A drone-assisted, self-free, large-scale MIMO system resource optimization method.
The method optimizes resource allocation in drone-assisted MIMO systems by integrating ground and aerial APs, addressing switching and energy issues, thereby improving coverage and user experience.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- NANJING UNIV OF POSTS & TELECOMM
- Filing Date
- 2023-11-23
- Publication Date
- 2026-04-23
AI Technical Summary
Existing drone-assisted, self-free large-scale MIMO systems face challenges in resource allocation, improper UAV scheduling leading to communication switching, energy consumption, and diverse user QoE needs, with conventional methods failing to optimize system resources and user experience.
A method involving user scheduling, drone positioning, and power distribution schemes, utilizing a collaborative service communication model between ground and aerial APs, optimized through iterative convex optimization techniques to ensure fair resource allocation and minimize energy consumption.
Improves wireless coverage, prevents arbitrary user connection switching, and ensures fair resource allocation, enhancing system stability and user communication performance.
Smart Images
Figure 2026513226000001_ABST
Abstract
Description
[Technical Field]
[0001] This invention belongs to the field of communication resource management, and more particularly to a method for optimizing resources in a drone-assisted, self-free, large-scale MIMO system. [Background technology]
[0002] Cell-free large-scale MIMO (Massive Multi-Input, Multi-Output) is expected to outperform 5G (B5G) wireless technology due to its higher spectral efficiency, higher energy efficiency, and better spatial diversity. However, the large amount of long-distance cabling (front-end transport requirements) between each access point (AP) and central processing unit (CPU) hinders its practical application. In recent years, UAVs have been increasingly considered as an aerial communications support tool due to their absolute advantages such as rapid deployment, controllable mobility, low cost, and high probability of achieving line-of-sight air-to-ground links. They are often used for rapid service restoration after some or all of infrastructure is damaged by natural disasters and in areas with congested communications. Therefore, adding drones as aerial APs can effectively mitigate the challenges posed by the large amount of long-distance cabling in cell-free large-scale MIMO systems. While UAVs can overcome the limitations of conventional cell-free large-scale MIMO, several remaining issues still need to be addressed.
[0003] One of the most critical issues in self-free large-scale MIMO systems that work with UAVs is resource allocation. First, since ground access points (APs) can already provide high-quality service to nearby users, improper UAV scheduling can cause communication switching among nearby users, increasing the system load and degrading the quality of service for users. Second, the performance and execution time of UAV systems are fundamentally limited by the limits of onboard power, and if more appropriate stopping points are not designed for UAVs, the UAVs will consume a large amount of energy through maneuvering, reducing their communication efficiency. Furthermore, if radio resources are insufficient, deploying UAVs to cover multiple furthest-end users may not meet the quality of service (QoE) requirements of some users. While it is easy to design multiple drones to meet the QoE of users, this could result in enormous resource consumption for the system as a whole. Finally, because users' QoE needs are diverse and randomly distributed, achieving rational resource scheduling and allocation is extremely challenging for the CPU when aiming to improve the user experience while ensuring fair performance among users. [Overview of the project] [Problems that the invention aims to solve]
[0004] The objective of the present invention is to provide a resource optimization method for a drone-assisted, self-free, large-scale MIMO system that satisfies the constraints of system equipment and the QoE needs of all users, namely, a method that allows the CPU to rationally allocate resources while considering system resource consumption, while enabling optimal UAV placement and user scheduling according to user requests. [Means for solving the problem]
[0005] In the resource optimization method for a drone-assisted self-free large-scale MIMO system according to the present invention, a drone is positioned as an aerial AP in the target area. Step 1 involves designing a user scheduling scheme in which ground and airborne APs work together to provide services to users in signal coverage dead zones by deploying drones as aerial APs in the target area. Step 2: Construct a collaborative service communication model based on a user scheduling scheme in which ground APs and air APs work together to provide services. Step 3 involves designing a resource allocation scheme that includes user scheduling, drone positioning, and power distribution to users, and maximizing the minimum downlink speed for users in the above-mentioned collaborative service communication model, and Step 4 involves verifying the feasibility of the collaborative service communication model using an optimization solution method, thereby constructing the collaborative service communication model and optimizing the configuration of communication resources based on the model.
[0006] Furthermore, step 1 specifically means, The AP selection method involves each user connecting to multiple APs within the service area at once, and using a binary variable {χ}. mk , χ uk The} table shows the scheduling status of the user's ground APs and air APs, respectively, and the user scheduling matrix χ∈￡ K×(U+1) Define a binary variable χ when user K is served by a ground AP point. mk The coefficient of is 1, and 0 otherwise, and the binary variable is χ when the user is served by an aerial AP point. uk This means that the condition is 1 if it is not 1, and 0 otherwise. Based on the heterogeneity of resources and the constraints of drone placement, it is taken into consideration to implement a cooperative access rule between airborne and ground AP points. This cooperative access rule stipulates that, during each resource allocation period, each user communicates with a single airborne or ground service point, the ground AP provides service to all users in the target area, and the airborne AP provides service to users within the radio coverage range of the target area. These are binary constraints.
number
[0007] Furthermore, step 2 specifically designs a channel model that combines the effects of small-scale fading and large-scale fading. Assuming that g mk ∈ £ represents the channel gain between the m-th AP and the k-th user, the channel transmission model is
Equation
Equation
number
number
number
number
number
number
number
number
number
number
[0008] Furthermore, step 3 specifically means, Based on a user scheduling scheme in which ground APs and air APs work together to provide service, in a scenario where both ground APs and air APs share channels, ground users are served by only one type of AP at a time, and user scheduling constraints apply.
number
number
number
[0009] Furthermore, step 4 specifically means, The above optimization problem (8) is a mixed-integer non-convex problem, and step 4-1 involves iteratively solving it by converting the non-convex problem into a convex problem using block coordinate descent and continuous convex approximation techniques, First, the binary variable of user scheduling is adjusted to a continuous variable between 0 and 1. Based on the optimization variables χ, q, and P(p,ρ) included in the problem, the block coordinate descent method is employed to decompose the new non-convex problem into three subproblems: user scheduling (χ), aerial AP placement (q), and power distribution (P(p,ρ)). Step 4-2 involves optimizing the user scheduling for the given aerial AP placement and power distribution based on continuous convex optimization techniques, optimizing the aerial AP placement for the given user scheduling and power distribution, and optimizing the complex power distribution for the given user scheduling and aerial AP placement by combining variable block decomposition. The problem of maximizing the minimum speed for users
number
number
number
number
number
number
number
number
number
number
number
number
number
number
[0010] Compared to the prior art, the present invention has the following remarkable advantages.
[0011] Firstly, this method improves the wireless coverage performance of the system by adding UAVs at aerial AP points to conventional cell-free large-scale MIMO. Secondly, this method allows for the rational configuration of UAV deployment and prevents arbitrary switching of user connections and wasted system resources. Thirdly, this method ensures the rationality and stability of the system by designing a fair resource allocation scheme that includes a user scheduling scheme, a drone positioning scheme, and a power distribution scheme to users, and iteratively determining the allocation between each scheme. [Brief explanation of the drawing]
[0012] [Figure 1] Figure 1 is a flowchart illustrating a resource optimization method for a drone-assisted, self-free, large-scale MIMO system. [Figure 2] Figure 2 shows a system model of a resource optimization method for a drone-assisted, self-free, large-scale MIMO system. [Modes for carrying out the invention]
[0013] The technical solutions of the present invention will be further described below with reference to the drawings.
[0014] A method for optimizing the resources of a drone-assisted, self-free, large-scale MIMO system, comprising the following steps as shown in Figure 1.
[0015] Since long-distance cables are used to connect each ground access point (AP) to the CPU, the limitations of cable length prevent users at long distances from accessing the APs. While easily deployable UAVs can act as airborne APs to provide service to users in coverage blind spots, adding UAVs increases the load on system resources to some extent. Furthermore, improper UAV placement can interfere with other ground users, affecting the overall system stability. Therefore, in the following sections, a user scheduling scheme and collaborative communication model will be designed from the perspective of designing highly efficient and equitable resource allocation, thereby ensuring that edge users can connect and communicate normally.
[0016] Step 101: Design user scheduling rules that enable ground access points (APs) and airborne APs to work together to provide services.
[0017] Step 101-1 AP selection method: Each user connects to multiple APs within the service area at one time, and the binary variable {χ mk , χ uk The} indicates the scheduling status of each user, and the user scheduling matrix χ∈￡ K×(U+1) Define a binary variable χ when user K is served by a ground AP point. mk The coefficient is 1 if it is not 1, and 0 otherwise, that is,
number
number
[0018] Step 101-2 In order to achieve coordinated operation between airborne AP points and ground AP points, the heterogeneity of resources and constraints on drone placement are further considered, and access rules are taken into account. These access rules stipulate that, during each resource allocation period, 1) each user can communicate with only one airborne or ground service point at a time, 2) ground APs can provide service to all users within a limited range, and 3) airborne APs can provide service to users within their radio coverage range. These rules are binary constraints.
number
[0019] Step 102: The large-scale fading coefficient in ground communications is mainly affected by path loss and shadow fading. Communication between UAVs and ground users is generally line-of-sight propagation, and large-scale fading is related to the altitude of the UAV and the angle it makes with the ground user. According to the user scheduling rule described above, the user's equipment receives only one type of service at a time. Therefore, we design the communication model for collaborative services affected by the user scheduling rule below.
[0020] Step 102-1 The channel model combines the effects of small-scale and large-scale fading, g mk If ∈£ represents the channel gain between the mth AP and the kth user, then the channel transmission model is
number
number
[0021] Step 102-2 The communication link between the aerial AP point and the user can obtain channel gain in a similar manner, i.e.,
number
number
number
number
number
number
[0022] Because the timescale of airborne AP deployment is far greater than the channel coherence time, leakage interference, network metrics, and actual constraints should all be derived on a large scale, i.e., E{g uk}=β uk That is the case.
[0023] Step 102-3 Assuming all channel conditions are known, the downlink transmits the signal to the user using conjugate beamforming. Therefore, the received signal for user k, considering user scheduling, is
number
number
[0024] Assume that the aerial AP provides service only to users within its communication coverage range and does not interfere with other user signals. If the aerial AP is hovering in a fixed position, the signals received by users within its communication range will be:
number
[0025] The total number of signals received by the user is:
number
[0026] Step 103: Aerial AP scheduling primarily serves users at long distances, and drone deployment should be conducted among multiple users at long distances, while simultaneously ensuring that service resources are not diverted by users at closer distances. Furthermore, drones should maintain a safe distance from each other to avoid collisions. In such cases, a fair resource allocation scheme must be designed to maximize user coverage using limited resources.
[0027] Step 103-1 If it is assumed that the terrestrial APs share channels with each other, the aerial APs share channels with each other, and there is no interference between the two, then terrestrial users are served by only one type of AP at a time. Therefore, user scheduling should
Number
[0028] Step 103-2 The placement of aerial APs is restricted to a certain extent by the geographical environment. A plurality of aerial APs maintain a safe physical distance ||q j -q u ||≤S af (u≠j). The system resources are limited, the transmission power of the AP cannot exceed its own maximum transmission power,
Number
[0029] Step 103-3 By maximizing the downlink communication speed of users in the above communication model and combining user scheduling constraints, drone placement constraints, and power distribution constraints, an optimization problem of resource allocation for a conventional cell-free MIMO system based on drone assistance is constructed to achieve coverage without dead spots for users in the system. The optimization problem is
Number
[0030] Step 104 Verify the feasibility of the proposed drone assistance model by an optimal solution method.
[0031] Step 104-1 Given that the user scheduling and associated optimization variables are binary variables, and that there are non-convex constraints on the drone placement variable q and the transmit power variable, the optimization problem is a mixed-integer non-convex problem. The non-convex problem is transformed into a convex problem by employing block coordinate descent and continuous convex approximation techniques and solved iteratively. Specifically, based on continuous convex optimization techniques, the user scheduling is optimized for a given aerial AP placement and power distribution, the aerial AP placement is optimized for a given user scheduling and power distribution, and a complex power distribution is optimized for a given user scheduling and aerial AP placement by combining this with variable block decomposition.
[0032] Step 104-2 To make the problem easier to handle, first, the binary variable of user scheduling is relaxed to a continuous variable between 0 and 1, and then the block coordinate descent method is employed to decompose the new non-convex problem into three sub-problems: user scheduling, aerial AP placement, and power distribution.
[0033] Step 104-3 Maximizing the minimum speed for the user
number
number
number
number
Number
[0034] Step 104-4 For the given χ, p, ρ, and any given local point q r the approximate problem of the original aerial AP placement problem is
Number
Number
Number
Number
Number
Number
[0035] Step 104-5 Since the power optimization problem for the ground AP is relatively complex, a variable block p m (m = 1…M) indicating the transmission power at the m-th AP point is introduced, and p -m represents the set of other variable blocks excluding the variable block at m. For the given χ and q, the approximate convex problem of the original power optimization problem is It should be noted that there are some unclear notations like "<图注: " in the original text which might need further clarification for a more accurate translation.
number
number
number
[0036] Step 104-5 In a typical block coordinate descent method, the sub-problem of updating each variable block must be solved optimally and accurately in each iteration, and user scheduling, placement optimization, and transmit power optimization are all lower bounds of the overall constraint function, i.e.,
number
Claims
1. A method for optimizing the resources of a drone-assisted, self-free, large-scale MIMO system, Drones are deployed as aerial APs in the target area. Step 1 involves designing a user scheduling scheme in which ground and air APs work together to provide services to users in signal coverage dead zones by deploying drones as aerial APs in the target area. Step 2: Construct a collaborative service communication model based on a user scheduling scheme in which ground APs and air APs work together to provide services. Step 3 involves designing a resource allocation scheme that includes user scheduling, drone positioning, and power distribution to users, and maximizing the minimum downlink speed for users in the above-mentioned collaborative service communication model, and A method for optimizing resources for a drone-assisted, self-free, large-scale MIMO system, characterized by performing step 4, which involves verifying the feasibility of the collaborative service communication model using an optimization solution method, thereby constructing a collaborative service communication model, and optimizing the configuration of communication resources based on the model.
2. Step 1, specifically, The AP selection method involves the user connecting to multiple APs within the service area at once, and a binary variable {χ}. ｍｋ , χ ｕｋ The} table shows the scheduling status of user k's ground AP and air AP, respectively, and the user scheduling matrix χ∈￡ Ｋ×（Ｕ＋１） Define the binary variable χ when user K is served by a ground AP point. ｍｋ The coefficient of is 1, and 0 otherwise, and the binary variable is χ when the user is served by an aerial AP point. ｕｋ This means that it is 1 if the condition is not met, and 0 if it is not met. Based on the heterogeneity of resources and the constraints of drone deployment, it is taken into consideration to implement a cooperative access rule between airborne and ground AP points. The cooperative access rule is that, during each resource allocation period, each user communicates with a single airborne or ground service point, the ground AP provides service to all users in the target area, and the airborne AP provides service to users within the radio coverage range of the target area. These are binary constraints. [Number 63] It was shown as, Here, w ｋ = [x ｋ , y ｋ Ｔ ∈ i ２×１ represents the horizontal coordinate of the k-th user, w ａ represents the horizontal coordinate of the terrestrial AP, q ｕ = [x ｕ , y ｕ Ｔ ∈ i ２×１ represents the horizontal position of the drone u, R ａ and R ｕ respectively represent the service ranges of the terrestrial AP and the aerial AP, R ａ is affected by the cable length, and R ｕ is determined by the flight altitude and remaining power of the drone, The resource optimization method of the cell-free massive MIMO system based on drone assistance according to claim 1, characterized in that it includes the above.
3. Step 2, specifically, We designed a channel model that combines the effects of small-scale and large-scale fading, g ｍｋ If ∈£ represents the channel gain between the m-th AP and the k-th user, then the channel transmission model is [Number 64] And here, β ｍｋ and h ｍｋ These figures represent the large-scale fading coefficient and small-scale fading factor between the m-th AP and the k-th user, respectively, where the small-scale fading is Rayleigh fading, i.e., h ｍｋ ~CN(0,1), and the large-scale fading coefficient β of the ground AP point. ｍｋ This is modeled as the product of path loss and shadow fading, i.e., β ｍｋ =pl ｍｋ ・s ｍｋ And here, s ｍｋ This demonstrates log-normal shadowing, and pl ｍｋ This is a three-stage path loss, 【Number 65】 It was shown as, Here, L is a constant determined by the carrier frequency, user and AP altitude, and d ｍｋ d is the horizontal distance between AP m and user k, and d ０ and d １ Step 2-1 is the reference distance, The communication link between the aerial AP point and the user is [Number 66] It was shown as, Here, a １ and a ２ These are the path loss indices for the line-of-sight line (LoS) and the non-line-of-sight line (NLoS), respectively. [Number 67] These are the probabilities of LoS links and NLoS links, respectively. [Number 68] When designing a communication system based on an aerial AP point, where is the Euclidean distance between the u-th drone and the k-th user, and H is the constant flight altitude of the drone, considering the randomness of the LoS and NLoS links, the LoS transmission probability between the u-th aerial AP and the k-th user is: [Number 69] And, Here, a and b are environmental parameters, and U is the set of aerial APs. [Number 70] is the elevation angle between the k-th user and the u-th aerial AP, and the LoS probability model increases with increasing altitude and elevation angle of the AP, and the average large-scale fading between the k-th user and the u-th aerial AP is [Number 71] It is shown as follows, and when the channel model is derived in a large form, E{g ｕｋ } = β ｕｋ Step 2-2 is as follows: In a scenario where all channel conditions are known, the downlink transmits the signal to the user using conjugate beamforming, and the received signal to user k, considering user scheduling, is [Number 72] And here, x ｉｋ This indicates a signal transmitted from the AP, and if the user is being served by a ground AP, [Number 73] And so, Here, p ｍｋ is the power distribution parameter of the ground AP in a cell-free large-scale MIMO downlink, and s ｋ The signal that is transmitted to the user, E{|s ｋ | ２ } satisfies 1, ω ｋ This is additive white Gaussian noise, In a scenario where an aerial AP provides service only to users within a certain range, if the aerial AP is hovering in a fixed position, the signals received by users within its communication range will be: [Number 74] And, Here ρ ｕｋ These are the power distribution parameters, The total number of signals received by the user [Number 75] A method for optimizing the resources of a drone-assisted, self-free, large-scale MIMO system according to claim 1, comprising steps 2-3 shown as follows:
4. Step 3, specifically, Based on a user scheduling scheme in which ground APs and air APs cooperate to provide services, in a scenario where both ground APs and air APs share channels, ground users are served by only one type of AP at a time, and user scheduling constraints apply. [Number 75] Step 3-1 that satisfies the requirements, The AP's transmit power is, [Number 76] Constrained by, Here, P ｍａｘ and ρ ｍａｘ Step 3-2, where the maximum transmit power of the AP and the drone are respectively, In the collaborative service communication model, we maximize the minimum downlink communication speed for users and, by combining user scheduling constraints, drone placement constraints, and power allocation constraints, we construct a resource allocation optimization problem for conventional self-free MIMO systems based on drone assistance, achieving blind spot coverage for users within the system. The optimization problem is: [Number 77] And, Here, in equation (8), (1) and (2) are constraints on user scheduling, constraint (3) is a restriction on the drone's placement, and (4) and (5) are constraints on power distribution, S ａｆ A method for optimizing the resources of a drone-assisted, self-free, large-scale MIMO system according to claim 1, comprising step 3-3, wherein is the maximum safe distance between multiple drones.
5. Step 4, specifically, The above optimization problem (8) is a mixed-integer non-convex problem, and step 4-1 involves iteratively solving it by converting the non-convex problem into a convex problem using block coordinate descent and continuous convex approximation techniques, First, the binary variable of user scheduling is adjusted to a continuous variable between 0 and 1. Based on the optimization variables χ, q, and P(p,ρ) included in the problem, the block coordinate descent method is adopted to decompose the new non-convex problem into three subproblems: user scheduling (χ), aerial AP arrangement (q), and power distribution P(p,ρ). Then, based on continuous convex optimization techniques, the user scheduling is optimized for the given aerial AP arrangement and power distribution, the aerial AP arrangement is optimized for the given user scheduling and power distribution, and by combining this with variable block decomposition, the complex power distribution is optimized for the given user scheduling and aerial AP arrangement (step 4-2). The problem of maximizing the minimum speed for users [Number 78] Expressed as a function, given the drone placement and power distribution, the convex optimization problem for user scheduling is: [Number 79] And, Here, [Number 80] The upper bound of the first-order Taylor expansion obtained by expanding the interference term, where in each iteration the original function is approximated by a function that is easily processed at a given local point, λ ｒ This is the target result in the r-th iteration, and χ ｒ If it is defined as the user result obtained in the r-th iteration, [Number 81] And so, Given q, p, and ρ, the problem is a convex problem, and step 4-3, which can be solved with a solver, Given χ, p, ρ and any given local point q ｒ In contrast, the convex optimization problem for the placement of airborne APs is: [Number 82] And, Here, [Number 83] And, [Number 84] R ｕ This is the value after scaling, [Number 85] Each [Number 86] These are the linear coefficients and constants of the equation, [Number 87] Step 4-4 is the result of the placement of airborne APs obtained in the r-th iteration, Variable block p to show the transmission power at point m of AP ｍ (m = 1...M) is introduced, p －ｍ Here we show the set of other variable blocks excluding the variable block at m, and for given χ and q, the convex optimization problem of power distribution is: [Number 88] And, Here, [Number 89] Each [Number 90] Steps 4-5, which are the upper bound of the first Taylor expansion, If each of the feasible sets in each subproblem is a subset of the feasible set, [Number 91] And so, A method for optimizing the resources of a drone-assisted, self-free, large-scale MIMO system according to claim 4, characterized in that the solution procedure converges, and improvements in system coverage performance and user communication performance are achieved through user scheduling, drone placement, and power distribution schemes.