Parameter configuration method for 6G smart reflector-assisted drone network

Abstract

This invention proposes a parameter configuration method for a 6G intelligent reflector-assisted UAV network, relating to the field of wireless communication networks, and particularly to network planning and network planning tools. Specifically, firstly, considering the occlusion effect of an IRS deployed on a building with unidirectional orientation and the user's location, the occlusion factor of relevant links is quantified. Under an unobstructed feasible set, the user selects a set of nearby IRS services based on the corresponding path loss index. The channel state information set under IRS assistance is obtained based on connection matching. Substituting this into the system model, optimal beamforming is jointly designed to obtain the optimized performance of the IRS-assisted UAV network in dense urban scenarios. Based on the performance under different parameter configurations, a superior IRS-UAV collaborative configuration is obtained under different occlusion scenarios to optimize and improve system efficiency.

Description

Technical Field

[0001] This invention relates to the field of wireless communication networks, and in particular to network planning, network planning tools, and a parameter configuration method for intelligent reflective surface (IRS)-assisted unmanned aerial vehicle (UAV) networks. Background Technology

[0002] Currently, 5G has basically achieved wide coverage and is being rapidly deployed globally, aiming to connect hundreds of millions of devices, achieve high capacity growth, and enable ubiquitous connectivity. However, due to the inherent uncontrollable fading and correlation in the channels, simply increasing the number of sites or antennas to expand the capacity of 5G networks has bottlenecks and cannot meet the needs of sustainable capacity growth in future wireless networks.

[0003] Consequently, research on future 6G networks has gradually unfolded. Among them, Intelligent Reflective Surfaces (IRS) have emerged as a candidate technology to overcome the challenges of highly complex, costly, and energy-intensive networks, standing out due to their unique characteristics of low cost, low energy consumption, programmability, and ease of deployment. IRS is essentially a metamaterial with programmable electromagnetic properties, composed of a large number of low-cost passive reflective units. Each reflective unit can independently generate amplitude and phase changes in the incident signal, thereby collaboratively achieving three-dimensional reflected beamforming.

[0004] Currently, research on IRS-assisted wireless communication systems is still in its early stages. The core of this research involves using IRS with reconfigurable reflective properties to cover the ground, buildings, drones, and even airships. Multiple independent reflective elements can enrich channel scattering conditions, enhancing the diversity multiplexing gain of the wireless communication system. IRS possesses the characteristic of generating ideal multipath effects, thus enabling the reconstruction of line-of-sight (LoS) links to improve situations where direct links are blocked, and expanding signal coverage in congested communication scenarios. IRS can also achieve signal propagation direction modulation and in-phase superposition in three-dimensional space, increasing received signal strength, suppressing co-channel interference, and improving transmission performance between communication devices. Furthermore, IRS can be used to improve physical layer security and enable synchronous wireless information and power transmission between miscellaneous devices in IoT networks.

[0005] As can be seen, IRS has great potential for enhancing coverage and capacity of future wireless networks. Similarly, to fully utilize its reconfigurable beyond-line-of-sight (LOS) links, the introduction of UAV communication networks into IRS has considerable untapped potential. Compared to traditional terrestrial communication networks, UAV networks have the advantage of establishing LoS communication links, flexible deployment, on-demand scheduling, and good mobility, making them capable of handling communication tasks in emergency situations and hotspot areas. However, UAV communication also has certain limitations and faces many challenges in real-world applications, especially in densely populated urban areas where numerous buildings, dense facilities, trees, and other tall structures obstruct the view. This significantly hinders the provision of visual communication services by UAVs, making communication connections prone to congestion and affecting user service quality, with edge users experiencing particularly severe performance impacts.

[0006] IRS (Internal Reflectors) can effectively compensate for Loss of Space (LoS) links, filling gaps and hotspots, and its low cost and energy efficiency do not lead to excessive energy consumption. Specifically, in dense urban environments where severe obstruction by buildings and other structures occurs, IRS-assisted drone communication networks can leverage the reconfigurable, rich-scattering characteristics of IRS in hotspot areas to improve channel capacity. Furthermore, when the LoS link between drones and users is congested, deploying one or more IRSs in the system can reconstruct multiple LoS links to fill gaps, establish more stable LoS connections with ground users, bypass obstacles to serve obstructed users, save drone propulsion energy, and reduce access latency. Therefore, based on intelligent reflectors and drone technology—two emerging communication technologies—and combined with relevant scenarios to identify pain points and challenges, exploring the future development direction of intelligent and green networks is of great practical significance. Summary of the Invention

[0007] This invention proposes a parameter configuration method for a 6G intelligent reflector-assisted UAV network. Specifically, firstly, considering the occlusion effect of an IRS deployed on a building with unidirectional orientation and the user's location, the occlusion factor of relevant links is quantified. Under the feasible set of unoccluded links, the user selects a set of nearby IRS services based on the corresponding path loss index. The channel state information set under IRS assistance is obtained based on the connection matching situation. Substituted into the system model, the optimal beamforming is jointly designed to obtain the optimized performance of the IRS-assisted UAV network in dense urban scenarios. Based on the performance under different parameter configurations, a better configuration mode for IRS and UAV collaboration under different occlusion scenarios is obtained to achieve system efficiency optimization and improvement.

[0008] The parameter configuration method for the 6G intelligent reflector-assisted UAV network of the present invention includes the following steps:

[0009] Step 200: Considering the one-sided orientation of the IRS, quantify the shading factor of the relevant links based on the shading effect of the IRS deployed on the building and the user's location.

[0010] Considering a scenario where buildings are rectangular and randomly distributed, and the IRS (Internet Retrieval System) is mounted at a certain height on nearby buildings according to deployment settings, there will be three types of links in the IRS-assisted drone network: drone-user, drone-IRS, and IRS-user. For each link, occlusion is determined based on the geometric positional relationship of its corresponding network elements. An occlusion factor w is introduced. A,B This characterizes whether the link AB related to network elements A and B is blocked. The blocking factor is determined by judging whether the link passes through a building in the 3D model. If it passes through a building, it is assumed that the building is generally impenetrable, resulting in severe signal loss, and therefore the blocking factor is set to 1. The Bernoulli random variable blocking factor w is then used. A,B ∈{0,1}.

[0011] Meanwhile, considering that IRS mounted on buildings is subject to practical physical limitations and can only serve users whose IRS and base station are located on the same side, specifically, relevant location parameters are obtained, and in the horizontal two-dimensional angle, cosη is determined. uav,irs cosη irs,user Whether the value is greater than 0 determines whether the drone and the user are on the same side of the extension line of the IRS plane, where cosη uav,irs Cosη represents the cosine of the angle between the vector from the IRS to the UAV and the perpendicular vector to the longer side of the IRS. irs,user This represents the cosine of the angle between the vector from the IRS to the user and the perpendicular vector to the longer side of the IRS.

[0012] In other words, for a user to successfully receive assistance through the IRS, both of the following conditions must be met simultaneously:

[0013] 1. The drone and the user are on the same side of the extended IRS plane, i.e., cosη uav,irs cosη irs,user >0;

[0014] 2. Both links from the IRS to the drone and from the IRS to the user are unblocked, meaning the corresponding blocking factor is zero.

[0015] Step 210: In the feasible set of unobstructed locations, the user selects a set of nearby IRS services based on the corresponding path loss metric.

[0016] Based on the occlusion factor and service status determined in step 200, the first step is to determine the occlusion factor w between the drone and the user. uav,user The 0 and 1 cases can be used to construct a serviceable set, where the set of users that can be served by drones is...

[0017]

[0018] Where k represents the user identifier and K represents the total number of users.

[0019] The set of users that each IRS can serve can be obtained from the IRS serviceability criteria. For the l-th, l=1,...,L IRS, the set of users it can serve is... This can be expressed mathematically as

[0020]

[0021] At the same time, there exists a feasible set of available IRSs for user k. For each user, obtaining the reference signal under the corresponding feasible IRS set is represented in the model as the path loss PL of the corresponding IRS cascaded link. l,k Based on path loss (i.e., signal strength), select L0 IRSs with better channel conditions and stronger signals to determine the set of IRSs serving user k. And update the user set of the IRS service accordingly.

[0022] Step 220: Obtain the IRS-assisted channel state information set based on the connection matching of each network element.

[0023] Based on the service set obtained from the connection matching in step 210, the connection status of the entire system can be determined. That is, the set of users who can normally receive drone or IRS-assisted services under the building occlusion effect can be defined as...

[0024] Let the channel state information between the UAV and user k be Considering the blocking effect, the actual direct channel state information between the UAV and user k is represented as follows: in

[0025]

[0026] Next, the Channel State Information (CSI) for each user under multi-IRS cooperative assistance is obtained, which can be expressed in the following form.

[0027]

[0028] This is an equivalent combination channel of the UAV-user direct connection channel and the IRS concatenated channel. This represents the channel state information between the UAV and the l-th IRS. This refers to the channel state information between the l-th IRS and user k. Let l be the phase matrix of the l-th IRS reflective unit. Each IRS has M units. Then... This refers to the limited channel state information used in subsequent beamforming processes, taking into account the blocking effect and the one-way service characteristics of the IRS.

[0029] Step 230: Substitute the system model and jointly design the optimal beamforming to obtain the optimized performance of the IRS-assisted UAV network in dense urban scenarios.

[0030] In scenarios where buildings obstruct the view, this invention, constrained by link obstruction, power limitations, and reflection phase constraints, optimizes the user-weighted rate sum as the target, where the optimization variable is the beamforming matrix {V}. k} and phase matrix {Φ l The optimization problem is as follows:

[0031]

[0032]

[0033]

[0034]

[0035] Where, ω k Let ω be the user weight coefficient, representing user fairness. When the optimization objective function is the network throughput, all users are fair, i.e., ω = 0. k All are equal to 1. In the constraint, P represents the total power of the base station, meaning the total power transmitted to users cannot exceed the base station's total power budget. If a user fails to receive service from either the drone or the IRS, the drone will not target them with service signals. The third constraint represents the unit modulus constraint of the IRS reflection phase, R... k Let k be the achievable rate for user k.

[0036] Note that the objective function in the optimization target is a non-convex function, and the phase constraint is also non-convex. This form is not easy to solve, so it is necessary to transform the target form and apply a beamforming matrix {V}. k} and phase matrix {Φ l The two optimization objectives are solved iteratively using an alternating optimization method. The beamforming matrix is ​​solved by exploring the relationship between users, rate, and mean square error, while the IRS phase is solved by manifold optimization Riemann conjugate gradient descent method to obtain the optimal value.

[0037] Step 240: Based on the performance results under different parameter configurations in step 230, obtain the optimal configuration for IRS and UAV collaboration under different occlusion scenarios, so as to optimize and improve system efficiency.

[0038] In specific scenarios, the optimal sum and rate performance under different parameter configurations guides the takeoff and landing altitude of drones and the deployment configuration of IRS. Specifically, it explores the performance trends of drones at different altitudes under different configuration scenarios, guides the configuration of IRS and drone parameters under specific scenario parameters, explores the performance differences between centralized and distributed deployment of IRS under different degrees of occlusion, and overcomes the impact of occlusion to maximize the spectrum efficiency of the IRS-assisted drone service system when the number of IRS units is limited.

[0039] Beneficial effects

[0040] This invention proposes a parameter configuration method for a 6G intelligent reflector-assisted UAV network. It establishes an occlusion and system channel model for the IRS-assisted UAV network in occlusion scenarios, considering the actual unidirectional service characteristics of the IRS and proposing a user-centric IRS connection matching strategy. A general and rate optimization algorithm under the proposed occlusion model is presented, using an alternating optimization method to handle non-convex complex problems. Optimization designs for UAV base station beamforming and IRS phase are given separately, obtaining the optimized system rate during the iterative process. In a multi-IRS multi-user system, inter-user interference is effectively suppressed, verifying the system performance gain of the IRS-assisted UAV architecture. Simulation analysis is conducted to examine the impact of various parameters on system and rate performance, providing guidance on the collaborative parameter configuration of UAVs and IRS to avoid ineffective IRS deployment and overhead, thereby improving system efficiency. Attached Figure Description

[0041] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0042] Figure 1 This is a schematic diagram of an example scenario of the present invention;

[0043] Figure 2 This is a schematic diagram of the system model of the intelligent reflective surface-assisted UAV network in dense urban areas according to the present invention;

[0044] Figure 3 This is a flowchart illustrating the implementation of the method in an example of the present invention;

[0045] Figure 4 This is a graph showing the changes in system throughput performance and rate performance under different numbers of intelligent reflector units.

[0046] Figure 5 This is a graph showing the relationship between system and rate performance at different UAV altitudes and different smart reflector deployment parameter configurations.

[0047] Figure 6 This is a graph showing the relationship between the number of users who can receive services normally under different building densities and the different deployment parameter configurations of smart reflective surfaces. Detailed Implementation

[0048] This invention proposes a parameter configuration method for a 6G intelligent reflector-assisted UAV network, with the network scenario shown in the attached figure. Figure 1 As shown, attached Figure 1 This demonstration showcases a network schematic illustrating how a drone base station assists edge users in a dense urban environment using intelligent reflective surfaces. Specifically, in dense urban scenarios, considering occlusion effects, the link between edge users and the drone base station is easily blocked. While increasing the probability of drones using these surfaces would significantly increase path loss, impacting system performance, intelligent reflective surfaces deployed on buildings can enhance signals and reconstruct the line-of-sight link, filling blind spots and expanding coverage. These intelligent reflective surfaces can be deployed in a distributed manner or centrally on a single board. Utilizing the multi-link gain from multiple units, and further enhancing the overall system performance through the joint design of active beamforming and passive phase control for both the drone base station and the intelligent reflective surfaces, the overall system performance can be significantly improved.

[0049] The algorithm flow for this case is attached. Figure 2 As shown, the specific implementation steps are as follows:

[0050] Step 300: Considering the one-sided orientation of the IRS, the occlusion factor of the relevant links is quantified based on the occlusion effect of the IRS deployed on the building and the user's location.

[0051] Specifically, a schematic diagram of the occlusion judgment model under this occlusion is attached. Figure 3 As shown. Considering a scenario where buildings are rectangular and randomly distributed, and the IRS is mounted at a certain height on a nearby building according to its deployment settings, there will be three types of links in the IRS-assisted drone network: drone-user, drone-IRS, and IRS-user. For each link, occlusion is determined based on the geometric positional relationship of its corresponding network elements.

[0052] Introducing occlusion factor w A,B This characterizes whether the link AB related to network element A and B is obstructed. For all buildings that pass through the projection of AB, it determines whether their height is greater than their corresponding height. Let the height h of network element A be... A The height h is less than B B According to geometric relationships, we can obtain Where d ALet A represent the horizontal distance from the obstructing building, and D represent the horizontal distance between network element A and network element B. The value of the obstruction factor is determined by judging whether the link passes through a building in the 3D model. If it passes through a building, it is assumed that the building is generally impenetrable, resulting in severe signal loss, and therefore the obstruction factor is judged to be 1. The Bernoulli random variable obstruction factor w is then... A,B ∈{0,1}.

[0053] Meanwhile, considering that IRS mounted on buildings is subject to practical physical limitations and can only serve users whose IRS and base station are located on the same side, specifically, relevant location parameters are obtained, and in the horizontal two-dimensional angle, cosη is determined. uav,irs cosη irs,user Whether the value is greater than 0 determines whether the drone and the user are on the same side of the extension line of the IRS plane, where cosη uav,irs Cosη represents the cosine of the angle between the vector from the IRS to the UAV and the perpendicular vector to the longer side of the IRS. irs,user This represents the cosine of the angle between the vector from the IRS to the user and the perpendicular vector to the longer side of the IRS.

[0054] In other words, for a user to successfully receive assistance through the IRS, both of the following conditions must be met simultaneously:

[0055] 1. The drone and the user are on the same side of the extended IRS plane, i.e., cosη uav,irs cosη irs,user >0;

[0056] 2. Both links from the IRS to the drone and from the IRS to the user are unblocked, meaning the corresponding blocking factor is zero.

[0057] Step 310: Under the feasible set of unobstructed locations, the user selects a set of nearby IRS services based on the corresponding path loss metric.

[0058] Based on the occlusion factor and service status determined in step 300, the first step is to determine the occlusion factor w between the drone and the user. uavuser The 0 and 1 cases can be used to construct a serviceable set, where the set of users that can be served by drones is...

[0059]

[0060] Where k represents the user identifier and K represents the total number of users.

[0061] The set of users that each IRS can serve can be obtained from the IRS serviceability criteria. For the l-th, l=1,...,L IRS, the set of users it can serve is... This can be expressed mathematically as

[0062]

[0063] At the same time, there exists a feasible set of available IRSs for user k. For each user, obtaining the reference signal under the corresponding feasible IRS set is represented in the model as the path loss PL of the corresponding IRS cascaded link. l,k Based on path loss (i.e., signal strength), select L0 IRSs with better channel conditions and stronger signals to determine the set of IRSs serving user k. And update the user set of the IRS service accordingly.

[0064] Step 320: Obtain the IRS-assisted channel state information set based on the connection matching situation.

[0065] Based on the service set obtained from the connection matching in step 310, the connection status of the entire system can be determined. That is, the set of users who can normally receive drone or IRS-assisted services under the building occlusion effect can be defined as...

[0066] Let the channel state information between the UAV and user k be Considering the blocking effect, the actual direct channel state information between the UAV and user k is represented as follows: in

[0067]

[0068] Next, the Channel State Information (CSI) for each user under multi-IRS cooperative assistance is obtained, which can be expressed in the following form.

[0069]

[0070] This is an equivalent combination channel of the UAV-user direct connection channel and the IRS concatenated channel. This represents the channel state information between the UAV and the l-th IRS. This refers to the channel state information between the l-th IRS and user k. Let l be the phase matrix of the l-th IRS reflective unit. Each IRS has M units. Then... This refers to the limited channel state information used in subsequent beamforming processes, taking into account the blocking effect and the one-way service characteristics of the IRS.

[0071] Step 330: Substitute the system model and jointly design the optimal beamforming to obtain the optimized performance of the IRS-assisted UAV network in dense urban scenarios.

[0072] After obtaining the relevant channel state information, the users who can receive the service are identified. Received power y k

[0073]

[0074] in The drone sends a signal vector to user k. For the corresponding beamforming matrix, d represents the number of data streams during transmission. Let K be the noise vector at user k, and assume it follows a cyclic complex Gaussian distribution. This represents the noise variance.

[0075] In scenarios where buildings obstruct the view, this invention, constrained by link obstruction, power limitations, and reflection phase constraints, optimizes the user-weighted rate sum as the target, where the optimization variable is the beamforming matrix {V}. k} and phase matrix {Φ l The optimization problem is as follows:

[0076]

[0077]

[0078]

[0079]

[0080] Where, ω k Let ω be the user weight coefficient, representing user fairness. When the optimization objective function is the network throughput, all users are fair, i.e., ω = 0. k All values ​​are equal to 1. In the constraint, P represents the total power of the base station, meaning the total power transmitted to users cannot exceed the base station's total power budget. If a user fails to receive service from either the drone or the IRS, the drone will not target them with service signals. The third constraint represents the unit modulus constraint of the IRS reflection phase.

[0081] The reachable rate of user k can then be expressed as:

[0082]

[0083] in This represents the sum of inter-user interference and noise received by user k.

[0084] Note that the objective function in the optimization target is a non-convex function, and the phase constraint is also non-convex. This form is not easy to solve, so it is necessary to transform the target form and apply a beamforming matrix {V}. k} and phase matrix {Φ l The two optimization objectives are solved iteratively using an alternating optimization method. The beamforming matrix is ​​solved by exploring the relationship between users, rate, and mean square error, while the IRS phase is solved by manifold optimization Riemann conjugate gradient descent method to obtain the optimal value.

[0085] Specifically, we first explore the relationship between users and rate with mean square error, transforming the original non-convex problem by considering a linear receiver filter. The estimated signal vector is The mean squared error matrix for each user is calculated as follows:

[0086]

[0087] Next, an auxiliary weight matrix associated with the user is introduced. The objective optimization problem is equivalently transformed into a mean squared error minimization problem.

[0088]

[0089]

[0090]

[0091]

[0092] With the beamforming matrix, phase matrix, and weighting matrix fixed as optimization variables, optimize the receiving matrix. Then, with the beamforming matrix, phase matrix, and optimized receiving matrix fixed, optimize the weighting matrix. Note that the objective function is concave with respect to both the receiving matrix and the weighting matrix. According to the first-order convexity condition for optimal solution, the optimized receiving matrix and weighting matrix can be expressed in the following forms:

[0093]

[0094]

[0095] Based on the optimized receiver matrix and weighting matrix, the beamforming matrix of the UAV base station is optimized. Considering the transmit power limitation, the optimization objective expression can be obtained by constructing a Lagrangian function and further simplifying it.

[0096]

[0097] in, Based on the first-order optimization conditions of the Lagrange function mentioned above, V k Taking the derivative and setting it to zero, we can obtain the optimal beamforming matrix V for user k. k Mathematically represented as

[0098]

[0099] μ is the multiplier term introduced by the corresponding power constraint, which can be updated using the bisection method. To satisfy the complementary relaxation condition of the power constraint, the beamforming matrix V... k Let μ be a function of μ. To maximize the satisfaction of the power constraint, the following condition must be met.

[0100] Next, with the receiver matrix, weighting matrix, and beamforming matrix fixed, the IRS phase matrix is ​​optimized. Ignoring irrelevant constant terms, the objective function can be reorganized into the following function.

[0101]

[0102] in, Substitute into the equivalent channel model After rearrangement, the optimization problem concerning the phase matrix can be reformulated as follows:

[0103]

[0104] in, q l =[[Q l ] 1,1 ,...,[Q l ] M,M ] T , ⊙ represents the Hadamard.

[0105] Due to the occlusion effect and users' choice of IRS, there is a different set of IRS services for each user, and each IRS also has its own set of users. Special attention to the definition Define the coefficient matrix

[0106] Due to the unit modulus constraint of the phase, this optimization problem remains a non-convex problem, and this constraint becomes the main obstacle to solving the problem. The Riemann conjugate gradient method in manifold optimization is used to solve it.

[0107] Finally, the optimized active beamforming matrix, phase matrix, receiver matrix, and weighting matrix are obtained. Through alternating iterations, the convergence value of the sum and rate is finally obtained, which is the optimized system performance.

[0108] The optimization process of the above joint beamforming algorithm can be summarized as follows:

[0109] The first step is to obtain the CSI set corresponding to the system based on the occlusion and matching conditions, and then construct the system model;

[0110] The second step is to make the outer loop D o =0, initialize the beamforming matrix and make it satisfy the power constraint condition.

[0111] The third step is to fix the phase matrix and beamforming matrix, and then calculate the optimized user-side receiving matrix and weighting matrix.

[0112] The fourth step involves fixing the optimized user reception matrix and the optimized weight matrix, and then performing beamforming design for users within the service cluster. Update μ using the binary search method;

[0113] Fifth, repeat steps three and four until the sum rate converges or the expected number of iterations is reached;

[0114] Step 6: Based on the optimized beamforming matrix, solve the phase optimization problem, starting with solving the objective function. Euclidean gradient

[0115] Step 7: Solve for the Riemann gradient gradf(φ) and obtain the update direction vector for the next iteration. β k For Polak-Ribiere parameters;

[0116] Step 8: Pull the updated vector back into the manifold space and obtain the updated phase. α k The step size for the Armijo backtracking search;

[0117] Step 9: Repeat steps 3 through 8 until the preset maximum number of iterations is reached or the objective function converges, to obtain the final optimized beamforming matrix and phase matrix, as well as the system's user and rate performance.

[0118] Step 340: Based on the performance results obtained from the optimization under different parameter configurations in step 330, a better configuration method for IRS and UAV collaboration under different occlusion scenarios is obtained to achieve system efficiency optimization and improvement.

[0119] In specific scenarios, the optimal sum and rate performance under different parameter configurations guides the takeoff and landing altitude of drones and the deployment configuration of IRS. Specifically, it explores the performance trends of drones at different altitudes under different configuration scenarios, guides the configuration of IRS and drone parameters under specific scenario parameters, explores the performance differences between centralized and distributed deployment of IRS under different degrees of occlusion, and overcomes the impact of occlusion to maximize the spectrum efficiency of the IRS-assisted drone service system when the number of IRS units is limited.

[0120] Simulation results are attached. Figure 4 Appendix Figure 5 With appendix Figure 6 As shown in the simulation, users are randomly distributed within a 600×600m area, located outside buildings, with a total of 20 users. The UAV is positioned at coordinates (300, 300) at a certain altitude, and has N antennas. t =8, assuming the IRS is deployed on the top floor of the building, i.e., has the same height as the building. Specific simulation parameters involved include building density λ. b =100, 200, 300 / km 2 With an average building height of 50m and a path loss coefficient ρ of 10 at a reference distance d0 = 1m. -3 The path loss exponents for direct and IRS cascaded links are respectively set as γ. direct =4.7, γ IRS =2.2, the Rician channel model used for small-scale fading, with the Rician factor ε set to 3.

[0121] Appendix Figure 4 The gains of the IRS optimization algorithm relative to the two baselines mentioned above—random phase and no IRS—are presented under different total numbers of IRS reflector units. First, it is evident that the designed optimization algorithm shows significant gains in both user and rate performance compared to the random phase and no IRS cases, regardless of the number of IRS units. Simultaneously, it can be observed that the performance is slightly improved in the random phase case compared to the no IRS case. With random phase, system performance also increases with the increase in the total number of reflector units, but the rate of increase gradually decreases. This is because the increased number of reflector units increases the number of spatially independent sub-channels, improves spatial degrees of freedom, enhances channel diversity gain, and increases channel capacity. Furthermore, it is observed that under the design of the optimization algorithm, L=1 represents all M... all Each reflective element is concentrated on a single IRS, which is a centralized configuration; L = 5, 10 represents all M... all The reflection units are evenly distributed across 5 to 10 IRSs, constituting a distributed system. In certain scenarios, distributed IRSs offer superior performance compared to centralized systems. Furthermore, because distributed IRSs have a higher connection probability than centralized systems, they can serve more users. Therefore, the performance of a distributed system increases with the total number of reflection units M. all With the increase in the number of reflection units, the growth rate is faster. However, when the number of reflection units is large, centralized systems can no longer provide a comparable performance gain compared to distributed systems, and there are certain bottleneck limitations.

[0122] From the appendix Figure 5As can be seen, the performance of deployments near the base station deteriorates more significantly with increasing drone altitude, while the performance of deployments near the user side is less affected. This is because as drone altitude increases, the increased distance leads to increased path loss. Deployments near the drone base station significantly affect the drone-IRS distance due to drone altitude, while deployments near the user side, where there is already a certain distance between the UAV and IRS, are less affected by drone altitude because the increase in drone altitude is relatively small compared to the horizontal distance between the drone and the IRS. When drone altitude is limited to a low level, deployments near the drone base station offer superior performance, achieving optimal performance even at low drone altitudes. For distributed deployments near the user side, due to the greater distance from the drone, the IRS has a certain probability of obstruction. Performance initially increases and then decreases with increasing drone altitude, but more dispersed units are not necessarily better. Optimal distributed deployment parameter configurations exist for different building densities.

[0123] Appendix Figure 6 This reflects the effect of different building densities λ b The number of users that can be served by the system under different centralized distribution and IRS locations. As can be seen, with a fixed drone altitude, the more distributed the IRS, the more users can be served normally, and the fewer users are unserved. Furthermore, as the building density increases, the number of users that can be served increases more significantly with the distribution of the IRS, but the overall number of users that can be served decreases. This is because the denser the buildings, the greater the probability of obstruction and the lower the probability of user connection. To some extent, this provides guidance on IRS deployment methods under different obstruction scenarios.

Claims

1. A parameter configuration method for a 6G smart reflector-assisted UAV network, specifically, firstly, based on the location of the smart reflector IRS deployed on a building and having a unidirectional orientation characteristic and the user's location, considering the occlusion effect, quantifying the occlusion factor of the relevant links and the physical limitation of the IRS's unidirectional reflection, determining the feasible set of IRSs that the user can establish a communication link with the assistance of the IRS, wherein the IRSs in the feasible set satisfy the following conditions: the UAV and the user are on the same side of the IRS plane extension line, both the links from the IRS to the UAV and from the IRS to the user are not blocked, and the corresponding occlusion factor is zero; Under the aforementioned feasible set of IRSs, the user selects a set of nearby IRS services based on the corresponding path loss index; and obtains a set of channel state information under IRS assistance based on the connection matching situation. By substituting the system model into the optimal beamforming design, the optimized performance of the IRS-assisted UAV network in dense urban scenarios is obtained. Based on the performance under different parameter configurations, the optimal configuration for IRS and UAV collaboration under different occlusion scenarios is obtained to optimize and improve system efficiency.

2. The parameter configuration method for a 6G intelligent reflector-assisted UAV network according to claim 1, characterized in that, Considering the unidirectional orientation characteristics of the IRS, occlusion is quantified. 3D geometric relationships are used to determine whether each link will be blocked by buildings, and an occlusion factor w is introduced to characterize this. Furthermore, for a user to successfully navigate via the IRS, the following two conditions must be met simultaneously: the drone and the user must be on the same side of the extended plane of the IRS, and cosη... uav,irs cosη irs,user >0, and both the IRS-to-drone and IRS-to-user links are unblocked, meaning the corresponding occlusion factor is zero. Here, cosη uav,irs Cosη represents the cosine of the angle between the vector from IRS to the UAV in the projection direction and the perpendicular vector to the longer side of IRS. irs,user This represents the cosine of the angle between the vector from the IRS to the user and the perpendicular vector to the longer side of the IRS.

3. The parameter configuration method for a 6G intelligent reflector-assisted UAV network according to claim 1, characterized in that, Users select the IRS service set from the feasible set. That is, the set of users that can be served by IRS l, l = 1, ..., L can be obtained first by the IRS occlusion judgment criterion. Where K represents the total number of users, and at the same time, for user k, there exists a feasible set of available IRSs. Among them, w uav,l and w l,k These represent the link blockage status from the drone to IRS1 and from IRS1 to user k, respectively. A value of 0 indicates no blockage. Each user obtains the reference signal under the corresponding feasible IRS set, which is represented in the model as the path loss PL of the corresponding IRS cascaded link. l,k Based on path loss (i.e., signal strength), select L0 IRSs with better channel conditions and stronger signals to determine the set of IRSs serving user k. And update the user set of the IRS service accordingly.

4. The parameter configuration method for a 6G intelligent reflector-assisted UAV network according to claim 1, characterized in that, Based on the connection matching of each network element, the channel state information set under IRS assistance is obtained. That is, considering the obstruction effect, the actual direct connection channel state information between the UAV and user k is represented as follows: in H represents the set of unobstructed users that can be served by the drone. k The channel state information between the UAV and user k, further used to obtain the channel state information (CSI) of each user under multi-IRS cooperative assistance, can be expressed in the following form. This is an equivalent combination channel of the UAV-user direct connection channel and the IRS concatenated channel, where G l This represents the channel state information between the UAV and the l-th IRS. Φ represents the channel state information between the l-th IRS and user k. l Let l be the phase matrix of the l-th IRS reflective unit. For the set of IRS serving user k.

5. The parameter configuration method for a 6G intelligent reflector-assisted UAV network according to claim 1, characterized in that, The optimal beamforming of the IRS-assisted UAV network in dense urban scenarios was jointly designed to optimize its performance. Specifically, based on the limited CSI information under occlusion conditions, the beamforming algorithm was used to iteratively optimize the active beamforming of the UAV base station and the passive phase of the IRS under the condition of uneven service sets, so as to obtain the system and rate under occlusion scenarios.

6. The parameter configuration method for a 6G intelligent reflector-assisted UAV network according to claim 5, characterized in that, When performing passive phase optimization of an IRS, under the condition of a finite service set, the optimization problem is conceived as follows: in, q l =[[Q l ] 1,1 ,...,[Q l ] M,M ] T , Furthermore, due to the occlusion effect and the user's choice of IRS, there is a different set of IRS services for each user, and each IRS also has its own set of users. Special definition Where ω k The weight coefficients corresponding to user k. W represents the channel state information between the l-th IRS and user k. k It is the auxiliary weight matrix associated with user k, U k For user k, a linear receiver filter, V k To solve the optimization problem for the corresponding beamforming matrix, the conjugate gradient descent method of manifold optimization is used.

7. The parameter configuration method for a 6G intelligent reflector-assisted UAV network according to claim 1, characterized in that, Based on the performance under different parameter configurations, the optimal configuration for IRS and drone collaboration under different occlusion scenarios is obtained. That is, based on the performance of optimization as the configurable parameters change, the optimal parameter configuration of drone and IRS is obtained, including drone altitude, IRS deployment location bias, and IRS size and number.

Images