A joint optimization method for low-power beam routing and operating modes in multi-STAR-RIS assisted communication

Abstract

This invention discloses a joint optimization method for low-power beam routing and operating modes in multi-STAR-RIS assisted communication, belonging to the field of 6G wireless communication technology. The invention first constructs a total power consumption model including base station transmit power and STAR-RIS hardware power consumption, and then builds a MINLP optimization problem with user QoS rate as a constraint. An alternating optimization framework is used to decouple continuous and discrete variables: the continuous subproblem utilizes block coordinate descent and semi-definite relaxation optimization to cascade phase shift, combined with the MRT strategy to calculate the minimum beam power; the discrete subproblem constructs a directed graph of the line-of-sight channel, uses K-shortest path to generate candidate routes, and adaptively selects the MS / ES low-power operating mode through ILP, iterating alternately until power convergence. This invention switches hardware operating modes according to the bidirectional transmission requirements of the path, balancing channel gain and hardware overhead, reducing the total system power consumption in multi-hop obstruction scenarios, and achieving 360° full-space coverage.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, specifically to a resource scheduling and routing method based on intelligent metasurface (RIS) technology for 6G mobile communication networks, and particularly to a low-power beam routing and joint optimization method for communication assisted by multi-simultaneous reflection and transmission intelligent surfaces (STAR-RIS). Background Technology

[0002] With the widespread deployment of fifth-generation (5G) mobile communication networks and the evolution towards sixth-generation (6G) communication, wireless communication networks face severe challenges in achieving full coverage, ultra-high spectrum efficiency, and green and low-carbon communication. Due to the short wavelength of high-frequency signals such as millimeter waves, they are easily blocked by obstacles such as buildings and vegetation, making it difficult for traditional single-hop communication to guarantee quality of service (QoS) in complex environments.

[0003] In recent years, Reconfigurable Intelligent Surfaces (RIS) have attracted widespread attention as an emerging technology capable of intelligently reconfiguring the wireless propagation environment. By establishing multi-hop cascaded links based on multiple RIS between base stations and users, severe signal obstruction problems can be effectively solved and coverage expanded. In this multi-RIS cascaded network, the signal originates from the base station, undergoes precise phase deflection and relay forwarding by multiple RIS nodes, and is finally directionally transmitted to the user end. This process of finding and establishing the optimal transmission path for the signal beam is called "beam routing." However, traditional RIS only possesses reflection capabilities, and their service area is limited to the "half-space" on the reflecting side. This results in significant physical constraints on the deployment location of equipment and the selection of signal routing paths in multi-hop beam routing, lacking flexibility.

[0004] To overcome the physical limitations of traditional RIS (Resonance Information System), Simultaneous Transmitting and Reflecting (STAR) RIS was developed. Unlike traditional RIS, STAR-RIS allows signals to propagate simultaneously through reflection and transmission modes, achieving 360-degree "full-space" signal coverage. This characteristic gives STAR-RIS a significant advantage in multi-hop communication, enabling it to bypass obstacles with fewer hops or more flexible paths.

[0005] Despite the significant potential of STAR-RIS in enhancing coverage and improving speed, existing technical solutions for beam routing design and algorithm research in multi-STAR-RIS assisted networks still have the following shortcomings: 1. Routing objectives focus on spectral efficiency while neglecting energy efficiency: Most existing beam routing studies focus on maximizing spectral efficiency or speed, often ignoring the energy efficiency of the system.

[0006] 2. Routing decisions ignore differences in hardware power consumption: In fact, while the introduction of STAR-RIS brings performance gains, it also introduces additional hardware power consumption. Specifically, STAR-RIS has different operating modes, such as Energy Splitting (ES) mode and Mode Switching (MS) mode. In ES mode, components must handle both reflected and transmitted signals simultaneously, resulting in higher circuit power consumption; while in MS mode, components only perform reflection or transmission, resulting in relatively lower power consumption. However, most existing routing decision-making methods focus on link performance optimization and do not take into account the differences in hardware power consumption under different STAR-RIS operating modes.

[0007] 3. Lack of a mathematical model for joint route and mode selection: In multi-hop routing scenarios, how to dynamically select the STAR-RIS operating mode (such as ES and MS switching) based on user distribution to achieve a balance between link gain and hardware power consumption remains an unresolved problem.

[0008] 4. Existing routing algorithms lack a low-power routing optimization framework that takes into account both physical channel and topological constraints: Existing literature attempts to use heuristic algorithms such as deep reinforcement learning to solve the routing energy efficiency problem of multi-hop STAR-RIS. However, due to the difficulty in establishing a unified mathematical model that simultaneously characterizes the physical channel features and the graph theory network topology, such methods often struggle to balance solution stability and optimization performance when faced with highly coupled complex constraints such as beam number allocation and node in-degree and out-degree restrictions.

[0009] Therefore, there is an urgent need for a method that can jointly optimize beam routing, STAR-RIS operating modes, phase shift matrix, and power allocation, namely, a low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication. This method aims to minimize the total system power consumption, including base station transmit power and STAR-RIS hardware power consumption, while meeting user QoS requirements, thereby achieving green and efficient communication. Summary of the Invention

[0010] To address the aforementioned issues, this invention proposes a low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication. This method incorporates the power consumption differences of different STAR-RIS operating modes into the optimization objective. By decoupling discrete and continuous coupled variables through an alternating optimization framework, it achieves joint optimization of beam routing, operating modes, phase shift matrix, and power allocation, thereby minimizing the total system power consumption while ensuring user QoS.

[0011] To achieve the above objectives, the technical solution of the present invention is implemented as follows: A low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication includes the following steps: S1. Construct a multi-STAR-RIS assisted downlink transmission communication system model and establish a total system power consumption model that includes base station transmit power and STAR-RIS hardware power consumption; S2. Based on the multi-STAR-RIS assisted downlink transmission communication system model, with the constraint of meeting the user's quality of service (QoS) requirements and the objective of minimizing the total system power consumption, a mixed integer nonlinear programming joint optimization problem is constructed. The joint optimization variables include beam routing, STAR-RIS operating mode, reflection and transmission phase shift matrix, number of base station beams, and beam power allocation parameters. S3. Solve the joint optimization problem using an alternating optimization framework: Decouple the original problem into a continuous variable optimization subproblem and a discrete variable optimization subproblem, and iterate the two subproblems alternately until the total power consumption of the system converges, and output the optimal beam routing, STAR-RIS working mode, phase shift matrix, number of base station beams and beam power allocation parameters.

[0012] Furthermore, the multi-STAR-RIS assisted downlink transmission communication system model in step S1 specifically includes: A configuration The base station uses a single transmitting antenna and employs multi-beam transmission technology to serve users; A fixed-position, simultaneously reflective and transmissive smart metasurface STAR-RIS is deployed between the base station and the user. Each STAR-RIS consists of N reconfigurable passive units, each of which can simultaneously reflect and transmit signals to build multi-hop cascaded links for auxiliary communication. In a user cluster consisting of multiple single-antenna users, where the direct link between the base station and the user is blocked, each user receives signals through only one multi-hop STAR-RIS path. The STAR-RIS operating modes include mode switching (MS) mode and energy splitting (ES) mode. The operating mode selection is represented by binary variables. In MS mode, the node operates only in total reflection or total transmission state, while in ES mode, the node performs both reflection and transmission.

[0013] Furthermore, the system total power consumption model in step S1 is specifically as follows: The total system power consumption consists of the sum of the base station's total transmit power and the static power consumption of all STAR-RIS hardware, expressed as:

[0014] in This represents the total transmit power of the base station. For the first The hardware power consumption of each STAR-RIS depends on its operating mode. The static power consumption of a single STAR-RIS in different operating modes is as follows: MS mode:

[0015] In ES mode:

[0016] In the formula, This is the base static power consumption of the STAR-RIS control module. The driving power consumption of a single passive unit in MS mode. This represents the drive power consumption of a single passive unit in ES mode, and > .

[0017] Furthermore, the constraints of the mixed-integer nonlinear programming joint optimization problem constructed in step S2 include: Constraints include: user reception rate not lower than the minimum QoS requirement; amplitude and phase constraints of the STAR-RIS unit; binary variable constraints of the operating mode; base station beam count constraints; maximum base station transmit power constraints; and non-negative beam power constraints.

[0018] Furthermore, the solution process for the alternating optimization framework in step S3 is as follows: S31. Variable decoupling: Decompose the original problem into a continuous variable subproblem of phase shift and beam power optimization, and a discrete variable subproblem of beam routing and mode decision. S32. Continuous variable optimization: With the current user beam routing, STAR-RIS operating mode and number of base station beams fixed, solve the continuous variable subproblem to obtain the optimal cascaded phase shift matrix and the corresponding beam power allocation result; S33. Discrete variable optimization: Fix the current optimal phase shift matrix, solve the discrete variable subproblem, and re-determine the user beam routing, STAR-RIS working mode and the number of base station beams; S34. Alternating Iteration: Repeatedly execute steps S32 and S33, using the optimal solution of the previous subproblem as the input of the next subproblem, until the difference in the total power consumption of the system between two adjacent iterations is less than the preset convergence threshold, or the maximum number of iterations is reached, and output the globally optimal configuration parameters.

[0019] Furthermore, in step S32, the optimization of the phase shift matrix is ​​specifically achieved in the following way: A block coordinate descent algorithm is used to decouple the phase shift of multi-hop cascaded nodes. In each iteration, only the phase shift of a single STAR-RIS node is optimized, while the phase shift parameters of all other nodes are fixed. For a single STAR-RIS node to be optimized, the users passing through the node are divided into a reflection user set and a transmission user set. A positive semidefinite matrix is ​​introduced to relax the original non-convex phase shift optimization problem into a positive semidefinite programming problem and solve it. Then, the optimal phase shift vector with rank 1 that satisfies the modulus constraint is recovered from the relaxed solution through the Gaussian randomization method, thereby maximizing the end-to-end equivalent channel gain. All STAR-RIS nodes are polled sequentially and iterated until the total channel gain of the system converges.

[0020] Furthermore, in step S32, the beam power allocation is calculated as follows: Based on the determined phase shift matrix, the base station adopts the maximum ratio transmission strategy to align the direction of the transmit beam with the direction of the user's equivalent channel. At this time, the end-to-end equivalent channel gain of the user is a constant. The signal-to-noise ratio (SNR) threshold is derived based on the user's minimum data rate requirement. Combined with the equivalent channel gain, the minimum beam transmit power required to meet QoS requirements is calculated. The expression is:

[0021] in, The set of users serving the q-th beam. Let u be the end-to-end equivalent channel gain for user u. To meet the minimum signal-to-noise ratio threshold required for the lowest data rate, the power of each beam is taken as the maximum power required to serve all users.

[0022] Furthermore, in step S33, candidate path screening is performed before discrete variable optimization, specifically including: Construct a directed graph of the line-of-sight channel, with the node set including the base station, all STAR-RIS nodes and all user nodes, and the directed edges consisting of non-zero channel links that satisfy the unidirectional transmission geometric constraints; Define the edge weight for each directed edge. The edge weight is taken as the negative logarithm of the squared modulus of the physical channel gain of the corresponding single-hop link, and its expression is:

[0023] in, Let be the physical channel matrix pointing from node i to node j. Denotes the Frobenius norm of a matrix; Minimizing the total path weight is equivalent to maximizing the end-to-end cascaded channel gain of the path; The K-shortest path algorithm is used to search for the top L paths with the smallest weights for each user on the directed graph, forming a candidate path set for that user.

[0024] Furthermore, after generating the candidate path set, the discrete variable optimization subproblem also completes the optimal solution through the following steps: Introducing three types of binary variables—user path selection, STAR-RIS activation, and base station beam triggering—along with the continuous variable of beam transmit power, an integer linear programming model is constructed with the objective function of minimizing the total system power consumption.

[0025] in Let j be the beam transmit power directed at the j-th STAR-RIS. As the activation variable for STAR-RIS, Choose a variable for the working mode. For beam triggering variables, This is the beam count penalty coefficient; Model constraints include user single-path constraints, beam triggering constraints, node activation constraints, reflection and transmission side capacity constraints, ES mode triggering constraints, user QoS power constraints, and variable domain constraints. The branch-and-bound method is used to solve the integer linear programming model to obtain the optimal combination of user routing, STAR-RIS operating mode and base station beam allocation.

[0026] Furthermore, the ES mode triggering constraint is specifically as follows: The high-power ES mode is triggered if and only if both a reflection path and a transmission path exist at the STAR-RIS node. When a node has only a reflection path or only a transmission path, it operates in a low-power MS mode, reducing system hardware power consumption through adaptive mode switching.

[0027] Beneficial effects: 1. Significantly improves system energy efficiency: This invention breaks through the limitations of traditional methods that only focus on spectral efficiency, incorporating hardware power consumption into the optimization objective. By establishing an accurate STAR-RIS power consumption model, the low-power MS mode is prioritized in routing decisions, and the high-power ES mode is only activated when necessary (i.e., when reflection / transmission conflicts occur), effectively balancing link gain and hardware overhead, and achieving system-level energy saving.

[0028] 2. Flexible full-space coverage: By utilizing the simultaneous reflection and transmission characteristics of STAR-RIS, it breaks through the physical limitations of traditional RIS "half-space" coverage, and can bypass obstacles through more flexible multi-hop paths to achieve full-space (360°) coverage of base station signals.

[0029] 3. Enhanced Adaptive Communication Capabilities in Complex Environments: This invention jointly optimizes beam routing and STAR-RIS operating mode configuration. Based on link status, path structure, and hardware power consumption differences under different operating modes, it adaptively selects appropriate operating modes and transmission paths for each STAR-RIS node on a multi-hop transmission path. Compared to fixed-mode or routing schemes based solely on link gain, this invention more effectively adapts to complex communication environments such as obstruction, multi-hop transmission, and heterogeneous node collaboration, improving system flexibility and environmental adaptability.

[0030] 4. Ensuring Quality of Service (QoS): While pursuing low power consumption, the algorithm consistently uses the minimum user data rate as a hard constraint. By maximizing the equivalent channel gain during continuous optimization phases, the system can meet users' QoS requirements with lower transmit power, thereby improving the reliability of the communication system. Attached Figure Description

[0031] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart of the main process of the low-power beam routing and working mode joint optimization method for multi-STAR-RIS assisted communication described in this embodiment of the invention. Figure 2 This is a schematic diagram illustrating the application scenario of the multi-STAR-RIS assisted communication system in this embodiment; Figure 3 This is a schematic diagram of the final beam routing topology for the embodiment; Figure 4 This is a comparison chart of simulation results showing the total system power consumption of the algorithm proposed in this embodiment and the comparison scheme as the number of users changes. Detailed Implementation

[0032] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0033] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0034] Example This embodiment uses Figure 2To study this scenario, a joint optimization method for low-power beam routing and operating modes in multi-STAR-RIS assisted communication is proposed. Under the premise of satisfying Quality of Service (QoS) constraints, the method aims to minimize the total system power consumption by jointly optimizing beam routing, STAR-RIS operating modes, phase shift matrix, number of base station beams, and beam power allocation. Since this optimization problem simultaneously involves discrete decision variables and continuous optimization variables, and they are coupled, it constitutes a mixed-integer nonlinear programming (MINLP) problem. Based on the Alternating Optimization (AO) framework, this problem is decomposed into two sub-problems: "phase shift and beam power optimization" and "beam routing and mode decision," which are solved alternately.

[0035] See Figure 1 The specific steps include the following: Step 1: Construction of System Model and Total Power Consumption Model 1.1 Communication System Scenarios like Figure 1 As shown, this embodiment considers a downlink transmission communication scenario based on multi-STAR-RIS assistance. The system includes a device equipped with... Base station (BS) with one antenna. A fixed-position STAR-RIS, and For single-antenna users, the direct link between the base station and the user is blocked by obstacles such as buildings, and the user can only receive signals through a multi-hop STAR-RIS cascaded link.

[0036] Each STAR-RIS consists of The system consists of passive units arranged in a uniform planar array (UPA). Each unit can simultaneously transmit (T) and reflect (R) the incident signal. The user set and the STAR-RIS set are denoted as follows: and .

[0037] 1.2 STAR-RIS Operating Modes and Phase Shift Model Each unit of STAR-RIS The amplitude and phase of the incident signal can be adjusted. Binary variables are introduced. To represent STAR-RIS The working mode, specifically...

[0038] Mode switching (MS) mode ( ): The node only operates in total internal reflection or total external reflection mode, and can only serve users on the reflection side or the external reflection side at any given time. or The circuit only needs to drive a single signal, resulting in low power consumption; the phase shift constraint is... , The matrix form is as follows: , or ,

[0039] Energy splitting (ES) mode The node performs both reflection and transmission, and can serve users on both the transmission and reflection sides simultaneously. However, circuit-driven power consumption is higher; it follows the energy conservation constraint: the reflection amplitude of a single unit With transmission amplitude satisfy The matrix form is as follows: , .

[0040] 1.3 Multi-hop beam routing model Due to the obstruction of direct links, the base station serves users via multi-hop STAR-RIS links. Routing variables are introduced. Indicates user The transmission path. Among them, Indicates the number of STAR-RIS in the path. Indicates the first in the path A selected STAR-RIS index, ,and For convenience, let's define... and These represent the base station and the user, respectively.

[0041] 1.4. System Total Power Consumption Model The total system power consumption consists of the base station's total transmit power and the static power consumption of all STAR-RIS hardware. The total base station transmit power is defined as... And among them , This refers to the transmit power of a single beam; The maximum allowable transmit power for the base station; the total system power consumption model is expressed as: ,in For the first The hardware power consumption of each STAR-RIS unit depending on its operating mode.

[0042] Assuming all STAR-RIS chips have the same size and number of cells, and all cells are fully operational, the power consumption per cell in "MS mode" is... In "ES mode", the power consumption of each unit is...

[0043] No. Power consumption per STAR-RIS node:

[0044] The total system power consumption consists of the base station transmit power and the STAR-RIS hardware static power consumption:

[0045] in, and These are the static power consumption of the STAR-RIS circuit in MS and ES modes, respectively.

[0046] Step 2: Construction of the joint optimization problem 2.1 Physical Channel Model All links use the line-of-sight (LoS) channel model, and the channel matrices for the three types of links are calculated respectively: Using a line-of-sight channel model, for STAR-RIS And the base station BS, whose array response vectors are respectively and :

[0047]

[0048] Based on this, the physical channels between each node are defined.

[0049] 1. Base station and the first Channels between STAR-RIS :

[0050] in, The carrier wavelength; For BS and STAR-RIS The straight-line distance between them; The path loss index; For STAR-RIS The array response vector; This is the response vector of the BS antenna array; To reach STAR-RIS Azimuth and elevation angles; The azimuth and elevation angles originating from BS.

[0051] 2. No. The first STAR-RIS and the first Channels between STAR-RIS;

[0052]

[0053] in, For STAR-RIS i and STAR-RIS The straight-line distance between them; For STAR-RIS The array response vector; For STAR-RIS The array response vector; To reach STAR-RIS Azimuth and elevation angles; For STAR-RIS The azimuth and elevation angles at departure.

[0054] 3. No. STAR-RIS and users The channel between them is:

[0055] , in, For STAR-RISj and users The straight-line distance between them; Let be the array response vector of STAR-RIS j; From STAR-RIS j to users The azimuth and elevation angles.

[0056] 2.2 Constructing a User Received Signal Model The base station uses multi-beam transmission technology, simultaneously transmitting Q independent beams, among which... Transmit signal ,in, Indicates assignment to the first The transmit power of each beam , This is the normalized beamforming vector. It is assumed that different active beams do not interfere with each other, and user-to-user interference within the same beam is not considered.

[0057] user Received signal:

[0058] user Received signal-to-noise ratio:

[0059] user Receive rate:

[0060] 2.3 MINLP Optimization Problem Construction Construct a mixed-integer nonlinear programming (MINLP) optimization problem with the objective of minimizing the total power consumption of the system.

[0061] Establish a system designed to minimize total power consumption. Optimization problem with the objective:

[0062] Where constraint C1 represents the user The rate must not be lower than its QoS requirements. Constraint C2 represents the amplitude and phase shift constraints of STAR-RIS. Constraint C3 represents the STAR-RIS mode selection restriction. Constraint C5 restricts the total transmit power of the BS from not exceeding the maximum power. Constraint C6 is the power allocation restriction of the BS active beam. Constraints C4 and C7 are the restrictions on the number of BS active beams.

[0063] Step S3: Solving the alternating optimization framework The highly coupled MINLP problem described above is decomposed into two sub-problems: "phase shift and beam power optimization (continuous variable)" and "beam routing and mode decision (discrete variable)," which are solved alternately. Due to the high coupling between discrete variables (beam routing, operating mode, number of beams) and continuous variables (phase shift matrix, beam power), directly finding the global optimum is computationally extremely complex. Therefore, this embodiment proposes an efficient solution framework based on Alternating Optimization (AO). First, under a fixed routing, the physical layer channel gain is maximized through continuous variable optimization, thereby reducing the transmit power required to meet QoS. Then, using this optimal channel state as a benchmark, a lower-power routing and mode are re-determined in the discrete domain. These two processes iterate cyclically until the total system power consumption converges.

[0064] The AO algorithm decomposes the original problem into the following two subproblems, which are solved alternately: Subproblem 1 (Phase Shift and Beam Power Optimization): Given a fixed user routing set, STAR-RIS operating mode, and number of beams, the STAR-RIS hardware power consumption is constant. Minimizing the total system power consumption is equivalent to minimizing the base station transmit power. While satisfying all user QoS rate constraints, this subproblem first optimizes the cascaded phase shift matrix using the Block Coordinate Descent (BCD) algorithm to maximize the end-to-end equivalent channel gain. Then, based on the optimal gain, it calculates the minimum beam transmit power required to satisfy the constraints.

[0065] Subproblem 2 (Beam Routing and Mode Decision): With the optimal phase shift matrix obtained from Subproblem 1 fixed (i.e., utilizing the current optimal beam alignment state), re-determine the user's beam routing, STAR-RIS activation state and operating mode (MS / ES), and the number of base station beams. The goal is to find a network topology that meets user requirements with the lowest total system power consumption (the sum of base station transmit power and STAR-RIS hardware power consumption) under the current channel gain conditions.

[0066] 3.1 Subproblem 1: Joint optimization of phase shift and beam power Under fixed user routing, operating mode, and number of beams, the phase shift matrix and base station beam power allocation of STAR-RIS are jointly optimized. By increasing the end-to-end equivalent channel gain, the total transmit power of the base station is minimized while meeting the minimum user rate requirements.

[0067] To facilitate solving the above optimization problem, we first derive the equivalent channel and received signal models of the system. Let... User The equivalent channel vector. If the user The beam routing is The equivalent channel can then be expressed as:

[0068] Due to users Only through a specific beam Received signal, of which Represented as user Assigned beam index. It is to serve users The transmit power of the beam, It is to serve users From the beamforming vector, the user's received signal-to-noise ratio expression can be further obtained:

[0069] To further simplify the optimization variables, given the phase shift matrix, the base station employs a Maximum Ratio Transmission (MRT) strategy, aligning the direction of the transmitted beam with the equivalent channel direction through conjugate transpose. According to the Cauchy-Schwarz inequality, the optimal beamforming vector for the base station is... At this time, the equivalent channel power gain Based on the above modeling, sub-problem 1 (phase shift and beam power optimization) can be solved in stages: Phase A: Cascaded Phase Shift Optimization Based on Block Coordinate Descent (BCD) and Semidefinite Relaxation To reduce the non-convexity of the optimization problem, a fixed amplitude strategy is adopted in the STAR-RIS ES mode, that is, the amplitudes of the reflection and transmission coefficients are set to be equal:

[0070] In ES mode, the reflection and transmission amplitudes of each unit are equal. Since all units introduce the same amplitude attenuation, this process can be equivalent to introducing a fixed amplitude attenuation factor at the link level. .

[0071] Because the signal may pass through multiple STAR-RIS hops, the end-to-end channel gain... This involves the product of multiple phase shift matrices. Directly optimizing all matrices together would lead to higher-order terms in the objective function, making it a non-convex optimization problem with multivariable coupling. Therefore, this stage employs the Block Coordinate Descent (BCD) algorithm, optimizing only the first phase shift matrix in each iteration. The phase shift of each STAR-RIS, while fixing the rest Each node.

[0072] (1) Decoupling of equivalent concatenated channels In the polling optimization stage STAR-RIS (denoted as STAR-RIS) When ), the phase shift of other nodes on the fixed path. For those passing through any user The path is defined as follows: Equivalent incident channel The signal originates from the base station and passes through the preceding nodes ( Reflection / transmission to reach The signal vector.

[0073] Equivalent output channel :from Departure, passing through subsequent nodes ( After cascading effects, the final result reaches the user. The reverse channel vector.

[0074] At this time, the user End-to-end equivalent scalar channel This can be further expressed as concerning the first STAR-RIS phase shift vectors The linear form of, where :

[0075] in, For Hadama accumulation, This is an auxiliary vector. Therefore, the channel gain corresponding to this hop is expressed as: .

[0076] (2) Construction of local SDR sub-problems according to The working mode divides users passing through this node into reflection sets. and transmission set .make They are respectively The reflection and transmission phase shift vectors are given. To maximize the channel gain for all relevant users, the following optimization problem is constructed:

[0077] Introducing positive semidefinite matrices and target matrix The problem can be relaxed to a semidefinite programming problem (SDP):

[0078] (3) Solving and Iteration The original non-convex optimization problem is transformed into a positive semidefinite programming (SDP) problem using the positive semidefinite relaxation (SDR) method, and the SDP problem is solved using the standard convex optimization method to obtain the optimal positive semidefinite matrix. Furthermore, the Gaussian randomization method is used to recover an approximate feasible solution with rank 1 from the matrix solution, thereby obtaining the corresponding optimal phase shift vector. The algorithm iterates through all the algorithms in turn. The process is repeated for several nodes in multiple rounds (InnerIterations) until the total channel gain of the system converges.

[0079] Phase B: Optimal Power Allocation Calculation After optimization in stage A, the phase shift matrices of all STAR-RIS are... It has been confirmed that each user path at this time... End-to-end equivalent channel gain It is a known constant.

[0080] And order Then the rate constraint is equivalent to:

[0081] Assuming each active beam of the base station points to only one first-hop STAR-RIS, paths to different first-hop STAR-RISes require different beams; if multiple user paths share the same first-hop STAR-RIS, they can share the same beam. One beam It may serve multiple users, let the first one be... The user set of beam services is Then the beam power All must be satisfied The needs of Chinese users:

[0082] Problem (P4) is transformed into:

[0083] This problem is a linear programming problem, and the optimal solution is obtained at the constraint boundary, that is, the minimum required value of each beam power:

[0084] At this time, the total transmission power of the base station is

[0085] 3.2 Sub-problem 2: Joint optimization of beam routing, operating mode and number of beams (1) Constructing a directed graph and generating candidate paths: Constructing a directed graph Node set Includes BS node (numbered 0) and all STAR-RIS nodes. and all user nodes The edge set E (directed edges) contains: if the channel from BS to STAR-RIS If it is non-zero, add an edge. If STAR-RIS To STAR-RIS channel Non-zero, and satisfies the unidirectional transmission geometric constraints (i.e., node) Distance from base station compared to node To further away (and prevent signal backflow), add an edge. If the STAR-RIS to user channel If it is non-zero, add an edge. .

[0086] To reduce the complexity of subsequent optimizations, this stage employs an "optimistic strategy" for path pre-screening, selecting the most promising paths based on physical channel quality. For each edge... Define weights This indicates the potential physical loss of the single-hop link:

[0087] in, Represents a node To the node The physical channel gain; for STAR-RIS nodes, it is assumed that their phases are aligned and the amplitude is 1. Since the edge weight is defined as the negative logarithm of the single-hop link channel gain, the total path weight is equal to the sum of the logarithms of the channel gains of each hop. Therefore, minimizing the total path weight is equivalent to maximizing the end-to-end cascaded channel gain of the path.

[0088] For each user In the constructed directed graph The weighted K-shortest path algorithm (such as Yen's algorithm) is used to search from the source node 0 to the target node. The former The path with the smallest weight. The path obtained above... The path is used as a user The set of candidate paths is denoted as:

[0089] (2) Construct and solve the integer linear programming (ILP) model For each candidate path Combined with the currently optimized phase shift matrix Calculate its true equivalent channel gain. Based on this calculation, the user is satisfied. Minimum beam transmit power required for QoS requirements:

[0090] At the same time, record the first hop STAR-RIS node of this path, denoted as... .

[0091] Introduce the following binary variables:

[0092]

[0093]

[0094]

[0095] Constructing an ILP optimization model:

[0096] Explanation of variable meanings: Indicates user Select path Otherwise, it is 0; STAR-RIS It is activated (i.e., at least one selected path passes through it), otherwise it is 0; This indicates that the base station is enabled, pointing to STAR-RIS. The active beam, otherwise 0; This indicates that the base station points to STAR-RIS. The beam transmission power.

[0097] Among them, C1 is the user single path constraint; C2 is the beam triggering constraint; C3 is the node activation constraint; C4 and C5 are the reflection side and transmission side capacity constraints, respectively; C6 is the ES mode triggering constraint; C7 is the user QoS constraint; and C8 is the binary variable constraint.

[0098] The solution is obtained by using the branch and bound method, which searches for the optimal combination of discrete variables within the feasible region that satisfies the constraints, thereby outputting the optimal routing scheme, STAR-RIS operating mode, and base station beam allocation strategy that minimizes the total power consumption of the system.

[0099] Figure 3 This is a schematic diagram of the final beam routing topology in this embodiment. The diagram clearly shows the optimal multi-hop link from which the base station output beam is transmitted to each user via multi-level STAR-RIS branching. It distinguishes between the reflection transmission branch and the transmission transmission branch of each metasurface, and intuitively presents the working mode of the globally optimal topology output by the ILP model and the matching of each STAR-RIS.

[0100] 3.3 Alternating Iterative Solution and Optimal Parameter Output The phase shift and beam power joint optimization in step S32 and the beam routing and operating mode joint optimization in step S33 are performed alternately and iteratively. In each iteration, the optimal solution obtained from the previous subproblem is used as the known constant input for the next subproblem. It is determined whether the difference in the total system power consumption between two adjacent iterations is less than the preset convergence threshold. If the convergence condition is met or the preset maximum number of iterations is reached, the iteration is stopped, and the final optimal global system configuration parameters (including beam routing scheme, STAR-RIS operating mode, phase shift matrix, and base station beam power allocation results) are output.

[0101] Simulation verification To verify the effectiveness of the proposed low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication systems, a simulation environment was built on the MATLAB platform for testing. The simulation scenario included one base station, 16 STAR-RIS nodes, and three single-antenna users. The base station was configured with 16 transmit antennas, and each STAR-RIS contained 64 reconfigurable elements. The carrier frequency was set to 8 GHz, the noise power to -105 dBm, and the user rate requirement to 4 bit / s / Hz.

[0102] During the simulation, a line-of-sight channel map was first constructed based on the node deployment locations, and a candidate route set was generated using the K-shortest path algorithm. Subsequently, an alternating optimization framework was employed for solution. In the continuous variable optimization phase, the phase shift matrix and beam power were jointly optimized using BCD and SDR methods. In the discrete variable optimization phase, ILP was used to jointly optimize beam routing, STAR-RIS operating mode, and the number of base station beams. The two sub-problems were iterated alternately until the total system power consumption converged, yielding the final optimization result.

[0103] Figure 4 The paper presents a comparison of the total system power consumption of the proposed algorithm with traditional single-hop schemes, heuristic routing schemes, and geometric phase-forming schemes. As shown in the figure, the joint optimization algorithm proposed in this invention achieves the lowest total system power consumption. This is because the method simultaneously considers the coupling relationship between beam routing, STAR-RIS operating mode, phase shift matrix, and power allocation. Under the premise of meeting user QoS requirements, it adaptively selects the optimal beam routing and STAR-RIS operating mode, effectively reducing base station transmit power and overall system power consumption. Simulation results demonstrate that this invention can significantly improve system energy utilization efficiency, verifying the effectiveness of the proposed method in achieving green and low-power communication in complex obstruction environments.

[0104] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication, characterized in that, Includes the following steps: S1. Construct a multi-STAR-RIS assisted downlink transmission communication system model and establish a total system power consumption model that includes base station transmit power and STAR-RIS hardware power consumption; S2. Based on the multi-STAR-RIS assisted downlink transmission communication system model, with the constraint of meeting the user's quality of service (QoS) requirements and the objective of minimizing the total system power consumption, a mixed integer nonlinear programming joint optimization problem is constructed. The joint optimization variables include beam routing, STAR-RIS operating mode, reflection and transmission phase shift matrix, number of base station beams, and beam power allocation parameters. S3. Solve the joint optimization problem using an alternating optimization framework: Decouple the original problem into a continuous variable optimization subproblem and a discrete variable optimization subproblem, and iterate the two subproblems alternately until the total power consumption of the system converges, and output the optimal beam routing, STAR-RIS working mode, phase shift matrix, number of base station beams and beam power allocation parameters.

2. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 1, characterized in that, The multi-STAR-RIS assisted downlink transmission communication system model in step S1 specifically includes: A configuration The base station uses a single transmitting antenna and employs multi-beam transmission technology to serve users; A fixed-position, simultaneously reflective and transmissive smart metasurface STAR-RIS is deployed between the base station and the user. Each STAR-RIS consists of N reconfigurable passive units, each of which can simultaneously reflect and transmit signals to build multi-hop cascaded links for auxiliary communication. In a user cluster consisting of multiple single-antenna users, where the direct link between the base station and the user is blocked, each user receives signals through only one multi-hop STAR-RIS path. The STAR-RIS operating modes include mode switching (MS) mode and energy splitting (ES) mode. The operating mode selection is represented by binary variables. In MS mode, the node operates only in total reflection or total transmission state, while in ES mode, the node performs both reflection and transmission.

3. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 2, characterized in that, The system total power consumption model in step S1 is as follows: The total system power consumption consists of the sum of the base station's total transmit power and the static power consumption of all STAR-RIS hardware, expressed as: in This represents the total transmit power of the base station. For the first The hardware power consumption of each STAR-RIS depends on its operating mode. The static power consumption of a single STAR-RIS in different operating modes is as follows: MS mode: In ES mode: In the formula, This refers to the basic static power consumption of the STAR-RIS control module. The driving power consumption of a single passive unit in MS mode. This represents the drive power consumption of a single passive unit in ES mode, and > .

4. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 1, characterized in that, The mixed-integer nonlinear programming joint optimization problem constructed in step S2 has the following constraints: Constraints include: user reception rate not lower than the minimum QoS requirement; amplitude and phase constraints of the STAR-RIS unit; binary variable constraints of the operating mode; base station beam count constraints; maximum base station transmit power constraints; and non-negative beam power constraints.

5. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 1, characterized in that, The specific solution process for the alternating optimization framework in step S3 is as follows: S31. Variable decoupling: Decompose the original problem into a continuous variable subproblem of phase shift and beam power optimization, and a discrete variable subproblem of beam routing and mode decision. S32. Continuous variable optimization: With the current user beam routing, STAR-RIS operating mode and number of base station beams fixed, solve the continuous variable subproblem to obtain the optimal cascaded phase shift matrix and the corresponding beam power allocation result; S33. Discrete variable optimization: Fix the current optimal phase shift matrix, solve the discrete variable subproblem, and re-determine the user beam routing, STAR-RIS working mode and the number of base station beams; S34. Alternating Iteration: Repeatedly execute steps S32 and S33, using the optimal solution of the previous subproblem as the input of the next subproblem, until the difference in the total power consumption of the system between two adjacent iterations is less than the preset convergence threshold, or the maximum number of iterations is reached, and output the globally optimal configuration parameters.

6. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 5, characterized in that, In step S32, the optimization of the phase shift matrix is ​​achieved in the following way: A block coordinate descent algorithm is used to decouple the phase shift of multi-hop cascaded nodes. In each iteration, only the phase shift of a single STAR-RIS node is optimized, while the phase shift parameters of all other nodes are fixed. For a single STAR-RIS node to be optimized, the users passing through the node are divided into a reflection user set and a transmission user set. A positive semidefinite matrix is ​​introduced to relax the original non-convex phase shift optimization problem into a positive semidefinite programming problem and solve it. Then, the optimal phase shift vector with rank 1 that satisfies the modulus constraint is recovered from the relaxed solution through the Gaussian randomization method, thereby maximizing the end-to-end equivalent channel gain. All STAR-RIS nodes are polled sequentially and iterated until the total channel gain of the system converges.

7. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 5, characterized in that, In step S32, the beam power allocation is calculated as follows: Based on the determined phase shift matrix, the base station adopts the maximum ratio transmission strategy to align the direction of the transmit beam with the direction of the user's equivalent channel. At this time, the end-to-end equivalent channel gain of the user is a constant. The signal-to-noise ratio (SNR) threshold is derived based on the user's minimum data rate requirement. Combined with the equivalent channel gain, the minimum beam transmit power required to meet QoS requirements is calculated. The expression is: in, The set of users serving the q-th beam. Let u be the end-to-end equivalent channel gain for user u. To meet the minimum signal-to-noise ratio threshold required for the lowest data rate, the power of each beam is taken as the maximum power required to serve all users.

8. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 5, characterized in that, In step S33, candidate path screening is performed before discrete variable optimization, specifically including: Construct a directed graph of the line-of-sight channel, with the node set including the base station, all STAR-RIS nodes and all user nodes, and the directed edges consisting of non-zero channel links that satisfy the unidirectional transmission geometric constraints; Define the edge weight for each directed edge. The edge weight is taken as the negative logarithm of the squared modulus of the physical channel gain of the corresponding single-hop link, and its expression is: in, Let be the physical channel matrix pointing from node i to node j. Denotes the Frobenius norm of a matrix; Minimizing the total path weight is equivalent to maximizing the end-to-end cascaded channel gain of the path; The K-shortest path algorithm is used to search for the top L paths with the smallest weights for each user on the directed graph, forming a candidate path set for that user.

9. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 8, characterized in that, After generating the candidate path set, the discrete variable optimization subproblem is further solved by the following steps: Introducing three types of binary variables—user path selection, STAR-RIS activation, and base station beam triggering—along with the continuous variable of beam transmit power, an integer linear programming model is constructed with the objective function of minimizing the total system power consumption. in Let j be the beam transmit power directed at the j-th STAR-RIS. As the activation variable for STAR-RIS, Choose a variable for the working mode. For beam triggering variables, This is the beam count penalty coefficient; Model constraints include user single-path constraints, beam triggering constraints, node activation constraints, reflection and transmission side capacity constraints, ES mode triggering constraints, user QoS power constraints, and variable domain constraints. The branch-and-bound method is used to solve the integer linear programming model to obtain the optimal combination of user routing, STAR-RIS operating mode and base station beam allocation.

10. The low-power beam routing and operating mode joint optimization method for multi-STAR-RIS assisted communication according to claim 9, characterized in that, The specific ES mode triggering constraint is as follows: The high-power ES mode is triggered if and only if both a reflection path and a transmission path exist at the STAR-RIS node. When a node has only a reflection path or only a transmission path, it operates in a low-power MS mode, reducing system hardware power consumption through adaptive mode switching.

Images