Method and apparatus for power controlling for network energy optimization

The method addresses energy inefficiency in ultra-dense networks by using AI-based deep reinforcement learning to optimize base station power control, enhancing energy efficiency and QoS satisfaction.

KR103044482B1Active Publication Date: 2026-11-06SAMSUNG ELECTRONICS CO LTD +1
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
KR1020200062499
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-05-25
Publication Date
2026-11-06
Estimated Expiration
2040-05-25

AI Technical Summary

Technical Problem

Ultra-dense networks face challenges such as severe inter-cell interference and low energy efficiency due to densely deployed small base stations with low or no traffic load, necessitating improved power control methods.

Method used

A base station power control method using artificial intelligence, specifically deep reinforcement learning, to determine active/sleep modes based on channel state information and QoS indicators, reducing computational complexity by involving a central unit to simplify the action space.

Benefits of technology

Maximizes network energy efficiency while satisfying QoS requirements by optimizing base station power consumption through intelligent mode switching, leveraging deep reinforcement learning to reduce computational complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure R1020200062499_ABST
    Figure R1020200062499_ABST
Patent Text Reader

Abstract

The present disclosure relates to a base station power control technique using artificial intelligence-based technology for improving the energy efficiency of a communication network. A method performed by a base station of a communication system according to one embodiment of the present disclosure may include: acquiring status information; determining an ASRI (active / sleep request indicator) based on at least a portion of the status information; transmitting at least one of the status information and the ASRI to a central unit; receiving power control information determined based on at least one of the status information and the ASRI from the central unit; and performing power control based on the received power control information.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present disclosure relates to a base station power control technique using artificial intelligence-based technology for improving the energy efficiency of a communication network. Background Technology

[0002] Efforts are being made to develop improved 5G or pre-5G communication systems to meet the increasing demand for wireless data traffic since the commercialization of 4G communication systems. For this reason, 5G or pre-5G communication systems are referred to as Beyond 4G Network communication systems or Post-LTE systems. To achieve high data transmission rates, the implementation of 5G communication systems in the mmWave band (e.g., the 60 GHz band) is being considered. To mitigate path loss and increase transmission distance in the mmWave band, technologies such as beamforming, massive MIMO, full Dimensional MIMO (FD-MIMO), array antennas, analog beamforming, and large-scale antennas are being discussed for 5G communication systems. In addition, to improve the network of the system, the development of technologies such as advanced small cell, advanced small cell, cloud radio access network (cloud RAN), ultra-dense network, Device to Device communication (D2D), wireless backhaul, moving network, cooperative communication, CoMP (Coordinated Multi-Points), and interference cancellation is taking place in 5G communication systems.In addition, advanced coding modulation (ACM) methods such as FQAM (Hybrid FSK and QAM Modulation) and SWSC (Sliding Window Superposition Coding), as well as advanced access technologies such as FBMC (Filter Bank Multi Carrier), NOMA (non-orthogonal multiple access), and SCMA (sparse code multiple access), are being developed in 5G systems.

[0003] Meanwhile, the Internet is evolving from a human-centric network where humans generate and consume information into an IoT (Internet of Things) network that processes information by exchanging it among distributed components, such as objects. IoE (Internet of Everything) technology, which combines IoT with Big Data processing technologies through connections with cloud servers, is also emerging. To implement IoT, technological elements such as sensing technology, wired and wireless communication and network infrastructure, service interface technology, and security technology are required; consequently, technologies such as sensor networks, Machine-to-Machine (M2M) communication, and Machine-Type Communication (MTC) are currently being researched to facilitate the connection of objects. In an IoT environment, intelligent IT services that create new value for human life by collecting and analyzing data generated from connected objects can be provided. Through the convergence and integration of existing IT technologies with various industries, IoT can be applied to fields such as smart homes, smart buildings, smart cities, smart or connected cars, smart grids, healthcare, smart home appliances, and advanced medical services.

[0004] Accordingly, various attempts are being made to apply 5G communication systems to IoT networks. For example, technologies such as sensor networks, Machine to Machine (M2M), and Machine Type Communication (MTC) are being implemented using 5G communication techniques such as beamforming, MIMO, and array antennas. The application of cloud RAN as a big data processing technology, as previously described, can also be considered an example of the convergence of 5G and IoT technologies.

[0005] As such, in 5G communication systems and the upcoming 6G communication systems, ultra-dense networks—which involve densely deploying small base stations—are attracting attention as a means to improve network capacity due to the rapid increase in the number of smartphones and Internet of Things (IoT) devices and data demands. While ultra-dense networks have the advantage of alleviating the burden on macro base stations while simultaneously improving network capacity, problems may arise such as severe inter-cell interference during base station placement due to the dense spacing between base stations, and low energy efficiency resulting from a large number of base stations with low or almost no traffic load. Consequently, technology is needed to address these issues. The problem to be solved

[0006] The present invention proposes a method for deactivating base stations with low traffic loads, i.e., switching them to sleep mode, in order to increase energy efficiency in ultra-dense networks. Additionally, it proposes a base station power control method that minimizes network power consumption while satisfying the terminal's QoS. means of solving the problem

[0007] A method performed by a base station of a communication system according to one embodiment of the present disclosure for achieving the aforementioned technical problem may include: acquiring status information; determining an ASRI (active / sleep request indicator) based on at least a portion of the status information; transmitting at least one of the status information and the ASRI to a central unit; receiving power control information determined based on at least one of the status information and the ASRI from the central unit; and performing power control based on the received power control information.

[0008] According to one embodiment, the status information may include at least one of channel status information with the terminal and the terminal's QoS (Quality of Service) indicator.

[0009] According to one embodiment, the ASRI may indicate at least one of active, sleep, and not decided.

[0010] According to one embodiment, when the ASRI indicates active or sleep, the power control information may be determined based on the ASRI.

[0011] According to one embodiment, if the ASRI indicates not decided, the power control information may be determined based on the state information.

[0012] Additionally, a method performed by a central unit of a communication system according to one embodiment of the present disclosure may include: receiving at least one of status information and an ASRI (active / sleep request indicator) from the base station; determining power control information for the base station based on at least one of the status information and ASRI received from the base station; and transmitting the power control information to the base station.

[0013] In addition, a base station of a communication system according to one embodiment of the present disclosure may include: a transceiver; and a control unit configured to acquire status information, determine an ASRI (active / sleep request indicator) based on at least a portion of the status information, transmit at least one of the status information and the ASRI to a central unit, receive power control information determined based on at least one of the status information and the ASRI from the central unit, and perform power control based on the received power control information.

[0014] Additionally, a central unit of a communication system according to one embodiment of the present disclosure may include: a transceiver; and a control unit configured to receive at least one of status information and an ASRI (active / sleep request indicator) from a base station, determine power control information for the base station based on at least one of the status information and ASRI received from the base station, and transmit the power control information to the base station. Effects of the invention

[0015] The present invention can maximize the energy efficiency of a network while satisfying the QoS of a terminal through base station power control in an ultra-dense network environment where base stations consume a large amount of power. Furthermore, according to the present invention, the computational complexity of a device performing power control can be significantly reduced by using an artificial neural network, and an optimal base station power control method can be provided through an active / sleep mode request signal transmitted by the base station. Brief explanation of the drawing

[0016] FIG. 1 is a diagram showing the structure of a communication network according to one embodiment of the present disclosure. FIG. 2 is a diagram illustrating deep reinforcement learning according to one embodiment of the present disclosure. FIG. 3 is a drawing for explaining a power control method in a communication network according to one embodiment of the present disclosure. FIG. 4 is a flowchart showing the operation of a base station according to one embodiment of the present disclosure. FIG. 5 is a flowchart showing the operation of a central unit according to one embodiment of the present disclosure. FIG. 6 is a flowchart illustrating a power control method in a communication network according to one embodiment of the present disclosure. FIG. 7 is a flowchart illustrating a power control method in a communication network according to another embodiment of the present disclosure. FIG. 8 is a diagram showing the configuration of a base station according to one embodiment of the present disclosure. FIG. 9 is a drawing showing the configuration of a central unit according to one embodiment of the present disclosure. Specific details for implementing the invention

[0017] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Furthermore, in describing the present invention, detailed descriptions of related known functions or configurations are omitted if it is determined that such detailed descriptions would unnecessarily obscure the essence of the invention. Additionally, the terms described below are defined considering their functions in the present invention, and these may vary depending on the intentions or conventions of the user or operator. Therefore, their definitions should be based on the content throughout this specification.

[0018] In addition, in specifically describing the embodiments of the present invention, the main gist of the present invention can be applied to other communication systems having similar technical backgrounds and channel types with slight modifications without significantly departing from the scope of the present invention, and this will be possible at the judgment of a person with skilled technical knowledge in the technical field of the present invention.

[0019] The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below but can be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Throughout the specification, the same reference numerals refer to the same components.

[0020] At this time, it will be understood that each block of the process flow diagrams and combinations of the flow diagrams can be executed by computer program instructions. Since these computer program instructions can be loaded into the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing equipment, the instructions executed through the processor of the computer or other programmable data processing equipment create means to perform the functions described in the flow diagram block(s). Since these computer program instructions can also be stored in computer-available or computer-readable memory that can be directed toward the computer or other programmable data processing equipment to implement the function in a specific way, the instructions stored in computer-available or computer-readable memory can also produce a manufactured item containing the means of instruction to perform the function described in the flow diagram block(s). Since computer program instructions can be loaded onto a computer or other programmable data processing equipment, instructions that perform a series of operation steps on the computer or other programmable data processing equipment to create a process executed by the computer can also provide steps for executing the functions described in the flowchart block(s).

[0021] Additionally, each block may represent a module, segment, or part of code containing one or more executable instructions for executing a specified logical function(s). It should also be noted that in some alternative execution examples, the functions mentioned in the blocks may occur out of order. For instance, two blocks described in succession may actually be executed substantially simultaneously, or the blocks may be executed in reverse order according to their corresponding functions.

[0022] In this embodiment, the term "part" refers to a software or hardware component, such as an FPGA or ASIC, and the "part" performs certain roles. However, the meaning of "part" is not limited to software or hardware. The "part" may be configured to reside in an addressable storage medium or configured to operate one or more processors. Accordingly, as an example, the "part" includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functions provided within the components and "parts" may be combined into a smaller number of components and "parts" or further separated into additional components and "parts." Furthermore, the components and "parts" may be implemented to operate one or more CPUs within a device or secure multimedia card.

[0023] Terms used in the following description to refer to signals, channels, control information, network entities, device components, nodes, messages, interfaces between network entities, and various identification information are examples provided for the convenience of explanation. Accordingly, this disclosure is not limited to the terms described below, and other terms referring to objects having equivalent technical meanings may be used.

[0024] With the recent surge in the number of smartphones and Internet of Things (IoT) devices and the resulting increase in data demand, ultra-dense networks—which involve the dense deployment of small base stations—are gaining attention as a means to enhance network capacity. While ultra-dense networks offer the advantage of improving network capacity while alleviating the burden on macro base stations, they present challenges such as severe inter-cell interference due to the dense spacing between base stations, and reduced energy efficiency resulting from the high number of base stations with low or almost no traffic load.

[0025] To improve energy efficiency in ultra-dense networks, a technique can be utilized to deactivate base stations with low traffic loads, i.e., switch them to sleep mode. In existing 4G LTE, networks were structured around macro base stations responsible for data and synchronization / control channel transmission, which limited the application of sleep mode switching techniques to them. However, in 5G NR and the upcoming 6G mobile communication, numerous small base stations are densely distributed around macro base stations. Since the coverage and number of connected terminals of these small base stations are limited, sleep mode switching techniques can be applied to small base stations with low traffic loads or no connected terminals. Conventional sleep mode techniques include heuristic methods and optimization-based methods. Heuristic methods apply sleep mode in order of decreasing transmission power or traffic load. However, these methods have disadvantages, such as poor energy efficiency due to frequent mode transitions, high computational complexity in ultra-dense networks, and difficulty in finding the optimal combination of base stations. Optimization-based techniques must solve the NP-hard problem of Mixed Integer Programming to determine the base station active / sleep mode. Although a solution can be obtained through exhaustive search, the computational complexity of the problem increases exponentially with the number of base stations in the network, resulting in a long time required for the decision and making it practically impossible to respond flexibly to a rapidly changing communication environment. In this invention, an optimal base station power control technique is proposed based on signals such as downlink channel state information between the base station and the terminal and the terminal's QoS indicator.By using machine learning, the complex relationship between input signals and power control that maximizes energy efficiency can be effectively learned based on data.

[0026] FIG. 1 is a diagram showing the structure of a communication network according to one embodiment of the present disclosure.

[0027] Referring to FIG. 1, a communication network (100) according to one embodiment of the present invention may be composed of a central unit (110) that performs power control for each base station (120a, 120b, 120c), and one or more base stations (120a, 120b, 120c) that operate in an active / sleep mode according to the control of the central unit (110) and transmit and receive signals to and from terminals within the coverage area.

[0028] The central unit (110) may be defined as any device equipped with a function or device for performing power control of a plurality of base stations (120a, 120b, 120c). For example, the central unit (110) may be implemented as a device existing independently of the base station as shown in FIG. 1 to perform power control for a plurality of base stations (120a, 120b, 120c), or it may be implemented within a macro base station (120a) connected to a plurality of small base stations (120b, 120c), although not shown in FIG. 1, to perform power control for each small base station (120b, 120c). According to one embodiment, a small base station (120b) may be connected to the central unit (110) through a macro base station (120a), or a small base station (120c) may be directly connected to the central unit (110). A plurality of base stations (120a, 120b, 120c) may include a macro base station (120a) or a small base station (120b, 120c) connected to a plurality of small base stations. In FIG. 1, for convenience of explanation, the communication network according to the present invention is illustrated to include both a macro base station (120a) and a small base station (120b, 120c) connected thereto, but this does not limit the scope of the present invention. In the following specification, power control performed by the central unit (110) means determining an active / sleep mode for each base station (120a, 120b, 120c) controlled by the central unit (110), or determining any condition (e.g., number of antennas used, transmission power level, number of available resource blocks, etc.) that can be used for power control of the base station determined to be in active mode.

[0029] In the communication network illustrated in FIG. 1, the problem of maximizing the energy efficiency of the network while satisfying certain conditions, such as the QoS of each terminal and the power conditions of the base station, through power control by the central unit (110) can be configured as an optimization problem based on mathematical formulas. As an example, it can be configured as a model to minimize the power consumption of the network under the condition that the data transmission rate received by each terminal satisfies the data requirements of the terminal. The power consumption of the base station can be composed of the total transmission power for the terminals connected to each base station, the maintenance power consumed in active / sleep mode respectively, and the transition power required when switching between active / sleep mode. The indices of the base station and the terminal, respectively , When referring to, the time unit Dog (e.g. The problem of minimizing base station power consumption during seconds can be expressed as Equation 1 below.

[0030] [Mathematical Formula 1]

[0031]

[0032]

[0033] Here The base station It means the sleep(0) or active(1) mode, The terminal It refers to the data transmission rate requirement. It is the sum of the power consumption of all base stations (e.g., transmit power, sustain power, switching power, etc.) is a terminal It refers to the received data transmission rate and can be determined by channel information between the base station and the terminal and transmission power. For example, assuming that a certain amount of power is allocated to each terminal when transmitting a signal to a base station, the received data capacity is can be applied. Here refers to the frequency band of the signal, and represents the signal-to-interference-plus-noise ratio. When transmitting data to terminals connected to a base station, each It can be expressed as a function of channel information between the base station and each terminal.

[0034] The present invention proposes a power control method utilizing artificial intelligence-based technology. To achieve the technical objectives proposed in the present invention, applicable artificial intelligence-based technology may be any technology including Convolutional Neural Network (CNN), Long-Term Short Memory (LSTM), Deep Reinforcement Learning (DRL), etc. In the following, the present invention discloses a method utilizing deep reinforcement learning (DRL) as an example, but this does not limit the scope of the present invention.

[0035] FIG. 2 is a diagram illustrating deep reinforcement learning according to one embodiment of the present disclosure.

[0036] Deep reinforcement learning is an algorithm that finds the best policy to achieve a goal based on trial and error, and the current state When given, the agent (210) performs an action Perform and, as a result of the action, receive a reward through the environment (220). and next state You can obtain. In this case, the optimal policy that maximizes the expected value of cumulative rewards ( ) can be expressed as in Mathematical Equation 2 below. Here is the discount factor( It represents the weight of future rewards relative to current rewards.

[0037] [Mathematical Formula 2]

[0038]

[0039] Deep reinforcement learning can obtain a policy that maximizes cumulative rewards by training a deep neural network within an agent using data (state, action, and reward) acquired in real time. For example, in the case of Q-learning, a reinforcement learning algorithm, the Q-values ​​of state-action pairs, expressed as Equation 3 below, must be known as indicators for action decisions.

[0040] [Mathematical Formula 3]

[0041]

[0042] To obtain the Q-value according to Equation 3, iterative updates must be performed using actions and the corresponding rewards; consequently, if the computational complexity of the learning is high and the difficulty of the task to be performed is high, it is very difficult to find the Q-values ​​for all possible state-action pairs through Q-learning. To solve this problem, a network capable of approximating the Q-value (deep Q-network: DQN) is used by incorporating deep neural networks. The parameters within the DQN are the loss function (L( It can be learned in a direction that minimizes )). of the mathematical formula refers to rewards based on a given state and behavior, Each represents the next state and next action.

[0043] [Mathematical Formula 4]

[0044]

[0045] In the present invention, the state information used by the central unit to determine power control (active / sleep mode) for each base station at time unit t is the state , and based on this, the central unit determines the power control method (active / sleep mode) for each base station and acts upon the generated power control information. It can be defined as.

[0046] According to one embodiment, the state information used by a central unit to determine power control (active / sleep mode) for each base station may include at least one of channel state information (CSI) between the base station and the terminal, the terminal's Quality of Service indicator, User perceived throughput, the amount of remaining traffic in the base station queue, the data fulfillment rate of each terminal to date, resource block scheduling information for terminals within the base station coverage, the intensity of interference received from a neighboring base station, the intensity of interference given to a neighboring base station, and traffic load. For example, state information It may include channel status information between the base station and the terminal and the QoS indicator of the terminal, which can be expressed mathematically as follows.

[0047] [Mathematical Formula 5]

[0048]

[0049]

[0050] For convenience of explanation, the state information has been defined as including the CSI and the QoS indicator of the terminal; however, this does not limit the present invention, and any information included in the aforementioned state information , It can be defined as a mathematical parameter through a method similar to, and state information It can also be defined to correspond to that.

[0051] In one embodiment, power control information for each base station determined based on state information, i.e., action can be defined to correspond to 1 when each base station is determined to be in active mode and 0 when determined to be in sleep mode. This can be expressed mathematically as follows.

[0052] [Mathematical Formula 6]

[0053]

[0054] In this case, action The number of cases considered to determine, that is, the size of the action space (A) It could be.

[0055] In one embodiment, power control information, i.e., behavior In addition to determining the active / sleep mode, the number of antennas (A) to be used by the base station set to active mode, the number of transmission power levels (T), and the number of allocated resource blocks (B), or a combination thereof, can be determined by considering any one of these as additional variables, and each case can be expressed mathematically as follows.

[0056] [Mathematical Formula 7]

[0057]

[0058]

[0059]

[0060] In each case, when C is the number of base stations determined in active mode The number of cases considered to determine, that is, the size of the action space (A) , , It could be.

[0061] In one embodiment, compensation is action As a result of the operation, the network's power consumption is lower, and the value is larger; it can be configured to impose a penalty if specific conditions required by the base station and terminal (e.g., terminal's data requirements, QoS, etc.) are not satisfied. An example of this can be expressed mathematically as follows.

[0062] [Mathematical Formula 8]

[0063]

[0064] is the maximum power used by the network when all base stations are in active mode, represents the power actually consumed at base station m. According to Equation 8, The reward has a higher value as the sum of the power actually consumed by the base stations constituting the network decreases, and if certain constraints are not satisfied, a penalty of -K may be imposed. Equation 8 is an example for mathematically defining the reward and can be defined in any form to induce behavior that maximizes energy efficiency while satisfying specific requirements.

[0065] The above describes a power control method based on deep reinforcement learning. However, when determining power control (active / sleep mode) for each base station based on the previously described method, the number of cases to be considered (size of the action space) increases exponentially as the number of base stations increases, which may lead to increased computational complexity and reduced learning efficiency.

[0066] Accordingly, the present invention proposes a method for reducing the size of the action space according to one embodiment, wherein each base station independently determines whether to operate in active mode or sleep mode in the next time unit and requests this from a central unit, and the central unit determines power control information for each base station by taking this into consideration.

[0067] FIG. 3 is a drawing for explaining a power control method in a communication network according to one embodiment of the present disclosure.

[0068] Referring to FIG. 3, the base station (310) can transmit information to the central unit (320) for the central unit (320) to determine power control information for the base station (310). At this time, FIG. 3 is illustrated as the base station (310) being directly connected to the central unit (320) to transmit information, but as described in FIG. 1, if the base station (310) corresponds to a small base station connected to the central unit (320) through a macro base station, such status information may be transmitted to the central unit (320) through the macro base station. Thus, the information transmitted by the base station (310) to the central unit (320) may include status information indicating the status of the base station (310) and terminals included in the coverage of the base station (310), and an active / sleep request indicator (ASRI) determined by the base station (310). Information transmitted from the base station (310) to the central unit (320) can be transmitted, for example, through a backhaul network.

[0069] According to one embodiment, the status information may include at least one of channel state information (CSI) between the base station and the terminal, the terminal's Quality of Service indicator (QoS indicator), User perceived throughput, the amount of remaining traffic in the base station queue, the data fulfillment rate of each terminal to date, resource block scheduling information for terminals within the base station coverage, the intensity of interference received from a neighboring base station, the intensity of interference given to a neighboring base station, and traffic load. Each of the information included in the status information may be obtained through any method as needed, such as by the base station (310) receiving from the terminal, by the base station (310) acquiring it independently, or by receiving it from a macro base station or a central unit (320). For example, the status information may include channel state information between the base station and the terminal and the terminal's QoS indicator.

[0070] According to one embodiment, an active / sleep request indicator (ASRI) is generated by a base station (310) and may be information indicating one of active, sleep, or not decided. According to one embodiment, the ASRI is generated based on the independent judgment of each base station (310) and transmitted to a central unit (320), and may be a signal requesting an active / sleep mode for the base station (310) in the next time unit or a specific time unit. If the base station (310) wishes to request the central unit (320) to control the base station (310) in an active mode in the next time unit or a specific time unit, the base station (310) may generate an ASRI indicating “active” and transmit it to the central unit (320). Alternatively, if the base station (310) wishes to request the central unit (320) to control the base station (310) in sleep mode at the next time unit or a specific time unit, the base station (320) may generate an ASRI indicating “sleep” and transmit it to the central unit (320). Alternatively, if the base station (310) finds it difficult to clearly determine on its own the necessity of active / sleep mode operation, the base station (310) may generate an ASRI indicating “not decided” and transmit it to the central unit (320).

[0071] In one embodiment, the base station (310) may generate an ASRI by determining the necessity of active mode / sleep mode operation based on at least some of the state information. The state information may include at least one of channel state information (CSI) between the base station and the terminal, the terminal's Quality of Service indicator (QoS indicator), User perceived throughput, the amount of remaining traffic in the base station queue, the data fulfillment rate of each terminal to date and resource block scheduling information for terminals within the base station coverage, the intensity of interference received from neighboring base stations, the intensity of interference given to neighboring base stations, and traffic load, and the base station (310) may generate an ASRI based on at least some or all of these. As an example, the base station (310) may generate an ASRI based on at least one of the signal and terminal scheduling related variables, such as the QoS indicator, User perceived throughput, the amount of remaining traffic in the base station queue, the data fulfillment rate of each terminal to date and resource block scheduling information for terminals within the small base station coverage.

[0072] In one embodiment, the ASRI may be generated proactively by each base station (310) and transmitted to the central unit (320) periodically or non-periodically. Alternatively, the ASRI may be generated and transmitted periodically or non-periodically by each base station (310) in response to a request from the central unit (320). Alternatively, the ASRI may be generated proactively by each base station (310) and transmitted to the central unit (320) periodically or non-periodically, and may also be generated and transmitted periodically or non-periodically by each base station (310) upon a request from the central unit (320).

[0073] If the base station (310) corresponds to a small base station connected to a macro base station, and it is determined that the data demand of terminals within the coverage area of ​​the base station (310) is high and the satisfaction rate is low, making it difficult to satisfy the QoS of the terminals using only the macro base station connected to the base station (310), an ASRI indicating “active” can be generated. Conversely, if it is determined that the data demand of terminals within the coverage area of ​​the base station (310) is low and the satisfaction rate is high, making it possible to satisfy the QoS of terminals within the coverage area of ​​the base station (310) using only the macro base station connected to the base station (310), an ASRI indicating “sleep” can be generated. Additionally, if the base station (310) cannot clearly determine whether it is active or sleep on its own, a “not decided” signal can be generated.

[0074] The central unit (320) may include a power control unit (322) that generates power control information based on status information, compensation, ASRI, etc. received from the base station (310), and an action space simplification unit (324). In one embodiment, the action space simplification unit (324) may perform the role of an action space simplifier to reduce the size of the action space that the power control unit (322) must consider when generating power control information, thereby providing a reduction in computational complexity and an improvement in performance.

[0075] The power control unit (322) can generate power control information based on at least one of the status information, compensation, ASRI, and information received from one or more base stations and the action space simplification unit (324). The power control information may be information indicating an active / sleep mode for each base station (310), or information indicating one or more of the number of antennas to be used (A), the number of transmission power levels (T), and the number of resource blocks (B) allocated for a base station indicated in active mode. The power control unit (322) can generate power control information for each base station by determining the action for which the largest Q-value is expected based on at least one of the status information, compensation, ASRI, and information received from the action space simplification unit (324) according to the method described above with reference to FIG. 2.

[0076] Meanwhile, the central unit (320) can reduce the action space based on the ASRI received from the base station (310) when generating power control information. As an example, the ASRI received from the base station (310) can form an action space consisting of a base station that indicates “not decided”. In this case, for base stations that determine and transmit an ASRI indicating “active,” the base station indicates “active,” and for base stations that determine and transmit an ASRI indicating “sleep,” the base station indicates “sleep.” In this case, since the central unit (320) does not necessarily need to perform calculations for determining the active / sleep mode based on state information for the base stations, it excludes them from the action space to be considered when generating power control information and can configure the action space only with the base stations that transmit an ASRI indicating “not decided.” That is, for base stations where the ASRI indicates “active” or “sleep,” the central unit (320) determines power control information as active or sleep based on the ASRI, and for base stations where the ASRI indicates “not decided,” the power control information can be determined based on state information according to the method described above.

[0077] Additionally, by introducing a deep neural network for action elimination within the action space simplification unit (324), the action space can be reduced by receiving CSI, QoS, and ASRI signals as inputs and identifying base stations that can be additionally excluded from the possibilities among small base stations where the ASRI signal is “not decided.” The power control unit (322) can maximize the efficiency of machine learning for generating power control information by using the action reducing notifier output by the action space simplification unit (324) as an input value. That is, the power control unit (322) within the central unit (320) takes the CSI between each base station and terminal, the terminals' QoS indicator, and the Action reducing notifier as inputs and determines the action expected to yield the largest Q-value, namely the base station active / sleep mode, as the action. Since the number of Q-value cases that the Deep Q-Network must predict is significantly reduced by using ASRI and the Action space simplifier, the computational complexity can be reduced. Specifically, when the Q-network receives a state and a specific action as inputs and predicts the Q-value of the corresponding state-action pair, the number of cases for determining the active / sleep mode that need to be considered can be dramatically reduced by using the action space simplification unit (324) compared to the case where only the conventional Q-Network is used. This has the significance of improving the learning speed of the artificial neural network inside the agent by preventing the active / sleep mode determination and the corresponding reward that cannot satisfy the terminal's QoS requirements from being used as training data for the DQN.

[0078] The central unit (320) can transmit power control information including active / sleep mode information of each determined base station, or information corresponding to each base station among the generated power control information, to each base station. Each base station that receives the power control information can transmit a reward based on the received power control information from each terminal and base station to the central unit (320). As an example, the reward transmitted by each base station can be determined based on the satisfaction of requirements and actual power consumption, and as shown in Equation 8, the value is larger as power consumption decreases, and can be set so that a penalty is imposed if specific conditions required by the base station and terminal (e.g., terminal's data requirement, QoS, etc.) are not satisfied.

[0079] According to one embodiment, each base station may transmit a (-) compensation value to the central unit when an active / sleep mode result is transmitted that cannot satisfy the terminal's QoS, such as when a mode determination result different from the ASRI signal determined by the base station itself is received, or when a sleep mode result is transmitted even though there is a terminal to transmit data. Additionally, a compensation variable (e.g., the base station's power consumption or ACK / NACK indicator) that has been pre-agreed upon with the central unit (320) may be transmitted to the central unit (320). Subsequently, the central unit (320) uses the received compensation variable value to [represent] the loss function of the DQN disclosed above through Equation 4. After calculation, the parameters of the artificial neural network can be updated.

[0080] FIG. 4 is a flowchart showing the operation of a base station according to one embodiment of the present disclosure.

[0081] In step 410, the base station may obtain state information for use in determining ASRI or for transmission to a central unit. The state information may include at least one of channel state information (CSI) between the base station and the terminal, the terminal's Quality of Service indicator, User perceived throughput, the amount of remaining traffic in the base station queue, the data fulfillment rate of each terminal to date, resource block scheduling information for terminals within the base station coverage, the intensity of interference received from a neighboring base station, the intensity of interference given to a neighboring base station, and traffic load. Each of the information included in the state information may be obtained through any method as needed, such as by the base station (310) receiving from the terminal, by the base station (310) acquiring it itself, or by receiving it from a macro base station or a central unit (320). For example, channel status information and the terminal's QoS indicator are received from a terminal, and terminal data fulfillment rate and resource block scheduling information for terminals within the base station coverage are received from a macro base station or a central unit, and the base station itself can obtain information such as the amount of remaining traffic in the base station queue. This is merely an example for obtaining status information and does not limit the present invention.

[0082] In step 420, the base station may generate an ASRI by determining the necessity of active mode / sleep mode operation based on at least some of the state information. The state information may include at least one of channel state information (CSI) between the base station and the terminal, the terminal's Quality of Service indicator (QoS indicator), User perceived throughput, the amount of remaining traffic in the base station queue, the data fulfillment rate of each terminal so far, resource block scheduling information for terminals within the base station coverage, the intensity of interference received from neighboring base stations, the intensity of interference given to neighboring base stations, and traffic load, and the base station may generate an ASRI based on at least some or all of these. As an example, the base station may generate an ASRI based on at least one of the signal and terminal scheduling related variables, such as the QoS indicator, User perceived throughput, the amount of remaining traffic in the base station queue, the data fulfillment rate of each terminal so far, and resource block scheduling information for terminals within the small base station coverage.

[0083] In step 430, the base station may transmit the previously acquired status information and ASRI to the central unit. In one embodiment, the base station may transmit the status information and ASRI to the central unit through a backhaul network.

[0084] In step 440, the base station may receive power control information (or information corresponding to the base station among the power control information) determined by the central unit based on the state information and ASRI from the central unit. The power control information may be information indicating an active / sleep mode for each base station (310), or information indicating one or more of the number of antennas to use (A), the number of transmission power levels (T), and the number of resource blocks (B) to be allocated for a base station indicated as an active mode.

[0085] In step 450, the base station can perform power control based on power control information received from the central unit.

[0086] Although not explicitly stated, the base station may transmit to the central unit a reward generated based on the satisfaction of requirements and actual power consumption, in accordance with the results of power control performance based on power control information received from the central unit. The reward may have a larger value as power consumption decreases, as shown in Equation 8, and may be configured to impose a penalty if specific conditions required by the base station and the terminal (e.g., terminal's data requirements, QoS, etc.) are not satisfied.

[0087] FIG. 5 is a flowchart showing the operation of a central unit according to one embodiment of the present disclosure.

[0088] In step 510, the central unit can receive status information and ASRI from the base station.

[0089] In step 520, the central unit may generate power control information based on the status information and ASRI received from the base station. In one embodiment, the central unit may generate power control information based on at least one of the status information, ASRI, and an action reducing notifier generated by the action space simplification unit (324). The power control information may be information indicating an active / sleep mode for each base station, or information indicating one or more of the number of antennas to use (A), the number of transmission power levels (T), and the number of resource blocks (B) allocated for a base station indicated as active mode. The central unit may generate power control information for each base station by determining the action for which the largest Q-value is expected based on at least one of the status information, ASRI, and an action reducing notifier generated by the action space simplification unit. In one example, if the ASRI received from the base station indicates “active” or “sleep,” the power control information corresponding to the base station may be determined as active or sleep based on the ASRI. Additionally, if the ASRI received from the base station is “not decided,” the base station is included in the action space considered for determining power control information based on state information, and can be determined based on the state information received from the base station. Meanwhile, although not illustrated, the central unit may generate power control information based further on the compensation according to the power control results of each base station performed in the previous time unit.

[0090] In step 530, the central unit can transmit the generated power control information or the information corresponding to each base station among the power control information to each base station.

[0091] FIG. 6 is a flowchart illustrating a power control method in a communication network according to one embodiment of the present disclosure.

[0092] Specifically, FIG. 6 is a flowchart illustrating a power control method when a base station (620) is directly connected to a central unit (630). In this case, the base station (620) may be a macro base station or a small base station directly connected to the central unit. Referring to FIG. 6, the base station (620) can obtain status information from a terminal (610) or a central unit (630) (601). Although FIG. 6 is illustrated as the base station obtaining status information from both the terminal and the central unit, depending on the configuration of the base station or the type of information to be obtained, the base station may obtain status information from either the terminal (610) or the central unit (630). Additionally, although not illustrated, the status information may include not only information obtained from the terminal (610) or the central unit (630) but also information obtained by the base station (620) itself. Based on the acquired status information, the base station (620) can determine the ASRI by deciding whether to request the central unit (630) to operate in active mode or to operate in sleep mode (602). If it is difficult for the base station (620) to independently decide whether to request active mode operation or sleep mode operation, the ASRI can be determined to indicate “not decided”. After determining the ASRI, the base station (620) can transmit the acquired status information and the determined ASRI to the central unit (630) (603). The central unit (630) can determine power control information based on the status information and ASRI received from one or more base stations (604). The central unit (630) transmits the determined power control information or the information among the power control information corresponding to the base station (620) to the base station (620) (605), and the base station (620) that receives this can perform power control based on the received power control information (607).Additionally, although not shown, the base station (620) can determine the reward to be transmitted to the central unit (630) based on the result of performing power control based on the received power control information and transmit it to the central unit (630). The central unit (630) can determine the power control information based on the reward as well as the status information and ASRI.

[0093] FIG. 7 is a flowchart illustrating a power control method in a communication network according to another embodiment of the present disclosure.

[0094] Specifically, FIG. 7 is a flowchart illustrating a power control method in which the base station subject to power control is connected to a central unit through a macro base station and corresponds to a small base station (720) controlled by the macro base station. Referring to FIG. 7, the small base station (720) can obtain status information from a terminal (710) (701). Additionally, the small base station (720) can obtain status information from a macro base station (730) (702). In one embodiment, at least some of the status information obtained from the macro base station (730) may be information determined based on at least some of the status information obtained by the small base station (720) from the terminal (710). Based on the obtained status information, the small base station (720) can determine the ASRI by deciding whether to request the central unit (740) to request the small base station (720) to operate in active mode or to operate in sleep mode (703). If it is difficult for the small base station (720) to independently decide whether to request active mode operation or sleep mode operation, it may decide to have the ASRI indicate “not decided”. After the ASRI is determined, the small base station (720) can transmit the acquired status information and the determined ASRI to the macro base station (730) (704). The macro base station (730) can transmit the acquired status information and the ASRI from the small base station (720) to the central unit (740) (705).

[0095] The central unit (740) can determine power control information based on status information and ASRI received from one or more base stations (706). The central unit (740) transmits the determined power control information or information corresponding to the small base station (720) controlled by the macro base station (730) to the macro base station (730) (707), and the macro base station (730) can transmit this back to the small base station (720) (708). The small base station (720) that receives this can perform power control based on the received power control information (709).

[0096] FIG. 8 is a diagram showing the configuration of a base station according to one embodiment of the present disclosure.

[0097] Referring to FIG. 8, the base station may include a transceiver (810), a memory (820), and a processor (830). According to the above-described embodiment, the transceiver (810), the memory (820), and the processor (830) of the base station may operate. However, the components of the base station are not limited to the above-described example. For example, the base station may include more components or fewer components than the above-described components. Furthermore, the transceiver (810), the memory (820), and the processor (830) may be implemented in the form of a single chip.

[0098] The transceiver (810) can transmit and receive signals with other components within the communication network, such as terminals, other base stations, and central units. Here, the signals may include control information and data. To this end, the transceiver (810) may be composed of an RF transmitter that up-converts and amplifies the frequency of a transmitted signal, and an RF receiver that low-noise amplifies a received signal and down-converts the frequency. However, this is merely one embodiment of the transceiver (810), and the components of the transceiver (810) are not limited to an RF transmitter and an RF receiver.

[0099] Additionally, the transceiver (810) can receive a signal through a wireless channel and output it to a processor (830), and transmit the signal output from the processor (830) through a wireless channel.

[0100] The memory (820) can store programs and data necessary for the operation of the base station. Additionally, the memory (820) can store control information or data included in signals transmitted and received by the base station. The memory (820) may be composed of a storage medium or a combination of storage media such as ROM, RAM, hard disk, CD-ROM, and DVD. Additionally, there may be multiple memory (820).

[0101] Additionally, the processor (830) can control a series of processes to enable the base station to operate according to the above-described embodiment. For example, the processor (830) can control the operation of the base station to acquire state information, determine ASRI, and control power according to the received power control information according to the above-described embodiment. There may be multiple processors (830), and the processor (830) can perform component control operations of the base station by executing a program stored in memory (820).

[0102] FIG. 9 is a drawing showing the configuration of a central unit according to one embodiment of the present disclosure.

[0103] Referring to FIG. 9, the central unit may include a transceiver (910), a memory (920), and a processor (930). According to the above-described embodiment, the transceiver (910), the memory (920), and the processor (930) of the central unit may operate. However, the components of the central unit are not limited to the above-described example. For example, the central unit may include more components or fewer components than the above-described components. Furthermore, the transceiver (910), the memory (920), and the processor (930) may be implemented in the form of a single chip.

[0104] The transceiver (910) can transmit and receive signals with other components within a communication network, such as terminals and base stations. Here, the signals may include control information and data. To this end, the transceiver (910) may be composed of an RF transmitter that up-converts and amplifies the frequency of a transmitted signal, and an RF receiver that low-noise amplifies a received signal and down-converts the frequency. However, this is merely one embodiment of the transceiver (910), and the components of the transceiver (910) are not limited to an RF transmitter and an RF receiver.

[0105] Additionally, the transceiver (910) can receive a signal through a wireless channel and output it to a processor (930), and transmit the signal output from the processor (930) through a wireless channel.

[0106] The memory (920) can store programs and data necessary for the operation of the central unit. Additionally, the memory (920) can store control information or data included in signals transmitted and received by the central unit. The memory (920) may be composed of a storage medium or a combination of storage media such as ROM, RAM, hard disk, CD-ROM, and DVD. Additionally, there may be multiple memory (920).

[0107] Additionally, the processor (930) can control a series of processes to enable the central unit to operate according to the above-described embodiment. For example, the processor (930) can control the operation of the central unit to acquire state information, determine ASRI, and control power according to the received power control information according to the above-described embodiment. There may be multiple processors (930), and the processor (930) can perform component control operations of the central unit by executing a program stored in memory (920). Additionally, the processor (930) may be configured to include the power control unit (322) and the action space simplification unit (324) described above.

[0108] Methods according to the claims or embodiments described in the specification of the present disclosure may be implemented in the form of hardware, software, or a combination of hardware and software.

[0109] When implemented in software, a computer-readable storage medium or computer program product storing one or more programs (software modules) may be provided. One or more programs stored in the computer-readable storage medium or computer program product are configured for execution by one or more processors within an electronic device. One or more programs include instructions that cause the electronic device to execute methods according to the claims or embodiments described in the specification of this disclosure.

[0110] Such programs (software modules, software) may be stored in random access memory, non-volatile memory including flash memory, ROM (Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), magnetic disc storage devices, CD-ROM (Compact Disc-ROM), Digital Versatile Discs (DVDs), or other forms of optical storage devices, magnetic cassettes. Alternatively, they may be stored in memory composed of some or all of these. Additionally, each constituent memory may include multiple units.

[0111] Additionally, the program may be stored on an attachable storage device accessible via a communication network such as the Internet, Intranet, Local Area Network (LAN), Wide LAN (WLAN), or Storage Area Network (SAN), or a combination thereof. Such a storage device may be connected to a device performing an embodiment of the present disclosure through an external port. Additionally, a separate storage device on a communication network may be connected to a device performing an embodiment of the present disclosure.

[0112] In the specific embodiments of the present disclosure described above, the components included in the present disclosure are expressed in a singular or plural form according to the specific embodiments presented. However, the singular or plural expression is selected to suit the situation presented for convenience of explanation, and the present disclosure is not limited to singular or plural components; even if a component is expressed in the plural, it may be composed of a singular form, and even if a component is expressed in the singular form, it may be composed of a plural form.

[0113] Meanwhile, the embodiments of the present disclosure disclosed in this specification and drawings are merely specific examples provided to facilitate the explanation of the technical content of the present disclosure and to aid in understanding the present disclosure, and are not intended to limit the scope of the present disclosure. That is, it is obvious to those skilled in the art that other variations based on the technical concept of the present disclosure are possible. Furthermore, each of the above embodiments may be combined and operated together as needed. For example, parts of one embodiment of the present disclosure and another embodiment may be combined to operate a base station and a terminal. In addition, the embodiments of the present disclosure are applicable to other communication systems, and other variations based on the technical concept of the embodiments may also be possible. For example, the embodiments may be applied to LTE systems, 5G or NR systems, etc.

Claims

Claim 1 A method performed by a base station of a communication system comprises: a step of obtaining status information from a terminal; a step of determining an ASRI (active / sleep request indicator) for requesting an operating mode of the base station based on the status information; a step of transmitting the status information and the ASRI to a central unit; a step of receiving power control information determined based on at least one of the status information and the ASRI from the central unit; and a step of performing power control based on the received power control information, wherein the ASRI indicates one of active, sleep, or not decided, and when the ASRI indicates active or sleep, the power control information is determined based on the ASRI, and when the ASRI indicates not decided, the power control information is determined based on the status information. Claim 2 A method according to claim 1, characterized in that the state information includes at least one of channel state information with the terminal and the terminal's QoS (Quality of Service) indicator. Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 A method performed by a central unit of a communication system comprises: receiving an active / sleep request indicator (ASRI) for requesting status information and an operating mode of the base station from a base station; determining power control information for the base station based on at least one of the status information and the ASRI received from the base station; and transmitting the power control information to the base station, wherein the ASRI indicates one of active, sleep, or not decided, and when the ASRI indicates active or sleep, the power control information is determined based on the ASRI, and when the ASRI indicates not decided, the power control information is determined based on the status information. Claim 7 A method according to claim 6, characterized in that the state information includes at least one of channel state information with the terminal and the terminal's QoS (Quality of Service) indicator. Claim 8 delete Claim 9 delete Claim 10 delete Claim 11 A base station of a communication system comprises: a transceiver; and a control unit configured to acquire status information from a terminal, determine an ASRI (active / sleep request indicator) for requesting an operation mode of the base station based on the status information, transmit the status information and the ASRI to a central unit, receive power control information determined based on at least one of the status information and the ASRI from the central unit, and perform power control based on the received power control information, wherein the ASRI indicates one of active, sleep, or not decided, and when the ASRI indicates active or sleep, the power control information is determined based on the ASRI, and when the ASRI indicates not decided, the power control information is determined based on the status information. Claim 12 A base station characterized in that, in claim 11, the above-mentioned status information includes at least one of channel status information with the terminal and the terminal's QoS (Quality of Service) indicator. Claim 13 delete Claim 14 delete Claim 15 delete Claim 16 A central unit of a communication system comprises: a transceiver; and a control unit configured to receive an ASRI (active / sleep request indicator) for requesting status information and an operating mode of the base station from a base station, determine power control information for the base station based on at least one of the status information and the ASRI received from the base station, and transmit the power control information to the base station, wherein the ASRI indicates one of active, sleep, or not decided, and when the ASRI indicates active or sleep, the power control information is determined based on the ASRI, and when the ASRI indicates not decided, the power control information is determined based on the status information. Claim 17 A central unit characterized in that, in claim 16, the above-mentioned status information includes at least one of channel status information with the terminal and the terminal's QoS (Quality of Service) indicator. Claim 18 delete Claim 19 delete Claim 20 delete

Citation Information

Patent Citations

  • Method and apparatus for performing handover in mobile communication system

    KR1020150021276A

  • System and method for controlling power in a communication system

    US20080220804A1

  • Method for reducing power consumption of base station in wireless communication system

    US20110170466A1