, which attenuation is related to the distance d [ n ] between the drone and the ground station, expressed as:

h_a[n]=exp(-ωd[n]) (8)

wherein ω is the attenuation coefficient of the laser link;

atmospheric turbulence causes power loss and random phase fluctuations, and is simulated by Gamma-Gamma distribution, expressed as:

Wherein, alpha and beta both represent fading parameters, h _t n is the mean value thereof To approximate the representation, there is

The pointing error loss is expressed as:

Where u is the radial distance from the beam center to the lens center, A ₀ is the fraction of the optical power acquired when the difference between the spot center and the detector center is zero, the expression is A ₀＝erf(v₁)erf(v₂), erf (. Cndot.) is the error function, Is the lower bound of the ratio of the aperture radius a of the receiver to the beam width omega _d at the distance dn,An upper bound for the ratio of a to omega _d,For the angle between the laser beam and the z axis in the direction of the unmanned plane and the ground station, θn is the angle between the projection of the laser beam on the x-y plane and the x axis, and the n-th moment pointing error coefficient k _g is expressed as:

In the formula, And θ are respectively the nth timeA value corresponding to θn;

omega _d is calculated by the following formula:

Where ω ₀ is the beam waist radius, λ is the wavelength of light, Referred to as the coherence length,Is the refractive index structural parameter of the laser beam;

the discrete time data reachability model of the laser link at the ground station is expressed as:

Where B _FSO is the laser link bandwidth, P _FSO is the peak power allowed by the ground station, For noise variance, the reachability coefficient k is expressed as:

wherein e represents a natural number, and the free parameter μ ^* is a formula Is a unique solution to (a);

2. Radio frequency link

The attenuation coefficient in the nth time slot of the radio frequency transmission link between the unmanned plane and the ground station is expressed as:

Wherein τ _RF [ n ] represents the influence of large-scale fading such as path loss and shadow fading, Representing the effects of small scale fading and its mean

Considering two different large scale attenuation models for line-of-sight and non-line-of-sight transmissions, τ _RF [ n ] is expressed as:

Where β ₀ is the received power at reference distance d ₀ =1, α _RF is the path loss index, and ε is the additional attenuation factor due to the non-line-of-sight link;

the desired representation of the radio frequency link attenuation coefficient is:

In the formula, Representing the actual probability of line-of-sight propagation between the drone and the ground station taking into account non-line-of-sight link attenuation, P _LoS [ n ] = 1/[1+C ·exp (-D [ n ] -C ]) ] representing the ideal line-of-sight propagation probability between the drone and the ground station, C and D being parameters dependent on propagation conditions, γ [ n ] being the angle between the line-of-sight link and the x-y plane;

The discrete time data reachability model for the radio frequency link at the ground station is expressed as:

Wherein B _RF is the radio frequency bandwidth, P _RF is the radio frequency transmitting power, and sigma ² _RF is the radio frequency noise variance;

the total achievable rate of the ground station is expressed as the sum of the laser link reachability and the radio link reachability:

R_T[n]＝R_FSO[n]+R_RF[n] (19)

thirdly, constructing a cloud model;

Constructing a thick cloud as a cylindrical obstacle, taking a coordinate q _cloud[n]＝[x_cloud[n],y_cloud[n],h_cloud]^T as a cloud center, wherein x _cloud[n],y_cloud[n],h_cloud respectively represents projection coordinates of the cloud center on an x axis, a y axis and a z axis, R _cloud and V _cloud respectively represent the radius of the bottom surface of the cylinder and the height of the cylinder, phi _cloud is the azimuth angle of the cloud, and is defined as the included angle between the connection point of any point on the circumference of the upper bottom surface and the circumference of the bottom surface of the cylinder and the projection point of the cloud center on the upper surface and the lower surface of the cylinder and the x axis;

the downlink from the early warning machine to the ground station is shielded by the edge of the upper bottom surface of the cylinder, and the uplink from the ground station to the early warning machine is shielded by the edge of the lower bottom surface of the cylinder;

And analyzing a feasible area deployed by the three-dimensional unmanned aerial vehicle by adopting a two-dimensional vertical section view, wherein an included angle between a connecting line of the ground station and the unmanned aerial vehicle and an x-y plane is a ground station elevation angle, which is expressed as:

The angle between the connection from the lower edge of the cylinder to the ground station and the x-y plane is the maximum elevation angle that keeps the line-of-sight link clear, expressed as:

the included angle between the connection of the early warning machine and the unmanned aerial vehicle and the x-y plane is the depression angle of the early warning machine, and is expressed as:

The included angle between the connection of the upper edge of the cylinder and the early warning machine and the x-y plane is the maximum depression angle of the early warning machine, and is expressed as:

feasible region for relay deployment of unmanned aerial vehicle Expressed as:

fourth, unmanned aerial vehicle energy consumption model construction;

Approximately decomposing the movement of the unmanned aerial vehicle into a tilting horizontal turning movement along an x-y plane and a flying height changing movement along a z axis, wherein v _xy and v _z are respectively speed vectors of the unmanned aerial vehicle on the x-y plane and the z plane, the propulsion energy consumption of the fixed wing unmanned aerial vehicle is related to the stress of the unmanned aerial vehicle in the movement process, the unmanned aerial vehicle receives four forces in the flying process, namely gravity W, resistance D, lift force L and thrust F, the gravity W=mg, m is the mass of the unmanned aerial vehicle, and g=9.8 m/s ² is gravity acceleration, and the following steps are obtained:

Lcosδ-W=ma_z (25)

Lsinδ=ma_⊥ (26)

F-D=ma_|| (27)

Where δ represents the unmanned aerial vehicle tilt angle, a _⊥ and a _|| represent acceleration components perpendicular and parallel to v _xy, respectively, a _z is the acceleration component in the direction of v _z, and dividing equation (26) and equation (25) to obtain:

Resistance is expressed as:

Wherein, c ₁ and c ₂ are two constant parameters related to the mass and wing area of the unmanned aerial vehicle, V represents the volume of the unmanned aerial vehicle, and the load coefficient k ^* is calculated by the formula (26), the formula (28) and the gravity expression w=mg, wherein m represents the mass of the unmanned aerial vehicle, and g represents the gravity acceleration:

If a _z =0, the load factor In this case, the unmanned aerial vehicle is tilted and steered horizontally to fly, conversely, if a _z is not equal to 0, the change in flying height will cause the change in the load factor, and the expression of the thrust is given by the formulas (27), (29), (30):

the instantaneous power required by the drone is expressed as:

Wherein a _⊥ and a _|| are represented by A _||＝a_xy ^Tv_xy/||v_xy |, wherein a _xy is the acceleration component of the unmanned aerial vehicle along v _xy, and therefore, a fixed wing unmanned aerial vehicle with a discrete-time three-dimensional trajectory q [ n ] under normal operating conditions, the total propulsion energy model is expressed as:

Wherein v _xy [ n ] and a _xy [ n ] are the unmanned speed and acceleration components, respectively, of the nth time slot along the x-y plane, and a _z [ n ] is the acceleration component along the z-axis;

assuming a _xy ^T[n]v_xy [ n ] =0, the energy model is simplified as:

If it is assumed that the initial velocity v [0] and the final velocity v [ N ] are the same, the last term in equation (34) is zero;

step five, optimizing the track of the unmanned aerial vehicle;

Step1, maximizing energy efficiency under constraint conditions;

The energy efficiency is defined as the ratio of the total data availability to the total energy consumption of the unmanned aerial vehicle in the flight time, and the energy efficiency of the unmanned aerial vehicle communication system is expressed as follows from the formula (19) and the formula (34):

the energy efficiency maximization problem of the constrained trajectory is expressed as:

constraint (1) - (6), (36 b)

0≤h[n]≤H_AWACS(n=1,...,N)(36c)

R_T[n]≥R_th(n=1,...,N)(36d)

Wherein, (36 b) shows that the unmanned plane track is required to meet the constraint of formulas (1) - (6), (36 c) is the practical constraint that the unmanned plane flying height is required to meet, (36 d) is the guarantee of the service quality requirement of the system, namely the sum of the transmission rates of the parallel links is required to be larger than the minimum accessibility threshold R _th, (36 e) is the constraint of the feasible region influenced by the cloud obstacle on the unmanned plane position;

Step2, a track optimization method;

A track planning method based on shared knowledge acquisition concretely comprises the following steps:

step1, initializing the maximum value G of the optimization iteration times, the population number N, the knowledge factor parameter k _f, the knowledge ratio parameter k _r and the greedy scaling factor l;

step2, randomly initializing a population x _i,i＝1,2,…N,x_i as the ith individual in the population;

step3. Calculate the fitness function value f (x _i) for each individual in the population, The concrete steps are as follows:

Wherein eta _E,η_R,η_φ,η_end is the energy efficiency target, transmission constraint, angle constraint and terminal constraint factor respectively, omega _E,ω_R,ω_φ,ω_end is the weight coefficient corresponding to 4 factors respectively, unmanned plane EE _T [ n ] represents the energy efficiency of the unmanned plane communication system at the nth moment, phi _AU,φ_AU,max is the elevation angle of the ground station and the maximum value thereof respectively, phi _GU,φ_GU,max is the depression angle of the early warning machine and the maximum value thereof respectively, q _I,q_F represents the initial position and the final position of the unmanned plane respectively, and v _I,v_F represents the initial speed and the final speed of the unmanned plane respectively;

step4, sequencing individuals in the population according to the sequence from small to large fitness value, wherein x _best,…x_i-1,x_i,x_i+1,…x_worst;x_best represents the individual with the smallest fitness value in the population, and x _worst represents the individual with the largest fitness value in the population;

step5. Primary acquisition and sharing knowledge phase, update each x _i, i=1, 2, according to the probability of k _r, by formula (38); updating x _i with the co-participation of the (i-1) th individual, the (i+1) th individual and the randomly selected individual in the population, thereby obtaining The primary acquisition and shared knowledge phase population update strategy is shown in equation (38):

wherein, For updated individuals, x _r is the randomly selected individual;

step6, sorting the updated population individuals according to the order of the fitness value from small to large, and then dividing the sorted individuals into 3 types, namely the best individual, the medium individual and the worst individual, wherein the best individual accounts for p, the worst individual accounts for p, and the medium individual accounts for 1-2p;

step7. update each x _i, i=1, 2, according to the probability of k _r, by equation (39);

Advanced acquisition and shared knowledge phase population update strategy is shown in equation (39)

Wherein x _p-best,x_m,x_p-worst is a randomly selected individual from the optimal individuals, a randomly selected individual from the worst individuals and a randomly selected individual from the medium individuals, and the calculation method of the selection probability pr_best _{i_1},pr_m_{i_2},pr_worst_{i_3} comprises the following steps:

wherein, N _best,N_m,N_worst is the optimal individual number, the middle individual number and the worst individual number respectively;

Rank_m _{i_2},Rank_m_{i_2},Rank_worst_{i_3} is the order of arrangement of each individual of the optimal individual number, the intermediate individual number, and the worst individual number, respectively, and is expressed as follows:

Rank_best_{i_1}＝l(N_best-i_1)+1 (43)

Rank_m_{i_2}＝l(N_m-i_2)+1 (44)

Rank_worst_{i_3}＝l(N_worst-i_3)+1 (45)

step8, updating the current optimal solution in the population according to the formula (46);

Then Then

Wherein, In order for the individual to be updated before,The fitness function value is used for the fitness function value; Is that Is a fitness function value of (a);

step9. Update the globally optimal solution according to equation (47);

Then Then

Wherein, As an individual who is globally optimal,The fitness function value is used for the fitness function value;

step10. Returning step4 until the number of optimization iterations reaches a maximum G;

step11. Output the globally optimal solution.

In one embodiment of the present invention, in the fifth Step2 and Step1, the initial values of the parameters are set to g=100, n=100, k _f＝0.5;k_r =0.9, and k=3.

In another embodiment of the present invention, ω _E＝10⁵,ω_R＝1,ω_φ＝10³,ω_end＝10³ is taken in steps five Step2 and Step 3.

In yet another embodiment of the present invention, in steps five Step2, step6, the optimal individual ratio p is 20%, the worst individual ratio p is 20%, and the medium individual ratio 1-2p is 80%

The invention utilizes the laser/radio frequency cooperative communication technology to construct the high-speed and high-reliability airborne optical communication link, fully considers constraint conditions such as cloud shielding, weather change, communication speed threshold, unmanned aerial vehicle speed, acceleration range and the like, optimizes the unmanned aerial vehicle relay flight track, and improves the energy efficiency of the system.

Drawings

Fig. 1 is a schematic diagram of an airborne laser/radio frequency communication system based on unmanned aerial vehicle relay assistance;

FIG. 2 is a view of a cloud model structure, wherein FIG. 2 (a) shows a three-dimensional view, and FIG. 2 (b) shows a front view;

FIG. 3 is a force analysis diagram of a fixed wing unmanned aerial vehicle;

fig. 4 is a trajectory of the maximization of the communication rate of the unmanned aerial vehicle, fig. 4 (a) shows a top view, fig. 4 (b) shows a front view;

Fig. 5 is a path for minimizing energy consumption of the unmanned aerial vehicle, fig. 5 (a) shows a top view, and fig. 5 (b) shows a front view;

fig. 6 is an unmanned energy efficiency maximizing trajectory, fig. 6 (a) shows a top view, fig. 6 (c) shows a side view, and fig. 6 (b) shows a front view;

FIG. 7 illustrates energy efficiency trajectories at different cloud centers;

FIG. 8 is an energy efficiency trace at different cloud movement speeds;

FIG. 9 is an analysis of the effect of weather on system performance, FIG. 9 (a) shows the effect of weather on energy efficiency, FIG. 9 (b) shows the effect of weather on communication rate, FIG. 9 (c) shows the effect of weather on energy consumption;

FIG. 10 is a graph of method performance versus performance.

Detailed Description

The invention is further described below with reference to the accompanying drawings:

An airborne laser/radio frequency communication system model based on the relay assistance of the fixed-wing unmanned aerial vehicle in the invention is shown in figure 1. The system comprises an early warning machine, an unmanned aerial vehicle relay and a ground station, wherein the early warning machine and the unmanned aerial vehicle relay are connected to the ground station through parallel laser/radio frequency mixed transmission links, thick clouds with high liquid water content values are regarded as barriers, the unmanned aerial vehicle relay is deployed at a position without cloud coverage for maintaining a video distance communication link of laser and radio frequency, shielding of the barriers is avoided, high-speed connection between the early warning machine and the ground station is maintained, and the unmanned aerial vehicle relay in the system works in a decoding and forwarding mode. The invention comprises the following specific steps.

Step one, constructing a system model;

And establishing a three-dimensional rectangular coordinate system, wherein the gravity center of the ground station is taken as an origin, the forward direction is taken as an X-axis forward direction, the forward direction is taken as a Y-axis forward direction, and the upward direction perpendicular to the X-Y plane is taken as a Z-axis forward direction. Assume that the three-dimensional position vectors of the ground station and the early warning machine are q _GS＝[0,0,0]^T and q _AWACS＝[0,0,H_AWACS]^T,H_AWACS respectively, which are Z-axis coordinates of the early warning machine. The system is regarded as a discrete time model, the movement time of the unmanned aerial vehicle is divided into N time slots with equal intervals, the time slot intervals are delta _t, and then the position coordinates of the unmanned aerial vehicle in the nth time slot are q [ N ] = [ x [ N ], y [ N ], h [ N ] ] ^T, n=0, 1.

The distance d [ n ] between the drone and the ground station may be expressed as:

The trajectory of the unmanned aerial vehicle may be represented by its position vector q [ n ], velocity vector v [ n ] and acceleration vector a [ n ]. The drone therefore has the following discrete state space model:

v[n+1]=v[n]+a[n]δ_t,n=0,...N (3)

The initial position q _I and the final position q _F, and the initial velocity v _I and the final velocity v _F of the unmanned plane are all determined in advance, and n=0 and n=n represent the initial time slot and the final time slot respectively, and then there are:

q[0]=q_I,q[N]＝q_F (4)

v[0]=v_I,v[N]＝v_F (5)

furthermore, the performance limitations of the drone should also be considered mathematically:

V_min≤||[n]||≤V_max,||a[n]||≤A_max (6)

Wherein V _max and V _min are respectively the maximum speed and the minimum speed of the unmanned aerial vehicle, a _max represents the maximum acceleration of the fixed-wing unmanned aerial vehicle, |·| represents the norm. In the present invention, we use standard fonts to represent scalar quantities and bold-faced words to represent vectors.

Step two, a system transmission link;

1. Laser link

h_FSO[n]＝h_a[n]h_t[n]h_p[n] (7)

Where h _a[n]、h_t [ n ] and h _p [ n ] represent atmospheric attenuation, turbulence fading and pointing error, respectively.

The scattering and absorption of atmospheric particles results in an atmospheric attenuation of the beam during propagation, which attenuation is related to the distance d n between the drone and the ground station, expressed as:

h_a[n]=exp(-ωd[n]) (8)

where ω is the attenuation coefficient of the laser link.

Atmospheric turbulence causes power loss and random phase fluctuations. The simulation was performed using the Gamma-Gamma distribution, expressed as:

Wherein, alpha and beta both represent fading parameters, h _t n can be used as the mean value To approximate the representation, there is

(Najafi,Marzieh,et al."Statistical modeling of the FSO fronthaul channel for UAV-based communications."IEEE Transactions on Communications 68.6(2020):3720-3736.).

The pointing error loss is expressed as:

In the formula, And θ are respectively the nth timeAnd the value corresponding to theta [ n ].

Omega _d is calculated by the following formula:

Where ω ₀ is the beam waist radius, λ is the wavelength of light, Referred to as the coherence length, C _n ² is the refractive index structural parameter of the laser beam.

The discrete-time data reachability model of the laser link at the ground station may be expressed as:

Where B _FSO is the laser link bandwidth, P _FSO is the peak power allowed by the ground station, For noise variance, the reachability coefficient k can be expressed as:

wherein e represents a natural number, and the free parameter μ ^* is a formula Is a unique solution to (c).

2. Radio frequency link

In addition, for a radio frequency link between the unmanned aerial vehicle and the ground station, due to shadow effect and reflection of obstacle signals, two different large-scale attenuation models of line-of-sight transmission and non-line-of-sight transmission need to be considered, τ _RF [ n ] can be expressed as:

Where β ₀ is the received power at reference distance d ₀ =1, α _RF is the path loss index, and ε is the additional attenuation factor due to the non-line-of-sight link.

In the formula, Representing the actual probability of line-of-sight propagation between the drone and the ground station taking into account non-line-of-sight link attenuation, P _LoS n=1/[ 1+C ·exp (-D [ n-C ]) ] represents the ideal line-of-sight propagation probability between the drone and the ground station, C and D being parameters (Lee,Ju-Hyung,et al."Spectral-efficient network design for high-altitude platform station networks with mixed RF/FSO system."IEEE Transactions on Wireless Communications 21.9(2022):7072-7087.),γ[n] that depend on propagation conditions being the angle between the line-of-sight link and the x-y plane.

The discrete time data reachability model for the radio frequency link at the ground station may be expressed as:

Where B _RF is the radio frequency bandwidth, P _RF is the radio frequency transmit power, and σ ² _RF is the radio frequency noise variance.

The communication between the drone and the ground station is a parallel laser/radio frequency hybrid link. As both links are active simultaneously, the ground station can acquire data from the laser and radio frequency transmitters on the drone. Thus, the total achievable rate of the ground station may be expressed as the sum of the laser link reachability and the radio frequency link reachability:

R_T[n]＝R_FSO[n]+R_RF[n] (19)

thirdly, constructing a cloud model;

The thick cloud is constructed as a cylindrical obstacle, and the coordinate q _cloud[n]＝[x_cloud[n],y_cloud[n],h_cloud]^T is taken as a cloud center, wherein x _cloud[n],y_cloud[n],h_cloud represents projection coordinates of the cloud center on an x axis, a y axis and a z axis respectively. R _cloud and V _cloud represent the cylinder bottom radius and the cylinder height, respectively. Phi _cloud is the azimuth angle of the cloud, and is defined as the included angle between the x-axis and the connection point of any point on the circumference of the upper and lower bottom surfaces of the cylinder and the projection point of the cloud center on the upper and lower surfaces of the cylinder, as shown in fig. 2. The size and location of the cloud and the speed of movement may be pre-detected by the weather radar.

As can be seen from fig. 2, the downlink from the pre-warning device to the ground station will be blocked by the upper bottom edge of the cylinder and the uplink from the ground station to the pre-warning device will be blocked by the lower bottom edge of the cylinder. In fig. 2 (a), the dashed line represents the critical state that the link is blocked, and the unmanned aerial vehicle relay is deployed in the shadow area, so that the communication between the early warning machine and the ground station can be ensured.

For ease of analysis, a two-dimensional vertical cross-section is used in fig. 2 (b) to analyze the feasible area of three-dimensional unmanned aerial vehicle deployment. The included angle between the connecting line of the ground station and the unmanned plane and the x-y plane is the elevation angle of the ground station, which is expressed as:

As shown in fig. 2 (b), the angle between the connection from the lower edge of the cylinder to the ground station and the x-y plane is the maximum elevation angle that keeps the line-of-sight link clear, expressed as:

Thus, a viable area for relay deployment of unmanned aerial vehicles Expressed as:

fourth, unmanned aerial vehicle energy consumption model construction;

The total energy consumption of the drone includes communication energy and propulsion energy. In a practical scenario the communication related energy (a few watts) of the drone is much smaller than the propulsion energy (a few hundred watts), so that it is ignored in the present invention, only the propulsion energy is considered.

The movement of the drone can be approximately decomposed into a banked horizontal turning (equal altitude flight) movement along the x-y plane and a flying height varying movement along the z-axis, where v _xy and v _z are the speed vectors of the drone on the x-y plane and the z-plane, respectively. The propulsion energy consumption of a fixed wing unmanned aerial vehicle is related to the stress of the fixed wing unmanned aerial vehicle during the movement process. As shown in fig. 3, the unmanned aerial vehicle is typically subjected to four forces during flight, gravity W, drag D, lift L, and thrust F. Gravity w=mg, m is the mass of the drone, g=9.8 m/s ² is the gravitational acceleration. The method comprises the following steps:

Lcosδ-W=ma_z (25)

Lsinδ=ma_⊥ (26)

F-D=ma_|| (27)

Where δ represents the unmanned tilt angle, a _⊥ and a _|| represent the acceleration components perpendicular and parallel to v _xy, respectively, and a _z is the acceleration component in the direction of v _z. The formula (26) and the formula (25) are divided to obtain:

Resistance is expressed as:

Wherein, c ₁ and c ₂ are two constant parameters, (Thibbotuwawa,Amila,et al."Energy consumption in unmanned aerial vehicles:A review of energy consumption models and their relation to the UAV routing."Information Systems Architecture and Technology:Proceedings of 39th International Conference on Information Systems Architecture and Technology–ISAT 2018:Part II.Springer International Publishing,2019.).V related to the mass and wing area of the unmanned aerial vehicle represents the unmanned aerial vehicle volume, the load coefficient k ^* is calculated by formula (26), formula (28) and gravity expression w=mg, where m represents the unmanned aerial vehicle mass, and g represents the gravitational acceleration:

thus, if a _z =0, the load factor In this case, the unmanned aerial vehicle is tilted horizontally to fly. Conversely, if a _z +.0, the change in fly height will cause a change in the load factor. From the expressions (27), (29), (30), the expression of the thrust force can be obtained as:

thus, the instantaneous power required by the drone can be expressed as:

Wherein a _⊥ and a _|| are represented by A _||＝a_xy ^Tv_xy/||v_xy |, wherein a _xy is the acceleration component of the unmanned aerial vehicle along v _xy. Thus, a fixed wing drone with a discrete-time three-dimensional trajectory q [ n ], under normal operating conditions, the total propulsion energy model is expressed as:

Where v _xy n and a _xy n are the unmanned speed and acceleration components, respectively, of the nth slot along the x-y plane and a _z n is the acceleration component along the z-axis.

Assuming a _xy ^T[n]v_xy [ n ] =0, the energy model can be reduced to:

Further, if it is assumed that the initial velocity v [0] and the final velocity v [ N ] are the same, the last term in equation (34) is zero.

Step five, optimizing the track of the unmanned aerial vehicle;

Step1, maximizing energy efficiency under constraint conditions;

Theoretically, the drone remains stationary at the nearest location to the ground station, and the highest data reachability may be obtained. However, it is almost impossible for a fixed wing unmanned aerial vehicle to hover at a strictly zero speed, and even if the unmanned aerial vehicle flies horizontally at a constant speed, the energy consumption is minimal, but the data accessibility is low when the unmanned aerial vehicle is far away from the ground station, and even a link break occurs. The aim of the invention is therefore not to pursue minimum energy consumption or maximum data accessibility. But to maximize communication energy efficiency.

Energy efficiency is defined as the ratio of the total data availability to the total energy consumption of the drone over the time of flight. As can be seen from the formulas (19) and (34), the energy efficiency of the unmanned aerial vehicle communication system can be expressed as:

the energy efficiency maximization problem of the constrained trajectory can be expressed as:

constraint (1) - (6), (36 b)

0≤h[n]≤H_AWACS(n=1,...,N)(36c)

R_T[n]≥R_th(n=1,...,N)(36d)

Wherein (36 b) indicates that the unmanned aerial vehicle trajectory is to satisfy the constraints of equations (1) - (6), and (36 c) is the actual constraint that the unmanned aerial vehicle flying height is to satisfy. (36d) Is a guarantee of the system for the quality of service requirement, i.e. the sum of the transmission rates of the parallel links needs to be larger than the minimum reachability threshold R _th. Further, (36 e) is a constraint on the position of the drone by the feasible region affected by the cloud obstacle.

Step2, a track optimization method;

When the unmanned aerial vehicle executes the communication relay task, the whole flight path is planned according to the communication task requirement and weather conditions, which can be regarded as a constraint optimization problem which is generally difficult to solve, and particularly has a plurality of non-convex constraints. To solve this problem, the present invention proposes a trajectory planning (Trajectory PLANNING GAIN SHARING knowledgebased, TPGSK) method based on obtaining shared knowledge.

The method comprises the following steps:

step1, initializing the maximum value G of the optimization iteration times, the population number N, the knowledge factor parameter k _f, the knowledge ratio parameter k _r and the greedy scaling factor l. The initial values of these parameters are set, for example, g=100, n=100, k _f＝0.5;k_r =0.9, and k=3.

Step2 randomly initializing population x _i,i＝1,2,…N,x_i to be the ith individual in the population.

Step3. Calculate the fitness function value f (x _i) for each individual in the population,The concrete steps are as follows:

Wherein η _E,η_R,η_φ,η_end is an energy efficiency target, a transmission constraint, an angle constraint and an endpoint constraint factor respectively, ω _E,ω_R,ω_φ,ω_end is a weight coefficient corresponding to 4 factors (ω _E＝10⁵,ω_R＝1,ω_φ＝10³,ω_end＝10³ is preferable according to different orders of magnitude of each factor), EE _T [ n ] represents an energy efficiency of the unmanned aerial vehicle communication system at the nth moment, R _T [ n ] represents a sum of transmission rates of parallel links at the nth moment, and R _th is a minimum accessibility threshold. Phi _AU,φ_AU,max is the elevation angle of the ground station and the maximum value thereof respectively, phi _GU,φ_GU,max is the depression angle of the early warning machine and the maximum value thereof respectively, q _I,q_F is the initial position and the final position of the unmanned aerial vehicle respectively, and v _I,v_F is the initial speed and the final speed of the unmanned aerial vehicle respectively.

Step4, the individuals in the population are ordered from small to large, wherein x _best,…x_i-1,x_i,x_i+1,…x_worst.x_best represents the individual with the smallest fitness value in the population, and x _worst represents the individual with the largest fitness value in the population.

Step5. primary acquisition and sharing knowledge phase, each x _i, i=1, 2,..n, is updated by equation (38) according to the probability of k _r. Updating x _i with the co-participation of the (i-1) th individual, the (i+1) th individual and the randomly selected individual in the population, thereby obtainingThe primary acquisition and shared knowledge phase population update strategy is shown in equation (38):

wherein, For updated individuals, x _r is a randomly selected individual.

Step6. The updated population individuals are ranked in order of fitness value from small to large, and then the ranked individuals are classified into 3 categories, namely, best individual, medium individual, worst individual, wherein the best individual accounts for p (e.g., 20%), the worst individual accounts for p (e.g., 20%), the medium individual accounts for 1-2p (e.g., 80%).

Step7. update each x _i, i=1, 2, according to the probability of k _r, by equation (39).

Wherein N _best,N_m,N_worst is the optimal individual number, the middle individual number and the worst individual number respectively.

Rank_m _{i_2},Rank_m_{i_2},Rank_worst_{i_3} is the order of arrangement of each individual in the optimal individual number, the intermediate individual number, and the worst individual number, respectively, and can be expressed as follows:

Rank_best_{i_1}＝l(N_best-i_1)+1 (43)

Rank_m_{i_2}＝l(N_m-i_2)+1 (44)

Rank_worst_{i_3}＝l(N_worst-i_3)+1 (45)

step8. Update the current optimal solution within the population according to equation (46).

ThenThenWherein, In order for the individual to be updated before,Is the fitness function value.Is thatIs a fitness function value of (a).

Step9 the globally optimal solution is updated according to equation (47).

ThenThen

Wherein, As an individual who is globally optimal,Is the fitness function value.

Step10. Returning to step4 until the number of optimization iterations reaches a maximum G.

Step11. Output the globally optimal solution.

And step six, simulating a result.

In order to verify the performance of the invention, MATLAB simulation software is adopted for track planning, and the maximum speed V _max = 100m/s, the minimum speed V _min = 3m/s, the maximum acceleration A _max＝5m/s², the initial position q _I＝[900,-300,1800]^T and the final position q _F＝[-700,600,1800]^T of the unmanned aerial vehicle are assumed. Slot size δ _t =0.4s, n=1000. The initial point of the cloud center is set to q _cloud＝[0,0,1800]^T,R_cloud＝500m,V_cloud =500 m. Other simulation parameters are shown in table 1.

Table 1 simulation parameter settings

The small aircrafts on the unmanned plane track are distributed at fixed time intervals, so that the degree of the density of the small aircrafts can reflect the speed of the unmanned plane. The denser the aircraft distribution on the trajectory, the longer it takes for the unmanned aerial vehicle to fly through the path, the slower the speed of the unmanned aerial vehicle, and vice versa.

Fig. 4 is a data reachability maximization trajectory that indicates that the drone is rapidly descending to near the ground station where the flight speed is slowed down, thereby maintaining a longer dwell time. To obtain maximum data reachability. This is because the data accessibility is closely related to the distance of the unmanned aerial vehicle from the ground, as shown in equations (13), (18). The trajectory of the drone in fig. 4 (b) is significantly shifted to the right down, avoiding the obstruction of the cloud.

Fig. 5 is an energy consumption minimization trace. Compared to fig. 4, it can be seen that the unmanned aerial vehicle in the energy minimization trajectory has a larger turning radius and smaller descending and ascending amplitudes. In addition, the red aircraft on the flight trajectory appears to be evenly distributed, indicating that the speed of the drone remains substantially uniform during the flight. This is because the propulsion energy consumption of the aircraft is proportional to the acceleration, as shown in equation (33). However, the energy minimization trajectory is much farther from the ground station most of the time, and thus the data reachability may be significantly lower. The slight drop in trajectory in fig. 5 (b) is caused by the constraint of the reachability threshold.

Fig. 6 shows an energy efficient maximization trajectory design where the drone is closer to the ground station while maintaining relatively gentle flight characteristics. This design aims to minimize energy consumption while ensuring optimal data accessibility. Further, as shown in fig. 6 (b), the unmanned aerial vehicle adopts a more energy-efficient spiral climbing mode than the mode in which the unmanned aerial vehicle directly ascends back to the destination in fig. 4.

Table 2 comparison of three track properties

Table 2 gives a detailed numerical comparison of these three different designs in terms of total data availability, energy consumption and energy efficiency. It can be observed that the energy efficiency maximization design proposed by the present invention achieves a good balance between maximizing data reachability and minimizing energy consumption. The energy efficiency maximization design achieves 120.16% of energy efficiency gain compared to the data rate maximization design, and 156.99% of energy efficiency gain compared to the energy consumption minimization design.

Fig. 7 is an energy efficiency maximization track of the unmanned aerial vehicle when the positions of the cloud centers are different, and tracks 1, 2 and 3 correspond to tracks when the cloud centers move to three different positions q _cloud＝[0,0,1800]^T、q_cloud2＝[100,100,1800]^T and q _cloud3＝[200,200,1800]^T. Along with the change of cloud center position, unmanned aerial vehicle deployment feasible regionAnd is constantly changing. When the cloud center is far away from the ground station, the feasible area in front of the cloud center moving direction is reduced, and the feasible area in back of the cloud center moving direction is increased. As shown in fig. 7, the range of motion of the drone in track 2 is significantly smaller relative to the range of motion on track 1. When the cloud center moves to q _cloud3, in order to ensure optimal energy efficiency, the track 3 bypasses from a position behind the cloud cover, because the deployment feasible area of the unmanned aerial vehicle behind the cloud center moving direction is larger.

The cloud moves in a speed direction at an angle of 30 DEG to the x-axis with speed magnitudes of 1km/h, 2km/h and 3km/h, and the energy efficiency trace is shown in FIG. 8. As the cloud moves, the deployment feasible region of the unmanned aerial vehicle changes. The faster the cloud moves, the farther from the ground station, the smaller its area of feasibility. Therefore, to avoid cloud occlusion, the range of motion of the drone in the z-axis is significantly reduced as the speed of cloud movement increases.

FIG. 9 is a graph of the effect of weather on system performance. Fig. 9 (a) simulations compare the maximum energy efficiency values for five weather conditions with laser/rf parallel links and rf only links with clear (ω=0.43×10 ^-3m^-1), haze (ω=4.2×10 ^-3m^-1), mist (ω=20×10 ^-3m^-1), medium mist (ω=42.2×10 ^-3m^-1), thick mist (ω=125×10 ^-3m^-1). In sunny days, the maximum energy efficiency of the unmanned aerial vehicle relay can reach 6×10 ⁸ bps/J. As the atmospheric attenuation value ω increases, the maximum energy efficiency rapidly decreases. Under dense fog conditions, the energy efficiency of the laser/rf parallel link is about 10 ⁴ bps/J, which is the same as a single rf link, indicating that the laser link has been completely broken by atmospheric effects. When ω is greater than 20×10 ^-3m^-1, the maximum energy efficiency of the laser/rf parallel link is very close to that of a single rf link.

Fig. 9 (b) and 9 (c) simulate the total data achievable rate and energy consumption curves, respectively, for different weather conditions. It is apparent that the trend of the data reachability curve in fig. 9 (b) is very consistent with the trend of the energy efficiency curve in fig. 9 (a). Whereas the energy consumption curve shown in fig. 9 (c) varies slightly over the range of about 1.8x10 ⁵ J to 2 x 10 ⁵ J, remaining substantially stable. This result indicates that the data accessibility is severely affected by weather, while unmanned energy is less sensitive to weather conditions.

Fig. 10 is a graph comparing energy efficiency of each time slot in the unmanned trajectory obtained by TPGSK and AGSK algorithms, respectively. As can be seen from the figure, the TPGSK method proposed by the present invention gives significantly higher energy efficiency than the comparative method AGSK.

Firstly, designing an airborne laser/radio frequency hybrid network structure based on unmanned aerial vehicle relay, secondly, comprehensively considering cloud cover, weather change data transmission rate threshold and other series of practical limitations faced by an aviation laser/radio frequency communication network, designing an unmanned aerial vehicle relay flight track optimization model which takes energy efficiency maximization as an objective function, modeling the cloud as an obstacle to obtain an unmanned aerial vehicle deployment feasibility domain which can ensure line-of-sight communication, particularly, deducing a fixed-wing unmanned aerial vehicle variable-height propulsion energy consumption formula, in addition, providing an efficient track planning method, verifying the effectiveness of the efficient track planning method, and finally, obtaining the track of maximum data accessibility, minimum energy consumption and optimal energy efficiency by the method, and analyzing the influence of weather change and cloud movement on the energy efficiency.

Claims

1. A method for optimizing the trajectory of drone relays in a laser/RF hybrid aviation communication network, which is used in an airborne laser/RF communication system based on fixed-wing drone relay assistance: parallel laser/RF hybrid transmission links are used for the AWACS to drone relay and the drone relay to the ground station; thick clouds with high liquid water content values are considered obstacles. In order to maintain the line-of-sight communication link of laser and RF, the drone relay is deployed in a position without cloud coverage, thereby bypassing the obstruction of obstacles and maintaining a high-speed connection between the AWACS and the ground station; the method specifically comprises the following steps:

Step 1: System model construction;

A three-dimensional rectangular coordinate system is established, with the center of gravity of the ground station as the origin, the east direction as the positive direction of the X-axis, the south direction as the positive direction of the Y-axis, and the direction perpendicular to the xy plane as the positive direction of the Z-axis. Assume that the three-dimensional position vectors of the ground station and the AWACS are q _GS = [0,0,0] ^T and q _AWACS = [0,0,H _AWACS ] ^T , respectively, and H _AWACS is the Z-axis coordinate of the AWACS. Consider the system as a discrete time model, divide the movement time of the UAV into N equally spaced time slots, and set the time slot interval to δ _t . Then, the position coordinates of the UAV in the nth time slot are q[n] = [x[n], y[n], h[n]] ^T , n = 0, 1, …, N, where x[n], y[n], h[n] represent the projection coordinates of the UAV on the x-axis, y-axis, and z-axis, respectively.

The distance d[n] between the UAV and the ground station is expressed as:

The trajectory of a drone is represented by its position vector q[n], velocity vector v[n], and acceleration vector a[n]; the drone has the following discrete state space model:

v[n+1]＝v[n]+a[n]δ _t ,n＝0,…N (3)

The initial position q _I and final position q _F of the drone as well as the initial velocity v _I and final velocity v _F are determined in advance. n = 0 and n = N represent the initial time slot and the final time slot respectively. Then:

q[0]＝q _I ,q[N]＝q _F (4)

v[0]＝v _I ,v[N]＝v _F (5)

Consider the performance limitations of a drone mathematically:

V _min ≤||v[n]||≤V _max ,||a[n]||≤A _max (6)

Where V _max and V _min are the maximum and minimum speeds of the UAV, respectively, A _max represents the maximum acceleration of the fixed-wing UAV, and ||·|| represents the norm;

Step 2: System transmission link;

1. Laser link

The fading experienced in the laser link transmission between the UAV and the ground station is composed of atmospheric attenuation, turbulence and pointing error. The attenuation coefficient in the nth time slot is expressed as:

h _FSO [n]＝h _a [n] h _t [n] h _p [n] (7)

Where _ha [n], _ht [n] and _hp [n] represent atmospheric attenuation, turbulence fading and pointing error, respectively;

The scattering and absorption of atmospheric particles cause atmospheric attenuation of the beam during propagation. This attenuation is related to the distance d[n] between the UAV and the ground station and is expressed as:

_ha [n]＝exp(-ωd[n]) (8)

Where ω is the attenuation coefficient of the laser link;

Atmospheric turbulence causes power loss and random phase fluctuations; this is simulated using the Gamma-Gamma distribution, expressed as:

Where α and β are both fading parameters, and h _t [n] is their mean To approximate, we have

The pointing error loss is expressed as:

Where u is the radial distance from the center of the beam to the center of the lens, A ₀ is the fraction of the optical power collected when the difference between the center of the spot and the center of the detector is zero, and its expression is A ₀ = erf(v ₁ )erf(v ₂ ), erf(·) is the error function, is the lower bound of the ratio of the receiver aperture radius a to the beam width ω _d at distance d[n], is the upper bound of the ratio of a to ω _d , is the angle between the laser beam in the direction of the UAV and the ground station and the z-axis, θ[n] is the angle between the projection of the laser beam on the xy plane and the x-axis, and the pointing error coefficient k _g at the nth moment is expressed as:

In the formula, and θ are the nth moment and the value corresponding to θ[n];

ω _d is calculated by the following formula:

Where _ω0 is the beam waist radius, λ is the wavelength of light, is called the coherence length, is the refractive index structure parameter of the laser beam;

Where B _FSO is the laser link bandwidth, P _FSO is the peak power allowed by the ground station, is the noise variance, and the accessibility coefficient k is expressed as:

Wherein, e represents a natural number, and the free parameter μ ^* is the unique solution of the formula α=1/μ ^* -e ^-μ* /(1-e ^-μ* );

2. RF link

The attenuation coefficient of the RF transmission link between the UAV and the ground station in the nth time slot is expressed as:

where τ _RF [n] represents the effect of large-scale fading such as path loss and shadow fading, represents the influence of small-scale fading, and its mean

Considering two different large-scale attenuation models of line-of-sight transmission and non-line-of-sight transmission, τ _RF [n] is expressed as:

Where β ₀ is the received power at the reference distance d ₀ = 1, α _RF is the path loss exponent, and ε is the additional attenuation factor due to the non-line-of-sight link;

The expected attenuation coefficient of the RF link is expressed as:

In the formula, represents the actual probability of line-of-sight propagation between the UAV and the ground station considering the attenuation of the non-line-of-sight link, P _LoS [n] = 1/[1+C exp(-D[γ[n]-C])] represents the ideal line-of-sight propagation probability between the UAV and the ground station, C and D are parameters that depend on the propagation conditions, and γ[n] is the angle between the line-of-sight link and the xy plane;

The discrete-time data reachability model of the RF link at the ground station is expressed as:

Where B _RF is the RF bandwidth, P _RF is the RF transmit power, and σ ² _RF is the RF noise variance;

The total achievable rate of the ground station is expressed as the sum of the laser link reachability and the RF link reachability:

R _T [n]＝R _FSO [n] + R _RF [n] (19)

Step 3: Cloud model construction;

Thick clouds are constructed as cylindrical obstacles, with the coordinates q _cloud [n] = [x _cloud [n], y _cloud [n], h _cloud ] ^T as the cloud center, where x _cloud [n], y _cloud [n], h _cloud represent the projection coordinates of the cloud center on the x-axis, y-axis and z-axis respectively; R _cloud and V _cloud represent the radius of the cylinder base and the height of the cylinder respectively; φ _cloud is the azimuth of the cloud, which is defined as the angle between the connection point of any point on the circumference of the upper and lower base of the cylinder and the projection point of the cloud center on the upper and lower surfaces of the cylinder and the x-axis; the size, position and moving speed of the cloud are pre-detected by the meteorological radar;

The downlink from the AWACS to the ground station will be blocked by the upper edge of the cylinder, and the uplink from the ground station to the AWACS will be blocked by the lower edge of the cylinder. Deploying drone relays in areas where neither is blocked can ensure communication between the AWACS and the ground station.

The feasible area of 3D UAV deployment is analyzed using a 2D vertical profile. The angle between the line connecting the ground station and the UAV and the x-y plane is the ground station elevation angle, which is expressed as:

The angle between the connection from the lower edge of the cylinder to the ground station and the x-y plane is the maximum elevation angle to keep the line-of-sight link open, expressed as:

The angle between the connection between the AWACS and the UAV and the x-y plane is the pitch angle of the AWACS, expressed as:

The angle between the connection between the upper edge of the cylinder and the AWACS and the x-y plane is the maximum pitch angle of the AWACS, expressed as:

Feasible areas for drone relay deployment It is expressed as:

Step 4: Construction of UAV energy consumption model;

The motion of the UAV is approximately decomposed into the inclined horizontal turning motion along the xy plane and the flight altitude change motion along the z axis, where _vxy and _vz are the velocity vectors of the UAV on the xy plane and the z plane respectively; the propulsion energy consumption of the fixed-wing UAV is related to the force it is subjected to during the motion process; the UAV is subjected to four forces during the flight: gravity W, resistance D, lift L and thrust F; gravity W = mg, m is the mass of the UAV, and g = 9.8m/ ^s2 is the acceleration of gravity; we get:

Lcosδ-W＝ma _z (25)

Lsinδ＝ma _⊥ (26)

FD＝ma _|| (27)

Where δ represents the tilt angle of the drone, a _⊥ and a _|| represent the acceleration components perpendicular to and parallel to v _xy , respectively, and a _z is the acceleration component along the v _z direction. Dividing equation (26) and equation (25), we get:

The resistance is expressed as:

Among them, _c1 and _c2 are two constant parameters related to the mass and wing area of the UAV; V represents the volume of the UAV, and the load coefficient k ^* is calculated by equation (26), equation (28) and the gravity expression W = mg, where m represents the mass of the UAV and g represents the gravitational acceleration:

If a _z = 0, the load factor In this case, the UAV tilts and turns horizontally. On the contrary, if a _z ≠ 0, the change of flight altitude will cause the change of load factor. According to equations (27), (29), and (30), the thrust expression is:

The instantaneous power required by the drone is expressed as:

In the formula, a _⊥ and a _|| are expressed as a _|| ＝a _xy ^T v _xy /||v _xy ||, where a _xy is the acceleration component of the UAV along v _xy ; therefore, for a fixed-wing UAV with a discrete-time three-dimensional trajectory q[n], under normal operation, the total propulsion energy model is expressed as:

Where v _xy [n] and a _xy [n] are the velocity and acceleration components of the UAV along the xy plane at the nth time slot, respectively, and a _z [n] is the acceleration component along the z axis;

Assuming that a _xy ^T [n]v _xy [n] = 0, the energy model is simplified to:

If it is assumed that the initial velocity v[0] and the final velocity v[N] are the same, the last term in equation (34) is zero;

Step 5: UAV trajectory optimization;

Step 1: Maximize energy efficiency under constraints;

Energy efficiency is defined as the ratio of the total data reachability to the total energy consumption of the UAV during the flight time. From equations (19) and (34), the energy efficiency of the UAV communication system is expressed as:

Constraints: (1)-(6), (36b)

0≤h[n]≤H _AWACS (n＝1,…,N) (36c)

R _T [n]≥R _th (n＝1,…,N) (36d)

Among them, (36b) indicates that the trajectory of the UAV must satisfy the constraints of formulas (1)-(6), (36c) is the actual constraint that the flight altitude of the UAV needs to meet; (36d) is the guarantee of the system's service quality requirements, that is, the sum of the transmission rates of the parallel links needs to be greater than the minimum reachability threshold R _th ; (36e) is the constraint on the UAV position in the feasible domain affected by cloud obstacles;

Step 2, trajectory optimization method;

A trajectory planning method based on acquiring shared knowledge is as follows:

Step 1. Initialize the maximum number of optimization iterations G, the population number N, the knowledge factor parameter k _f , the knowledge ratio parameter k _r , and the greediness ratio factor l;

Step 2. Randomly initialize the population x _i , i = 1, 2, ... N, x _i is the i-th individual in the population;

Step 3. Calculate the fitness function value for each individual in the population Specifically expressed as:

Among them, η _E ,η _R ,η _φ ,η _end are energy efficiency target, transmission constraint, angle constraint and end point constraint factors respectively; ω _E ,ω _R ,ω _φ ,ω _end are weight coefficients corresponding to the four factors respectively; UAV _EET [n] represents the energy efficiency of the UAV communication system at the nth moment; φ _AU ,φ _AU,max are the elevation angle of the ground station and its maximum value respectively; φ _GU ,φ _GU,max are the depression angle of the early warning aircraft and its maximum value respectively; q _I ,q _F represent the initial position and final position of the UAV respectively; v _I ,v _F represent the initial speed and final speed of the UAV respectively;

Step 4. Sort the individuals in the population in ascending order of fitness value. The sorting results are as follows: x _best ,…xi _-1 , _xi , _xi+1 ,…x _worst ; x _best represents the individual with the smallest fitness value in the population, and x _worst represents the individual with the largest fitness value in the population;

Step 5. In the primary stage of acquiring and sharing knowledge, according to the probability of k _r , update each x _i , i = 1, 2, ... N, through formula (38); use the i-1th individual, i+1th individual and randomly selected individuals in the population to jointly participate in updating x _i , so as to obtain The population update strategy in the primary acquisition and sharing knowledge stage is shown in formula (38):

in, is the updated individual, and x _r is a randomly selected individual;

Step 6. Sort the updated population individuals in order of fitness value from small to large, and then divide the sorted individuals into three categories, namely the best individuals, the middle individuals, and the worst individuals, where the best individuals account for p, the worst individuals account for p, and the middle individuals account for 1-2p;

Step 7. Update each x _i , i = 1, 2, ... N, according to the probability of k _r using formula (39);

The population update strategy in the advanced knowledge acquisition and sharing stage is shown in formula (39):

Among them, x _p-best , x _m , x _p-worst are individuals randomly selected from the best individuals, individuals randomly selected from the worst individuals, and individuals randomly selected from the middle individuals, respectively. The calculation method of their selection probabilities pr_best _{i_1} , pr_m _{i_2} , pr_worst _{i_3} is:

Among them, N _best , N _m , and N _worst are the best number of individuals, the middle number of individuals, and the worst number of individuals respectively;

Rank_m _{i_2} , Rank_m _{i_2} , Rank_worst _{i_3} are the ranking order of each individual in the optimal number of individuals, the middle number of individuals, and the worst number of individuals, respectively, expressed as follows:

Rank_best _{i_1} =l(N _best -i_1)+1 (43)

Rank_m _{i_2} =l(N _m -i_2)+1 (44)

Rank_worst _{i_3} =l(N _worst -i_3)+1 (45)

step8. Update the current optimal solution in the population according to formula (46);

in, For the individual before the update, is its fitness function value; for The fitness function value of

step9. Update the global optimal solution according to formula (47);

in, is the globally optimal individual, is its fitness function value;

step10. Return to step4 until the number of optimization iterations reaches the maximum value G;

step11. Output the global optimal solution.

2. The method for optimizing the trajectory of a UAV relay in a laser/RF hybrid aviation communication network according to claim 1, wherein in step 5, Step 2 and Step 1, the initial values of the parameters are set as follows: G = 100; N = 100; k _f = 0.5; k _r = 0.9; k = 3.

3. The method for optimizing the UAV relay trajectory in the laser/RF hybrid aviation communication network according to claim 1, characterized in that, in step 5, Step 2 and Step 3, ω _E = 10 ⁵ , ω _R = 1, ω _φ = 10 ³ , ω _end = 10 ³ .

4. The method for optimizing the UAV relay trajectory in the laser/RF hybrid aviation communication network as described in claim 1 is characterized in that in step 5 Step 2 and step 6, the best individual proportion p is 20%, the worst individual proportion p is 20%, and the medium individual proportion 1-2p is 80%.