Multi-service QoS cross-band bandwidth allocation method and system based on deep reinforcement learning

Through deep reinforcement learning methods, the effective capacity model and multi-task hierarchical architecture are constructed, combined with MADDPG and DDPG algorithms, the problems of poor user experience and low effective capacity in multi-band environments in wireless communication systems are solved, and efficient resources are achieved Improve management and user experience.

CN119052861AActive Publication Date: 2024-11-29DALIAN UNIV

Patent Information

Application Number
CN202411263649.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-10
Publication Date
2024-11-29
Estimated Expiration
2044-09-10

AI Technical Summary

Technical Problem

The existing wireless communication systems fail to fully consider the characteristics of different frequency bands and service QoS differences in the multi-band environment, resulting in poor user experience and low effective capacity.

Method used

Using a multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning, the effective capacity model and multi-task hierarchical architecture are constructed, combined with MADDPG and DDPG algorithms, and the initial and refined allocation of bandwidth resources is performed at the high and low levels respectively to meet the different QoS requirements for the service.

Benefits of technology

It achieves high average effective capacity and stable performance under different number of users, balances the global optimization and local optimization requirements of resource allocation, reduces latency, and improves SINR and user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119052861A_ABST
    Figure CN119052861A_ABST
Patent Text Reader

Abstract

The invention discloses a multi-service QoS cross-band bandwidth allocation method and system based on deep reinforcement learning, and relates to the technical field of wireless communication. Comprising the following steps: constructing an effective capacity model to define an overall optimization target of bandwidth resources; on the basis of the effective capacity model, a multi-task hierarchical architecture based on deep reinforcement learning is constructed, and the architecture preliminarily allocates total bandwidth resources of Sub-6GHz and millimeter wave frequency bands through a high-level network; fine allocation is carried out on specific bandwidth resources in each frequency band through a low-layer network; a shared presentation layer, a strategy head and a value head are introduced into a high-level network and a low-level network respectively, the shared presentation layer captures common features of different frequency bands, the strategy head generates a specific distribution strategy, and the value head evaluates the distribution effect. According to the method, global optimization and local optimization requirements in resource allocation are successfully balanced through combination of global and local strategies, and the method has outstanding performance in the aspects of reducing delay, improving SINR and improving user experience.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular to a multi-service QoS cross-band bandwidth allocation method and system based on deep reinforcement learning. Background Art

[0002] With the rapid growth of global mobile data traffic, modern communication networks are facing unprecedented challenges. Users' demand for high-speed data transmission, extremely low latency and high reliability continues to surge, and traditional spectrum resources can no longer meet this demand. In this context, 5G communication systems continue to explore new spectrum resource boundaries and seek more efficient bandwidth utilization strategies. [1] , to meet this challenge. Especially in a complex and ever-changing multi-service environment, ensuring the quality assurance (QoS) of various services has become a core task. The Sub-6GHz band has been widely used in the industry due to its excellent coverage capabilities, but its limited bandwidth can hardly support the future demand for ultra-high data rates. In order to break this bottleneck, millimeter wave (mmWave) communication technology came into being, and its extremely high available bandwidth has opened a new era for large-scale data transmission. However, millimeter wave signals are easily affected by physical obstacles, and their coverage range is relatively limited. Therefore, Sub-6GHz and millimeter wave technology are cleverly combined to build a cross-band collaborative networking architecture. [2] , which has become an effective strategy to solve the current dilemma. Under this architecture, how to intelligently allocate bandwidth resources among different frequency bands to accurately meet the user's service quality assurance (QoS) requirements? [3] , which has become a key problem that the industry needs to solve urgently. The significant differences between millimeter wave and Sub-6GHz bands in propagation characteristics, transmission distance, and interference sensitivity have greatly increased the complexity and challenge of cross-band resource allocation. An effective allocation strategy not only needs to consider the physical characteristics of the frequency band, but also needs to respond to the differentiated QoS requirements of different services to optimize the user experience.

[0003] When implementing QoS-based bandwidth allocation strategies in a cross-band environment, the main challenge facing current technologies is that they fail to fully take into account the complex QoS differences between multiple services, making it difficult to effectively guarantee the service quality requirements of specific services, especially the delay and reliability indicators of some key services. [4] The focus is on the transmission of specific services, and the reliability and rate of information transmission are improved through spectrum sharing, but the important delay performance requirements are not fully considered. However, in actual application scenarios, the network environment usually needs to meet the QoS requirements of multiple services at the same time. This requires not only optimizing the performance of a single service, but also comprehensively considering the requirements of multiple services in terms of transmission rate, packet delay, packet loss rate, etc. [5-8]The weighted QoS deviation utility function is used to integrate multiple indicators to meet the diverse QoS requirements of users. However, since multiple weight parameters need to be adjusted, the algorithm complexity is increased and it is difficult to ensure the accuracy of the configuration. [9]

[10] A utility function based on delay deviation was designed. Although it performs well in processing delay-sensitive services, it ignores other key QoS indicators. In general, methods based on QoS utility functions have achieved some results in improving multi-service QoS guarantees, but they usually only focus on a single performance indicator and cannot fully describe user needs. In contrast, the effective capacity model shows unique advantages.

[11] The effective capacity is used to evaluate the diversified QoS requirements of users, and a spectrum allocation algorithm that can meet the multi-faceted QoS requirements of users is designed accordingly. The outstanding feature of the effective capacity model is its comprehensiveness, which covers not only the transmission rate, but also key performance indicators such as delay and packet loss rate. This makes it more valuable in practical applications than the utility function that only focuses on a single indicator (such as Shannon capacity or delay).

[0004] In addition to building an efficient QoS model, researchers have also done a lot of work on specific allocation schemes.

[12] An improved Bald Eagle Search (IBES) algorithm is proposed, which comprehensively considers factors such as user service QoS, channel state information, and resource periodicity. However, the QoS analysis in this paper is based on fixed parameters set in advance, lacks flexibility, and is difficult to adapt to the changing situations in practical applications.

[13] The proposed ideal fractional programming method (Ideal FP) assumes that the global channel state information (CSI) can be obtained instantaneously, which is difficult and cannot adapt to the complex environment of reality.

[14] The effective capacity is used to optimize the access right allocation of relay nodes through a forward auction algorithm, although the problem of multiple concurrent auction requests in ultra-dense network environments has not been explored in depth. With the development of artificial intelligence, machine learning technology has been widely used in bandwidth allocation in wireless communication systems. [15-17] The optimization of bandwidth allocation is carried out by combining multi-agent collaboration and reinforcement learning, but the frequency band characteristics are not fully considered, which limits the room for improvement of its allocation effect.

[18] An improved Deep Deterministic Policy Gradient (DDPG) algorithm is proposed, which comprehensively considers resources such as computing, storage, and network bandwidth. This method effectively handles complex resource allocation problems, but due to its high complexity, it may face real-time and computing resource limitation problems in actual deployment.

[19] A Multi-Agent Double Deep Q Network (MADDQN) and Prioritized Multi-agent Deep Deterministic Policy Gradient (P-MADDPG) were developed to solve the bandwidth allocation subproblem for each user, but the research mainly focused on specific network settings and conditions without considering the scalability and adaptability of the algorithm in other environments or networks of different sizes.

[20] A hybrid deep deterministic policy gradient and deep Q network (Deep QNetwork, DQN) method is proposed to more comprehensively address the resource allocation problem, but the generalization and stability under different network scales or configurations have not been fully verified.

[0005] In summary, although these methods perform well in a single frequency band or uniform spectrum environment, they often fail to fully consider the characteristics and requirements of different frequency bands in a complex multi-band environment. These methods usually treat all frequency bands as homogeneous resources, ignoring the differences in coverage, propagation loss, and data rate between frequency bands. At the same time, they lack detailed modeling of frequency band characteristics during the optimization process and fail to implement personalized resource allocation strategies for each frequency band. Therefore, despite the overall improvement, it is still difficult to fully tap the potential and advantages of resource allocation in a multi-band environment.

[0006] [1] N. Prodromos, D. Diasakos, V. Kokkinos, A. Gkamas, C. Bouras and P. Pouyioutas, "Dynamic Bandwidth Allocation in MIMO 5G Networks," 2024International Wireless Communications and Mobile Computing(IWCMC),AyiaNapa,Cyprus,2024,pp.97-102,doi:10.1109 / IWCMC61514.2024.10592577.

[0007] [2]Lopatka J, Paso T, Massin R, et al.Multi band efficient networks for ad hoc communications[J]. Procedia Computer Science, 2022, 205:88-96.

[0008] [3]T.Eckert and S.Bryant,“Quality of Service(QoS),”in FutureNetworks,Services and Management:Underlay and Overlay,Edge,Applications,Slicing,Cloud,Space,AI / ML,and Quantum Computing,M.Toy,Ed.,Cham:SpringerInternational Publishing,2021,pp.309–344.doi:10.1007 / 978-3-030-81961-3_11.

[0009] [4]L.Liang,S.Xie,G.Y.Li,Z.Ding and X.Yu,"Graph-Based Resource Sharingin Vehicular Communication,"in IEEE Transactions on Wireless Communications,vol.17,no.7,pp.4579-4592,July 2018,doi:10.1109 / TWC.2018.2827958.

[0010] [5]L.Li,N.Deng,W.Ren,B.Kou,W.Zhou and S.Yu,"Multi-Service ResourceAllocation in Future Network With Wireless Virtualization,"in IEEE Access,vol.6,pp.53854-53868,2018,doi:10.1109 / ACCESS.2018.2871506.

[0011] [6]Liu,Q.;Li,R.;Li,Y.;Wang,P.;Sun,J.Adaptive Bandwidth Allocation forMassive MIMO Systems Based on Multiple Services.Appl.Sci.2023,13,9861.https: / / doi.org / 10.3390 / app13179861

[0012] [7]Bendaoud F.Network selection in a heterogeneous wirelessenvironment based on path prediction and usermobility[M] / / Comprehensive Guideto Heterogeneous Networks.Academic Press,2023:59-86.

[0013] [8]Huang,J.;Yang,Z.;Xie,J.;Zhang,H.;Li,Z.Joint Power and BandwidthAllocation in Collocated MIMO Radar Based on the Quality of ServiceFramework.Electronics 2023,12,2567.

[0014] https: / / doi.org / 10.3390 / electronics12122567

[0015] [9]W.Huang,L.Ding,D.Meng,J.-N.Hwang,Y.Xu and W.Zhang,"QoE-BasedResource Allocation for Heterogeneous Multi-Radio Communication in Software-Defined Vehicle Networks,"in IEEE Access,vol.6,pp.3387-3399,2018,doi:10.1109 / ACCESS.2018.2800036.

[0016]

[10] P.V.P K,K.Jagannathan,D.B and K.Milleth,"Resource Allocation forQoS Enforcement in 5G:Trading off Fairness and Delay-Awareness,"2024 NationalConference on Communications(NCC),Chennai,India,2024,pp.1-6,doi:10.1109 / NCC60321.2024.10485929.

[0017]

[11] X.Mi,L.Xiao,M.Zhao,X.Xu and J.Wang,"Statistical QoS-DrivenResource Allocation and Source Adaptation for D2D Communications UnderlayingOFDMA-Based Cellular Networks,"in IEEE Access,vol.5,pp.3981-3999,2017,doi:10.1109 / ACCESS.2017.2679113.

[0018]

[12] Liu,Q.;Li,R.;Li,Y.;Wang,P.;Sun,J.Adaptive Bandwidth Allocationfor Massive MIMO Systems Based on Multiple Services.Appl.Sci.2023,13,9861.https: / / doi.org / 10.3390 / app13179861

[0019]

[13] K.Shen and W.Yu,"Fractional Programming for CommunicationSystems—Part I:Power Control and Beamforming,"in IEEE Transactions on SignalProcessing,vol.66,no.10,pp.2616-2630,15 May15,2018,doi:10.1109 / TSP.2018.2812733.

[0020]

[14] X.Huang,F.She,K.Wu and M.Jiang,"Device-to-Device ContentPlacement and Delivery Exploiting Joint Matching and Auction Theories,"2020International Conference on Wireless Communications and Signal Processing(WCSP),Nanjing,China,2020,pp.795-800,doi:10.1109 / WCSP49889.2020.9299732.

[0021]

[15] J.Wang,X.Zhang,X.He and Y.Sun,"Bandwidth Allocation andTrajectory Control in UAV-Assisted IoV Edge Computing Using MultiagentReinforcement Learning,"in IEEE Transactions on Reliability,vol.72,no.2,pp.599-608,June 2023,doi:10.1109 / TR.2022.3192020.

[0022]

[16] H.Albinsaid,K.Singh,S.Biswas and C.-P.Li,"Multi-AgentReinforcement Learning-Based Distributed Dynamic Spectrum Access,"in IEEETransactions on Cognitive Communications andNetworking,vol.8,no.2,pp.1174-1185,June 2022,doi:10.1109 / TCCN.2021.3120996.

[0023]

[17] B.Lim and M.Vu,"Distributed Multi-Agent Deep Q-Learning for LoadBalancing User Association in Dense Networks,"in IEEEWirelessCommunicationsLetters,vol.12,no.7,pp.1120-1124,July 2023,doi:10.1109 / LWC.2023.3250492.

[0024]

[18] Q.Liu, H.Zhang, X.Zhang and D.Yuan, "Improved DDPG Based Two-Timescale Multi-Dimensional Resource Allocation for Multi-Access EdgeComputing Networks," in IEEE Transactions on Vehicular Technology, vol.73, no.6, pp.9153-9158, June2024, doi:10.1109 / TVT.2024.3360943.

[0025]

[19] Zhang, C.; Lv, T.; Huang, P.; Lin, Z.; Zeng, J.; Ren, Y.

[0026]

[20] H.Hu, D.Wu, F.Zhou, Technology, vol.72, no.8, pp.10696-10709, Aug.2023, doi:10.1109 / TVT.2023.3253905. Summary of the invention

[0027] In order to solve the problem that bandwidth allocation in wireless communication systems fails to fully consider the QoS differences of multiple services and the characteristics of different frequency bands (Sub-6GHz and millimeter wave), resulting in poor user experience and low effective capacity, the present invention proposes a multi-service QoS cross-band bandwidth allocation method and system based on deep reinforcement learning.

[0028] According to a first aspect of an embodiment of the present disclosure, a multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is provided, comprising the following steps:

[0029] Build an effective capacity model, comprehensively consider key performance indicators such as transmission rate, delay and packet loss rate of different services, and clarify the overall optimization goal of bandwidth resources;

[0030] Based on the effective capacity model, a multi-task hierarchical architecture based on deep reinforcement learning is constructed. This architecture makes a preliminary allocation of the overall bandwidth resources of the Sub-6GHz and millimeter wave bands through the high-level network to ensure the reasonable allocation of bandwidth resources across the frequency bands. The specific bandwidth resources within each frequency band are then finely allocated through the low-level network to meet the QoS requirements of different services.

[0031] A shared representation layer, a policy head and a value head are introduced into the high-level network and the low-level network respectively. The shared representation layer captures the common characteristics of different frequency bands, the policy head generates specific allocation strategies, and the value head evaluates the allocation effect to achieve more accurate resource allocation.

[0032] According to a second aspect of an embodiment of the present disclosure, a multi-service QoS cross-band bandwidth allocation system based on deep reinforcement learning is provided, including:

[0033] Model building module: constructs an effective capacity model, comprehensively considers key performance indicators such as transmission rate, delay and packet loss rate of different services, and clarifies the overall optimization goal of bandwidth resources;

[0034] The architecture design module uses the effective capacity model as the basis to build a multi-task hierarchical architecture based on deep reinforcement learning. This architecture makes a preliminary allocation of the overall bandwidth resources of the Sub-6GHz and millimeter wave bands through the high-level network to ensure the reasonable allocation of bandwidth resources across frequency bands. It then makes a refined allocation of specific bandwidth resources within each frequency band through the low-level network to meet the QoS requirements of different services.

[0035] The allocation module introduces a shared representation layer, a policy head, and a value head in the high-level network and the low-level network, respectively. The shared representation layer captures the common characteristics of different frequency bands, the policy head generates a specific allocation strategy, and the value head evaluates the allocation effect to achieve more accurate resource allocation.

[0036] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored and running on the memory, wherein when the processor executes the program, the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is implemented.

[0037] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is implemented.

[0038] Compared with the prior art, the above technical solution adopted by the present invention has the following advantages: The present invention proposes a multi-service QoS cross-band bandwidth allocation method and system based on deep reinforcement learning, which aims to solve the problems of poor user experience and low effective capacity caused by the failure to fully consider the differences in multi-service QoS and the characteristics of different frequency bands in the current wireless communication system. By constructing an effective capacity model, designing a two-layer deep reinforcement learning architecture, and introducing a shared representation layer, a policy head, and a value head at the high and low layers respectively, effective management of Sub-6GHz and millimeter wave frequency band resources is achieved. The high-level strategy uses the MADDPG algorithm to perform rough bandwidth allocation across frequency bands, and the low-level strategy uses the DDPG algorithm to perform fine-grained allocation of bandwidth within each frequency band to further meet the QoS requirements of each service.

[0039] The present invention can not only maintain a high average effective capacity under different numbers of users, but also better cope with the network pressure caused by the increase in the number of users, showing more stable performance. In addition, through the combination of global and local strategies, it successfully balances the global optimization and local optimization requirements in resource allocation, and performs outstandingly in reducing latency, improving SINR and improving user experience, providing a solid foundation for further improving system throughput in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings in the specification, which constitute a part of the present application, are used to provide further understanding of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0041] Figure 1 This is a diagram of the Sub-6GHz and millimeter wave collaborative networking scenario;

[0042] Figure 2 This is a schematic diagram of a multi-task hierarchical architecture based on deep reinforcement learning;

[0043] Figure 3 Generate action decision flow chart for Actor network;

[0044] Figure 4 This is the flowchart of Critic network value assessment. DETAILED DESCRIPTION

[0045] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0046] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0047] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0048] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the methods and systems according to various embodiments of the present disclosure. It should be noted that each box in the flowchart or block diagram can represent a module, a program segment, or a part of a code, and the module, program segment, or a part of a code may include one or more executable instructions for implementing the logical functions specified in each embodiment. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, or they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the flowchart and / or block diagram, and the combination of boxes in the flowchart and / or block diagram can be implemented using a dedicated hardware-based system that performs a specified function or operation, or can be implemented using a combination of dedicated hardware and computer instructions.

[0049] Embodiment 1:

[0050] This embodiment provides a multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning, including the following steps:

[0051] S1. Build an effective capacity model, comprehensively consider the key performance indicators of different services, and clarify the overall optimization goal of bandwidth resources;

[0052] Specifically, in order to achieve multi-service QoS guarantee, low-frequency base stations (LF-BS) below 6 GHz and millimeter-wave high-frequency base stations (HF-BS) are integrated to achieve high- and low-frequency collaborative networking, such as Figure 1 As shown; the Sub-6GHZ frequency band of LF-BS is equally divided into N s The millimeter wave frequency band of HF-BS is equally divided into N bandwidth resource blocks. mvThere are bandwidth resource blocks, and each bandwidth resource block has services with different QoS requirements. The user may have a communication link with the LF-BS or HF-BS, and the link m consists of a receiver m and its transmitter m. The transmission link between the user and the base station uses an orthogonal channel, so during the downlink transmission process, there is no downlink interference between cells within the coverage area of ​​each HF-BS microcell.

[0053] Sub-6GHz and millimeter wave bands have different requirements for antennas, which is reflected in the calculation of signal-to-interference-noise ratio. Sub-6GHz bands usually only consider antenna power gain, because this is sufficient to describe relatively uniform signal radiation and reception characteristics. In the millimeter wave band, the directional antenna gain of the transmitting and receiving ends needs to be considered to ensure that signals can be effectively sent and received in the case of short wavelengths and complex propagation environments, thereby improving system reliability and performance.

[0054] The signal to interference plus noise ratio (SINR) at the receiver m of the low-frequency base station LF-BS can be expressed as:

[0055]

[0056] in, is the power gain of the Sub-6GHz low-frequency cellular antenna, N s is the power spectral density of the additive white Gaussian noise (AWGN) of the cellular spectrum resources, B s is the bandwidth of cellular user spectrum resources; p m represents the transmission power on link m, h m is the channel gain from user m to LF-BS, is the sum of interferences from other links to link m.

[0057] The signal-to-interference-to-noise ratio at the receiver m of the high-frequency base station HF-BS is expressed as:

[0058]

[0059] in, and are the directional antenna gains of the HF-BS transmitting and receiving users respectively. mv is the additive white Gaussian noise power spectral density of millimeter wave spectrum resources, B mv The bandwidth of the spectrum resources for millimeter wave users;

[0060] The effective capacity model comprehensively considers key indicators such as transmission rate, delay and packet loss rate, and quantifies the average data transmission rate of links in different frequency bands through effective capacity. It describes the average data transmission rate that can be achieved under given wireless channel conditions to meet different service QoS requirements:

[0061]

[0062] Where T represents the time slot length, T D Indicates the packet delay, set t delay is the maximum tolerable delay of the data packet (when the data packet exceeds t delay When θ is greater than θ, it is considered lost), R represents the link transmission rate, and ε represents the maximum allowed packet loss rate. θ is a QoS guarantee parameter, covering key performance indicators such as transmission rate, delay, and packet loss rate. For real-time data traffic transmission services, the delay needs to be strictly controlled, that is, θ→0, and the effective capacity is expressed as interruption capacity; while for non-real-time data transmission services, the focus is on achieving high throughput while allowing loose delay constraints, that is, θ→∞, and the effective capacity will become the ergodic capacity of each state. The relationship between the QoS guarantee parameter θ and the packet arrival rate, delay, and packet loss rate is expressed as follows:

[0063]

[0064] Therefore, the effective capacity model of LF-BS and HF-BS is constructed based on the above definition:

[0065]

[0066]

[0067]

[0068] in, is the delay QoS constraint parameter of the link. m represents the communication mode selected by the link, α m =0 means LS-BS communication is selected, α m =1 means HF-BS communication is selected, 0<α m <1 indicates the ratio of mmWave bandwidth allocated to HF-BS on the link.

[0069] As a preferred embodiment, in order to further ensure transmission reliability, a data overflow / loss constraint is introduced based on the effective capacity model. This constraint serves as an important restriction to ensure that the resource allocation scheme meets the reliability requirements of the actual system.

[0070] Assume that the data packet is infinitely small and the source node on link m sends data at a constant rate λ mGeneration / arrival; these packets are temporarily stored in a First In First Out (FIFO) queue before transmission. The probability of packet loss on link m can be expressed as:

[0071]

[0072] Among them, T m represents the data delay of link m, Q m represents the cache queue length of link m, represents the maximum allowable data delay of service b, Indicates the maximum allowable cache queue length for service b, data transmission exceeds the maximum delay threshold or the data cached by the node exceeds the maximum queue length threshold When , data is lost. m ) represents the probability that the link m cache is not empty:

[0073]

[0074] Therefore, the constraint condition of the data overflow / loss probability of the link is expressed as:

[0075]

[0076] Among them, ε b Indicates the maximum data overflow / loss probability threshold allowed by service b.

[0077] It should be noted that the present invention expresses the bandwidth resource allocation problem as an optimization problem, which is specifically described as follows: In order to improve the upper limit of system throughput while ensuring the multi-service QoS requirements, the present invention takes the maximization of link effective capacity as the overall optimization goal of bandwidth resources, and constructs a cross-band bandwidth allocation problem based on multi-service QoS. The problem takes the overflow / loss probability of the transmission delay exceeding the service delay threshold, the link capacity being higher than the data arrival rate, and the minimum communication quality guarantee as constraints;

[0078]

[0079] in, and They represent link m in Sub6-GHz bandwidth resource block n. s and bandwidth resource block n on mmWave mv The allocation matrix on θ = [θ m ] is the QoS guarantee parameter configuration matrix.

[0080] S2. Based on the effective capacity model, a multi-task hierarchical architecture based on deep reinforcement learning is constructed. This architecture makes a preliminary allocation of the overall bandwidth resources of the Sub-6GHz and millimeter wave bands through the high-level network to ensure the reasonable allocation of bandwidth resources across the frequency bands. The specific bandwidth resources within each frequency band are then finely allocated through the low-level network to meet the QoS requirements of different services.

[0081] The optimization problem described in this paper is a non-convex optimization problem. The proposed multi-service QoS cross-band bandwidth allocation method aims to maximize the total effective capacity of the network while meeting the QoS requirements of different service types. To achieve this goal, a multi-task hierarchical deep reinforcement learning architecture is designed, which combines the MADDPG and DDPG algorithms to optimize high-level and low-level decision-making requirements respectively; each communication link acts as an independent intelligent agent.

[0082] In the high-level network decision-making stage, the main task is to make a preliminary allocation of the total bandwidth resources of the Sub-6GHz and millimeter wave bands. This stage requires not only coordinating resource allocation among multiple communication links, but also fully considering the characteristics of the frequency band and the dynamic needs of each link, and giving priority to allocating resources to those links that need the most bandwidth. Therefore, the MADDPG algorithm is chosen. MADDPG is good at handling collaboration and competition in multi-agent environments. Based on global information, it can generate the optimal allocation strategy across links through the strategy network and value network of each link agent, and preliminarily allocate the total bandwidth roughly.

[0083] In the low-level network decision-making stage, the task changes to fine-grained allocation of specific bandwidth resources within each frequency band. The decisions involved in this stage are more localized, focusing on optimizing resource allocation for each link within its own frequency band. Since the decision of each link at this time is more focused on the bandwidth allocation within its own frequency band, the simpler and more efficient DDPG algorithm is selected. DDPG is suitable for single-agent environments and can quickly generate action decisions to further optimize bandwidth allocation based on the resource requirements of each link within a specific frequency band. Through independent policy networks and value networks, DDPG refines the bandwidth resources allocated at a high level and accurately meets its quality of service requirements.

[0084] In order to realize multi-task hierarchical bandwidth resource allocation, the state space, action space and reward function are first defined:

[0085] Its state space S includes: the receiver's demand matrix θ m , instantaneous channel state information h m , the tolerable delay threshold of the current service b of the link transmitter and the maximum acceptable data overflow / loss probability The instantaneous traffic arrival rate of the link ism The state space of high-level and low-level decisions is the same, but the low-level is more detailed to the state of specific links and frequency bands.

[0086] Its action space A includes: The action space of the high-level network is defined as the total bandwidth resource allocation strategy of the link in different frequency bands, ρ m is the allocation ratio of Sub6-GHz bandwidth resources of LF-BS, α m It represents the allocation ratio of millimeter wave bandwidth resources of HF-BS. The action space of the low-level network is expressed as A = {ρ, α}, which independently optimizes the specific resource allocation of each frequency band to adapt to the characteristics and requirements of different frequency bands.

[0087] Its reward function R includes: The high-level network instantaneous reward is designed as a global effective capacity indicator, the purpose is to provide the highest possible system transmission rate while meeting the QoS requirements of different services. Therefore, the instantaneous reward of the tth time slot can be expressed as:

[0088]

[0089] The instantaneous reward of the low-level network is designed to be the average effective capacity that the agent and its neighboring agents can achieve. Therefore, the instantaneous reward at the tth time slot can be expressed as:

[0090]

[0091] in, is the sum of the number of the current agent and the observed neighbor agents.

[0092] Multi-task learning (MTL) is a method that improves the generalization ability of a model by learning multiple related tasks simultaneously. In the bandwidth resource allocation problem, multi-task learning can effectively handle the QoS requirements of multiple services and improve resource utilization efficiency. Traditional single-task training methods often ignore valuable information in related tasks, while multi-task learning allows the model to learn multiple related tasks, mine common information between tasks, and share representations between related tasks, so that the model can better generalize to the original task, thereby improving performance on specific indicators.

[0093] The core of MTL is to process multiple tasks in parallel by sharing the underlying representation learning, thereby improving the performance of each task, capturing the commonalities and relationships between tasks, and ultimately improving the performance of the overall model.

[0094] S3. A shared representation layer, a policy head, and a value head are introduced into the high-level network and the low-level network respectively. The shared representation layer captures the common features of different frequency bands, the policy head generates a specific allocation strategy, and the value head evaluates the allocation effect to achieve more accurate resource allocation.

[0095] Specifically, in order to achieve the goal of multi-task learning, the present invention designs an architecture with a shared representation layer (SharedRepresentation Layer), a policy head (Policy Head) and a value head (Value Head), such as Figure 2 shown.

[0096] The shared representation layer extracts and shares basic features, allowing multiple tasks or agents to use the same feature representation during the feature learning process. This sharing mechanism improves the learning efficiency and generalization ability of the model, reduces the risk of overfitting, and enables the model to not only perform well on training data, but also better adapt to new data. In the bandwidth resource allocation task, the shared representation layer helps improve task performance, capture the inherent laws of the data, and improve robustness to noise and abnormal data. In addition, it also promotes transfer learning, so that common features can be more effectively applied to new tasks or data distributions, thereby further enhancing the generalization ability of the model. The shared representation layer extracts useful features from the input state through the neural network layer. Specifically, for the input state s, the shared representation layer uses a parameterized function To extract features:

[0097]

[0098] here, represents the features extracted from the input state s, Represents the neural network parameters of the shared representation layer.

[0099] ① Shared representation layer of high-level networks: In high-level decision-making, MADDPG needs to optimize spectrum resource allocation strategies from a global perspective. The shared representation layer extracts common features from the input states of different links, which are used for high-level resource allocation decisions.

[0100]

[0101] ② Shared representation layer of the lower-level network: In the lower-level decision-making, DDPG is responsible for optimizing resource allocation within a specific frequency band. Adding a shared representation layer can help the lower-level model better understand the characteristics and patterns within the frequency band and improve the optimization effect of the specific allocation strategy. The shared representation layer extracts the characteristics within the specific frequency band, which are used to optimize resource allocation within the specific frequency band.

[0102]

[0103] Both DDPG and MADDPG use two main networks: Actor network and Critic network. The features extracted by the shared representation layer are then passed to the strategy head and value head to generate specific strategies and estimate the value of actions. The strategy head and value head are the specific implementation parts of the Actor and Critic networks. They are responsible for generating strategies (decision-making) and evaluating values ​​(evaluation) respectively.

[0104] The role of the Actor network is to generate specific actions based on the current state. After the introduction of the shared representation layer, the Actor network can use the shared feature representation to generate more consistent and coordinated action decisions. The shared representation layer can be regarded as a general feature extractor, and these features are then passed to the strategy head to generate actions, such as Figure 3 shown.

[0105] The role of the Critic network is to evaluate the value of the state-action pair. After the introduction of the shared representation layer, the Critic network can use the shared feature representation to make more accurate value assessments. The shared representation layer is also used to extract state features, which are then passed to the value head together with the action to evaluate the Q value, so that the Critic network can better evaluate the value of the current action, such as Figure 4 shown.

[0106] DDPG used for low-level network decision-making combines the advantages of deep Q networks and deterministic policy gradients (DPG). It directly outputs deterministic actions through the policy network and evaluates these actions using the Q-value network. The policy network and Q-value network are approximated using neural networks, where the policy network generates the best action a based on the current state s (generating the resource allocation strategy for a specific frequency band through the policy head), and the Q-value network evaluates the value of the state-action pair (s, a) (evaluating the value of the current action through the value head). DDPG uses an experience replay mechanism to break data correlation and improve training stability by randomly sampling from the experience (s, a, r, s') stored in the experience replay pool D. At the same time, DDPG uses a soft update method to update the target network to further improve the stability and performance of the algorithm. The network parameters of the Actor network and Critic network of DDPG are defined as θ μ and θ Q , the target network parameters are θ μ' and θ Q' In the Actor network, let μ(a|s) be the probability of taking action a in the current state s, and the Critic network gives the Q value based on s and a. Learning goal: maximize the expected sum:

[0107]

[0108] Here, γ∈(0,1) is the discount factor for the relative importance of future rewards to the current strategy.

[0109] Through the policy header Generate resource allocation strategy for specific frequency bands.

[0110] By value header Evaluate the expected total reward of the current action.

[0111] Therefore, the total expected value under strategy μ can be written as:

[0112]

[0113] β is the behavior decision. During the exploration process, (s, a, r, s') is stored in the experience replay D in the form of a four-tuple. s' is the next state that s transfers to through action a. During the training process, small batch samples are extracted from D to update the parameters.

[0114] The loss function of updating the Critic network is expressed as:

[0115]

[0116] The goal

[0117] In MADDPG, which is used for high-level network decision making, each agent has its own Actor and Critic network. The Critic can access the information of all agents during training, which helps to make better decisions in environments with other agents. First, initialize the four network parameters like DDPG. For agent m, the network parameters can be expressed as

[0118] Through the policy header Generate resource allocation strategy for specific frequency bands.

[0119] By value header Evaluate the expected total reward of the current action.

[0120] The loss function of updating the Critic network is expressed as:

[0121]

[0122] The goal

[0123] Embodiment 2:

[0124] This embodiment provides a multi-service QoS cross-band bandwidth allocation system based on deep reinforcement learning, including:

[0125] Model building module: constructs an effective capacity model, comprehensively considers key performance indicators such as transmission rate, delay and packet loss rate of different services, and clarifies the overall optimization goal of bandwidth resources;

[0126] The architecture design module uses the effective capacity model as the basis to build a multi-task hierarchical architecture based on deep reinforcement learning. This architecture makes a preliminary allocation of the overall bandwidth resources of the Sub-6GHz and millimeter wave bands through the high-level network to ensure the reasonable allocation of bandwidth resources across frequency bands. It then makes a refined allocation of specific bandwidth resources within each frequency band through the low-level network to meet the QoS requirements of different services.

[0127] The allocation module introduces a shared representation layer, a policy head, and a value head in the high-level network and the low-level network, respectively. The shared representation layer captures the common characteristics of different frequency bands, the policy head generates a specific allocation strategy, and the value head evaluates the allocation effect to achieve more accurate resource allocation.

[0128] Embodiment three:

[0129] An electronic device includes a memory, a processor, and a computer program stored and running on the memory, wherein the processor implements the above-mentioned multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning when executing the program, including:

[0130] Build an effective capacity model, comprehensively consider the key performance indicators of different services, and clarify the overall optimization goal of bandwidth resources;

[0131] Based on the effective capacity model, a multi-task hierarchical architecture based on deep reinforcement learning is constructed. This architecture makes a preliminary allocation of the overall bandwidth resources of the Sub-6GHz and millimeter wave bands through the high-level network to ensure the reasonable allocation of bandwidth resources across the frequency bands. The specific bandwidth resources within each frequency band are then finely allocated through the low-level network to meet the QoS requirements of different services.

[0132] A shared representation layer, a policy head and a value head are introduced into the high-level network and the low-level network respectively. The shared representation layer captures the common characteristics of different frequency bands, the policy head generates specific allocation strategies, and the value head evaluates the allocation effect to achieve more accurate resource allocation.

[0133] Embodiment 4:

[0134] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning, including:

[0135] Build an effective capacity model, comprehensively consider the key performance indicators of different services, and clarify the overall optimization goal of bandwidth resources;

[0136] Based on the effective capacity model, a multi-task hierarchical architecture based on deep reinforcement learning is constructed. This architecture makes a preliminary allocation of the overall bandwidth resources of the Sub-6GHz and millimeter wave bands through the high-level network to ensure the reasonable allocation of bandwidth resources across the frequency bands. The specific bandwidth resources within each frequency band are then finely allocated through the low-level network to meet the QoS requirements of different services.

[0137] A shared representation layer, a policy head and a value head are introduced into the high-level network and the low-level network respectively. The shared representation layer captures the common characteristics of different frequency bands, the policy head generates specific allocation strategies, and the value head evaluates the allocation effect to achieve more accurate resource allocation.

[0138] Before implementing the above multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning, the basic parameters of the wireless communication system are set, including frequency band division, user demand, channel status, etc. The parameters of the multi-task hierarchical architecture of deep reinforcement learning are initialized, including the weights of high-level and low-level policy networks.

[0139] It should be noted that the present invention designs the high-level reward function as a global effective capacity indicator, aiming to optimize the total transmission rate of the system. The low-level reward function is designed as the average effective capacity of a specific link and neighboring links. Using the experience replay mechanism, historical data is randomly sampled for model training to optimize the policy network and value network. A soft update mechanism is used to update the target network to enhance the stability and performance of the algorithm. In each time slot, bandwidth resources are preliminarily allocated according to the high-level strategy, and then the specific frequency band is finely adjusted through the low-level strategy.

[0140] Taking a specific scenario as an example, suppose that multiple users request different types of services (such as video streaming, online games, etc.) at the same time. The steps are as follows:

[0141] 1) User demand analysis: Collect each user’s QoS requirements (such as bandwidth and latency requirements).

[0142] 2) Preliminary resource allocation: The high-level strategy preliminarily allocates bandwidth resources based on user demand and channel status.

[0143] 3) Refined allocation: The low-level strategy further adjusts the bandwidth allocation of each frequency band to meet the QoS requirements of specific users.

[0144] 4) Performance evaluation: Monitor network performance in real time, evaluate effective capacity and latency, and optimize strategies based on feedback.

[0145] In summary, the present invention can efficiently implement cross-band intelligent bandwidth resource management and is applicable to various application scenarios of modern wireless communication systems.

[0146] Those skilled in the art should understand that the modules or steps of the present disclosure can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present disclosure is not limited to any specific combination of hardware and software.

[0147] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0148] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning, characterized in that: The following steps are involved: Build an effective capacity model, comprehensively consider the key performance indicators of different services, and clarify the overall optimization goal of bandwidth resources; Based on the effective capacity model, a multi-task hierarchical architecture based on deep reinforcement learning is constructed. This architecture makes a preliminary allocation of the overall bandwidth resources of the Sub-6GHz and millimeter wave bands through the high-level network to ensure the reasonable allocation of bandwidth resources across the frequency bands. The specific bandwidth resources within each frequency band are then finely allocated through the low-level network to meet the QoS requirements of different services. A shared representation layer, a policy head and a value head are introduced into the high-level network and the low-level network respectively. The shared representation layer captures the common characteristics of different frequency bands, the policy head generates specific allocation strategies, and the value head evaluates the allocation effect to achieve more accurate resource allocation.

2. According to claim 1, the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is characterized in that: The effective capacity model is constructed as follows: The low-frequency base station LF-BS below 6GHz and the millimeter-wave high-frequency base station HF-BS are integrated to realize high- and low-frequency collaborative networking; the signal-to-interference-to-noise ratio at the receiver m of the low-frequency base station LF-BS is expressed as: in, is the power gain of the Sub-6GHz low-frequency cellular antenna, N s is the additive white Gaussian noise power spectral density of cellular spectrum resources, B s is the bandwidth of cellular user spectrum resources; p m represents the transmission power on link m, h m is the channel gain from user m to LF-BS, is the sum of interferences from other links to link m; The signal-to-interference-to-noise ratio at the receiver m of the high-frequency base station HF-BS is expressed as: in, and are the directional antenna gains of the HF-BS transmitting and receiving users respectively; N mv is the additive white Gaussian noise power spectral density of millimeter wave spectrum resources, B mv The bandwidth of spectrum resources for millimeter wave users; Effective capacity describes the average data transmission rate that can be achieved by meeting the QoS requirements of different services under given wireless channel conditions: Where T represents the time slot length, T D Indicates the packet delay, set t delay is the maximum tolerable delay of the data packet, R is the link transmission rate, ε is the maximum allowed packet loss rate; θ is the QoS guarantee parameter, covering the key performance indicators of transmission rate, delay, and packet loss rate, and their relationship is expressed as follows: Therefore, the effective capacity model of LF-BS and HF-BS is: in, is the delay QoS constraint parameter of the link; α m represents the communication mode selected by the link, α m =0 means LS-BS communication is selected, α m =1 means HF-BS communication is selected, 0<α m <1 indicates the ratio of mmWave bandwidth allocated to HF-BS on the link.

3. According to claim 2, the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is characterized in that: The data overflow / loss constraint is introduced into the effective capacity model. Specifically, the probability of data packet loss on link m is expressed as: Among them, T m represents the data delay of link m, Q m represents the cache queue length of link m, represents the maximum allowable data delay of service b, Indicates the maximum allowable cache queue length for service b, data transmission exceeds the maximum delay threshold or the data cached by the node exceeds the maximum queue length threshold When , data is lost; l(λ m ) represents the probability that the link m cache is not empty: Therefore, the constraint condition of the data overflow / loss probability of the link is expressed as: Among them, ε b Indicates the maximum data overflow / loss probability threshold allowed by service b.

4. According to claim 1, the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is characterized in that: Taking the maximization of link effective capacity as the overall optimization goal of bandwidth resources, the cross-band bandwidth allocation problem of multi-service QoS is solved. The problem is based on the overflow / loss probability of transmission delay exceeding the service delay threshold, link capacity exceeding the data arrival rate, and minimum communication quality guarantee as constraints: in, and They represent link m in Sub6-GHz bandwidth resource block n. s and bandwidth resource block n on mmWave mv The allocation matrix on θ = [θ m ] is the QoS guarantee parameter configuration matrix.

5. According to claim 1, the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is characterized in that: The state space, action space and reward function are defined separately in the multi-task hierarchical architecture based on deep reinforcement learning; Its state space S includes: the receiver's demand matrix θ m , instantaneous channel state information h m , the tolerable delay threshold of the current service b of the link transmitter and the maximum acceptable data overflow / loss probability The instantaneous traffic arrival rate of the link is m ; The state space of high-level network decision making is the same as that of low-level network decision making, but the low-level network is refined to the state of specific links and frequency bands; Its action space A includes: The action space of the high-level network is defined as the total bandwidth resource allocation strategy of the link in different frequency bands, ρ m is the allocation ratio of Sub6-GHz bandwidth resources of LF-BS, α m represents the allocation ratio of millimeter wave bandwidth resources of HF-BS; the action space of the low-level network is expressed as A = {ρ, α}, which independently optimizes the specific resource allocation of each frequency band to adapt to the characteristics and requirements of different frequency bands; Its reward function R includes: The instantaneous reward of the tth time slot of the high-level network is expressed as: The instantaneous reward of the tth time slot of the low-level network is expressed as: in, is the sum of the number of the current agent and the observed neighbor agents.

6. According to claim 1, the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is characterized in that: For the input state s, the shared representation layer is parameterized by the function To extract features: in, represents the features extracted from the input state s, Neural network parameters representing the shared representation layer; The shared representation layer of the high-level network extracts common features from the input states of different links. These features are used for high-level resource allocation decisions: The shared representation layer of the lower-level network extracts features within a specific frequency band, which are used to optimize resource allocation within the specific frequency band.

7. The multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning according to claim 1 or 6, characterized in that: The features extracted by the shared representation layer are passed to the strategy head and the value head. The strategy head and the value head correspond to the Actor network and the Critic network in deep reinforcement learning, which are responsible for generating strategies and evaluating values ​​respectively. The DDPG algorithm is used in the low-level network, and the network parameters of the Actor network and the Critic network are defined as θ μ and θ Q , the target network parameters are θ μ ′ and θ Q′ ; In the Actor network, let μ(a|s) be the probability of taking action a in the current state s, and the Critic network gives the Q value based on s and a; the learning goal is to maximize the expected sum: Among them, γ∈(0,1) is the discount factor of the relative importance of future rewards to the current strategy; Through the policy header Generate resource allocation strategy for specific frequency bands; By value header Evaluate the expected total reward of the current action; The total expected value under strategy μ can be written as: Among them, β is the behavior decision, which is stored in the experience replay D in the form of a four-tuple during the exploration process (s, a, r, s'), and s' is the next state that s transfers to through action a; The loss function of updating the Critic network is expressed as: The goal The MADDPG algorithm is used in the high-level network. Each agent has a corresponding Actor network and Critic network. The network parameters for agent m are expressed as Through the policy header Generate resource allocation strategy for specific frequency bands; By value header Evaluate the expected total reward of the current action; The loss function of updating the Critic network is expressed as: The goal 8. A multi-service QoS cross-band bandwidth allocation system based on deep reinforcement learning, characterized in that: include: Model building module: constructs an effective capacity model, comprehensively considers key performance indicators such as transmission rate, delay and packet loss rate of different services, and clarifies the overall optimization goal of bandwidth resources; The architecture design module uses the effective capacity model as the basis to build a multi-task hierarchical architecture based on deep reinforcement learning. This architecture makes a preliminary allocation of the overall bandwidth resources of the Sub-6GHz and millimeter wave bands through the high-level network to ensure the reasonable allocation of bandwidth resources across frequency bands. It then makes a refined allocation of specific bandwidth resources within each frequency band through the low-level network to meet the QoS requirements of different services. The allocation module introduces a shared representation layer, a policy head, and a value head in the high-level network and the low-level network, respectively. The shared representation layer captures the common characteristics of different frequency bands, the policy head generates a specific allocation strategy, and the value head evaluates the allocation effect to achieve more accurate resource allocation.

9. An electronic device comprising a memory, a processor and a computer program stored and running on the memory, characterized in that: When the processor executes the program, the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning is implemented as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the multi-service QoS cross-band bandwidth allocation method based on deep reinforcement learning as described in any one of claims 1-7.

Citation Information

Patent Citations

  • Multi-band network resource allocation method based on ML-MADDPG

    CN120475414A

  • Intelligence and Learning in O-RAN for 5G and 6G Cellular Networks

    US20220167236A1

Cited By

  • Port multi-agent task allocation method and system based on reinforcement learning

    CN119273104A

  • A port multi-agent task allocation method and system based on reinforcement learning

    CN119273104B

  • Data transmission optimization method and system for WiFi (Wireless Fidelity) router

    CN120166430A

  • Intelligent roaming decision-making system and method and intelligent screen

    CN120730295A

  • Method and system for allocating edge computing resources of electromagnetic safety driving data

    CN121568175A