CN113038583A - Inter-cell downlink interference control method, device and system suitable for ultra-dense network - Google Patents
Inter-cell downlink interference control method, device and system suitable for ultra-dense network Download PDFInfo
- Publication number
- CN113038583A CN113038583A CN202110262914.5A CN202110262914A CN113038583A CN 113038583 A CN113038583 A CN 113038583A CN 202110262914 A CN202110262914 A CN 202110262914A CN 113038583 A CN113038583 A CN 113038583A
- Authority
- CN
- China
- Prior art keywords
- base station
- network
- user
- cell
- downlink
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 230000005540 biological transmission Effects 0.000 claims abstract description 46
- 230000002787 reinforcement Effects 0.000 claims abstract description 43
- 238000012549 training Methods 0.000 claims description 19
- 230000008569 process Effects 0.000 claims description 15
- 230000009471 action Effects 0.000 claims description 13
- 230000008901 benefit Effects 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 13
- 238000005265 energy consumption Methods 0.000 claims description 10
- 150000001875 compounds Chemical class 0.000 claims description 9
- 238000011217 control strategy Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000009826 distribution Methods 0.000 claims description 6
- 230000006399 behavior Effects 0.000 claims description 4
- 230000007774 longterm Effects 0.000 claims description 4
- 239000013068 control sample Substances 0.000 claims description 3
- 238000005562 fading Methods 0.000 claims description 3
- 230000002452 interceptive effect Effects 0.000 claims description 3
- 238000003012 network analysis Methods 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 10
- 238000004590 computer program Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010295 mobile communication Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000013526 transfer learning Methods 0.000 description 2
- 230000001413 cellular effect Effects 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000035755 proliferation Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
- H04W52/02—Power saving arrangements
- H04W52/0203—Power saving arrangements in the radio access network or backbone network of wireless communication networks
- H04W52/0206—Power saving arrangements in the radio access network or backbone network of wireless communication networks in access points, e.g. base stations
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
- H04W52/04—TPC
- H04W52/06—TPC algorithms
- H04W52/14—Separate analysis of uplink or downlink
- H04W52/146—Uplink power control
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
- H04W52/04—TPC
- H04W52/18—TPC being performed according to specific parameters
- H04W52/24—TPC being performed according to specific parameters using SIR [Signal to Interference Ratio] or other wireless path parameters
- H04W52/241—TPC being performed according to specific parameters using SIR [Signal to Interference Ratio] or other wireless path parameters taking into account channel quality metrics, e.g. SIR, SNR, CIR, Eb/lo
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
- H04W52/04—TPC
- H04W52/18—TPC being performed according to specific parameters
- H04W52/24—TPC being performed according to specific parameters using SIR [Signal to Interference Ratio] or other wireless path parameters
- H04W52/243—TPC being performed according to specific parameters using SIR [Signal to Interference Ratio] or other wireless path parameters taking into account interferences
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Quality & Reliability (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention provides a method, a device and a system for controlling downlink interference among cells, which are suitable for an ultra-dense network, wherein the method comprises the following steps: acquiring ultra-dense network parameter data of a network cell where a base station is located and current downlink signal to interference plus noise ratio data r of a user served by the base station; based on the acquired data, obtaining the average user density rho of the ultra-dense network, the current user channel state h of the base station and the network throughput f of the network cell where the base station is located; and inputting the obtained r, rho, h and f as current node state parameters into a pre-trained reinforcement learning network to obtain downlink transmission power data of the base station to the served users, which is output by the reinforcement learning network, and further implementing corresponding downlink transmission power to each user served by the base station. The invention can optimize the downlink transmitting power of the base station to the served user and the interference between downlink cells under the condition of ultra-dense network deployment, and improve the user access service quality.
Description
Technical Field
The invention relates to the field of base station transmitting power control in the communication technology, in particular to a method, a device and a system for controlling downlink interference among cells, which are suitable for an ultra-dense network.
Background
With the explosive growth of the number of intelligent terminals and communication services, 4G has been insufficient to meet the increasing demands of multi-device connections and high data rates, larger bandwidths, lower delays and low interference. To address these challenges, 5G is considered the most promising technology, and 5G has also been introduced for commercial decades. The 5G is a stable, high-speed, reliable evolution of mobile communication systems, is a continuation and enhancement of the 4G, and is intended to interconnect billions of objects, provide an extremely high data transmission rate for multimedia applications, and handle the accompanying proliferation of mobile network traffic.
Exponential growth and availability of various forms of data are the main driving force for continuous development of the communication industry, so that wireless base stations are increasingly Ultra-Dense and show a large-scale overlapping coverage trend, a mobile communication Network is further developed towards miniaturization of base stations and Ultra-Dense cell in the future, and an Ultra-Dense Network (UDN) is generated at present. The super-dense network forms a multilayer heterogeneous network together through a macro base station cell and a large number of micro cells, and the basic idea is to enable an access node to be as close to an end user as possible, and to achieve the super-dense network simply by densely deploying small cells in a hot spot area generating a large amount of flow, wherein the small cells are used as access nodes, the transmission power is small, and the coverage area is small. Ultra-dense network deployment enhances spatial reuse and significantly increases the user capacity of 5G mobile networks by using low-cost and low-power small cell base stations, and enables any user to be very close to a cell at any time and any place, thereby improving the user access service quality.
Due to the scarcity of frequency spectrum and the uncoordinated network infrastructure, the problem of how to control the interference between downlink cells in a time-varying channel state is inevitably brought by the ultra-dense deployment of small and medium cells in a 5G cellular network. Therefore, it is important how to provide a practical interference control scheme under the ultra-dense network deployment, which takes into account the reduction of signaling and computational overhead of the BS.
Disclosure of Invention
The invention aims to provide a method, a device and a system for controlling downlink interference among cells, which are suitable for an ultra-dense network. The technical scheme adopted by the invention is as follows.
In one aspect, the present invention provides a method for controlling inter-cell downlink interference applicable to an ultra-dense network base station, including:
acquiring ultra-dense network parameter data of a network cell where a base station is located and current downlink signal to interference plus noise ratio data r of a user served by the base station;
based on the acquired data, obtaining the average user density rho of the ultra-dense network, the current user channel state h of the base station and the network throughput f of the network cell where the base station is located;
inputting the obtained r, rho, h and f as current node state parameters into a pre-trained reinforcement learning network to obtain downlink transmission power data of a base station to a served user, wherein the downlink transmission power data is output by the reinforcement learning network;
and according to the obtained downlink transmission power data, implementing corresponding downlink transmission power for each user served by the base station.
Optionally, the ultra-dense network parameter data includes a current active user number of a network cell in which the base station is located, and the average user density is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,representing the number of active users of network cell i at time k, ω is a constant variable,representing the area region phi in network cell iiIs measured in the probability distribution of the number of users. The average user density is calculated by adopting the number of users in the network cell where the base station is located, so that the state of the base station can be reflected more accurately.
Above scheme, area region Φ in cell iiThe number of users of (1) changes with time, the invention is to rho(k)Is based on the average user density p of the supposition ultra-dense network(k)Subject to the two-dimensional poisson process.
In the invention, the channel state h of each current user of the base station can be obtained by utilizing channel estimation function estimation, the ultra-dense network parameter data also comprises a base station pilot frequency sequence, and the calculation of the channel estimation function is carried out based on the pilot frequency sequence.
Optionally, the network throughput f of the network cell in which the base station is located is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,downlink transmit power representing time instances when base station implements k-1 to user nThen, the downlink signal to interference plus noise ratio fed back by the user n at the time k; n represents the number of users in the network cell in which the base station is located.
Optionally, the training process of the reinforcement learning network includes:
s1, initializing reinforcement learning network parameters, wherein: the reinforcement learning network parameters comprise a learning rate, a discount factor, a V value and a Q value;
s2, for the base station interference control sample in the ultra-dense network with the known target state, determining the transmitting power set omega of the base station and the average user density rho corresponding to the current time k(k)Throughput f of the network cell in which the base station is located(k)Downlink signal to interference plus noise ratio for users served by a base stationAnd channel stateObtaining the node state of the base station at the time k
S3, node state S based on time k(k)Selecting downlink transmission power for users served by the base station according to a preset base station power control strategy
S4, acquiring base station simulation implementation X(k)Downlink signal to interference plus noise ratio of post-user feedback
S5, calculating the energy consumption and the inter-cell interference of the network cell where the base station is located, and determining that the base station implements the action X on the node k(k)Benefit u of(K);
S6, representing the long-term discount reward of BS taking action X at state S by Q function Q (S, X), based on node state S at time k(k)And behavior X(k)Updating the Q value of the reinforcement learning network;
s7, judging whether the base station reaches the target state on the current node: if so, stopping iteration and finishing the training of the reinforcement learning network; if the target state is not reached, the process goes to steps S2-S7 to implement X based on the updated Q value(k)Downlink signal to interference plus noise ratio of post-user feedbackAnd performing iterative training of the reinforcement learning network.
Optionally, the target status includes a predetermined channel status of a user served by the base station and a downlink signal to interference plus noise ratio of the user. And (4) reaching the target state, namely, the result data of a certain iteration in the training process conforms to each parameter data in the target state or is in a set reasonable range.
In the above scheme, the initialized Q value sequence may adopt an interference control experience sequence obtained by using a transfer learning method in a similar application scenario, so as to reduce an exploration process of a reinforcement learning network on random interference control at the beginning. The reinforcement learning network is initialized to a value of 0.
If the maximum transmission power P of the base station is given, the transmission power of the BS to the user n will be quantized to L +1 orders of magnitude, that is, in the present invention, the downlink transmission power set of the base station to the served user n is:
Ω={jP/L}0≤j≤L
where Ω represents the transmit power set, P represents the maximum transmit power for a given base station, and L represents the feasible transmit power level.
The maximum transmission power of the base station should be changed according to the type of network cell, such as femto cell, micro cell, pico cell, etc., to meet the coverage and service requirements.
Optionally, the downlink transmit power selected for the user served by the base station is selected according to the following formula:
in the formula, epsilon is random arbitrary small positive number, epsilon is less than 1, theta represents the optimal transmitting power,representing the power control strategy adopted by the base station when conducting the next action delta omega exploration,
optionally, the base station implements action X on node k(k)Benefit u of(K)Calculated according to the following formula:
in the formula, a first term represents the sum of downlink signal-to-interference-plus-noise ratios of users, a second term represents the energy consumption of a base station, and a third term represents the total inter-cell interference; b is the system bandwidth, σ is the noise power of the receiver, CsIn terms of a unit of transmission cost,the number of users served by base station i, G represents the number of neighbouring cells of the cell in which the base station is located,represents the power of the interfering signal transmitted from the ith base station and received by user n,which represents the power gain of the channel and,the interference factor represents the inter-cell interference of the base station in the ith (i is more than or equal to 1 and less than or equal to G) adjacent cell to the user m of the cell in which the benefit base station to be calculated is located, and the expression is as follows:
in the formula (I), the compound is shown in the specification,representing the path loss from base station i to user m,represents the large-scale fading gain from base station i to user m, and η represents the number of network cells per unit area in the ultra-dense network.
In step S6, the Q value is updated using the following formula, namely, the Sarsa method:
and after the Q value is updated, if the node state reaches the target state at the moment, the iteration is stopped, and the training is stopped, otherwise, the next iteration training is carried out based on the updated Q value and the finally obtained signal to interference and noise ratio.
In a second aspect, the present invention provides an inter-cell downlink interference control apparatus suitable for an ultra-dense network base station, including:
the data acquisition module is configured to acquire ultra-dense network parameter data of a network cell where the base station is located and current downlink signal to interference plus noise ratio data r of a user served by the base station;
the network analysis module is configured to obtain the average user density rho of the ultra-dense network, the current user channel state h of the base station and the network throughput f of the network cell where the base station is located based on the obtained data;
the control strategy calculation module is configured to input the obtained r, rho, h and f as node state parameters at the moment k to a pre-trained reinforcement learning network to obtain downlink transmission power data of a base station output by the reinforcement learning network to a served user;
and the control output module is configured to implement corresponding downlink transmission power for each user served by the base station according to the obtained downlink transmission power data.
The specific functional implementation of each functional module refers to the relevant content of the method of the first aspect.
In a third aspect, the present invention provides an inter-cell downlink interference control system suitable for an ultra-dense network, where the ultra-dense network includes multiple network cells, each network cell is provided with a base station, and users served by each base station include users in a cell where the base station is located and users in an adjacent cell;
each base station respectively executes the inter-cell downlink interference control method of the first aspect to obtain and implement downlink transmission power for the served users.
Advantageous effects
The invention is suitable for the downlink interference control method among the cells of the ultra-dense network base station, through introducing the reinforcement learning algorithm, make the base station of every cell carry on the downlink transmission power distribution, needn't know the channel state and its interference distribution among the adjacent cells, only according to downlink signal and interference plus noise ratio and downlink channel state that users feedback, can utilize the reinforcement learning network trained in advance to get the downlink transmission power to users of the next time slot, achieve the effects of optimizing the base station transmission power and inhibiting the interference among the cells.
In addition, referring to the training process of the reinforcement learning network, the behavior X is implemented on the node k at the computing base station(k)Benefit u of(K)When the reinforcement learning network obtained by iterative training is applied, the obtained downlink transmission power scheme can optimize the benefit of the base station, save the energy consumption of the base station of the ultra-dense small cell and reduce the network overhead.
Drawings
Fig. 1 is a flowchart illustrating an embodiment of an inter-cell downlink interference control method for an ultra-dense network base station according to the present invention;
fig. 2 is a schematic diagram of a super-dense network architecture to which the inter-cell downlink interference control method of the present invention is applied, wherein: BS stands for a base station, User stands for a User, and cell stands for a cell;
fig. 3 is a schematic diagram illustrating an iterative training process of the reinforcement learning network.
Detailed Description
The following further description is made in conjunction with the accompanying drawings and the specific embodiments.
Example 1
This embodiment introduces an inter-cell downlink interference control method suitable for an ultra-dense network base station, as shown in fig. 1, the method includes:
acquiring ultra-dense network parameter data of a network cell where a base station is located and current downlink signal to interference plus noise ratio data r of a user served by the base station;
based on the acquired data, obtaining the average user density rho of the ultra-dense network, the current user channel state h of the base station and the network throughput f of the network cell where the base station is located;
inputting the obtained r, rho, h and f as current node state parameters into a pre-trained reinforcement learning network to obtain downlink transmission power data of a base station to a served user, wherein the downlink transmission power data is output by the reinforcement learning network;
corresponding downlink transmit power is applied to each user served by the base station according to the output of the reinforcement learning network.
When the method is applied, the base stations configured in all cells in the ultra-dense network respectively execute the method, downlink transmission power is distributed to the served users, and downlink interference control optimization among the cells in the ultra-dense network can be realized.
In the method of the present embodiment, the node state parameters input to the pre-trained reinforcement learning network are specifically determined as follows.
The ultra-dense network parameter data of the network cell in which the base station is positioned comprises the current active user number of the network cell in which the base station is positionedAnd each base station adopts the number of the active users of the cell corresponding to the base station when calculating the average user density in the interference control process.
Suppose that cell i has at time kAverage user density rho of individual active users, ultra-dense small cells(k)Obeying a two-dimensional Poisson procedure, taking into account the area region phi in the cell iiThe number of users in the table is changed with time, and the probability distribution can be expressed as follows:
in the formula (I), the compound is shown in the specification,representing the number of active users of network cell i at time k, ω is a constant variable,representing the area region phi in network cell iiIs measured in the probability distribution of the number of users. The average user density is calculated by adopting the number of users in the network cell where the base station is located, so that the state of the base station can be reflected more accurately.
Thus, the average user density ρ(k)The determination can be made by the above equation (1) according to the number of real-time active users of the cell in which the base station is located that can be determined.
The channel state h of each current user of the base station can be obtained by utilizing channel estimation function estimation, the ultra-dense network parameter data also comprises a base station pilot frequency sequence, and the calculation of the channel estimation function is carried out based on the pilot frequency sequence.
The base station can acquire the signal-to-noise ratio of the signal and the interference plus the noise of the downlink of the user from the feedback channel in real timeBased onThe network throughput f of the network cell in which the base station is located is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,downlink transmit power representing time instances when base station implements k-1 to user nThen, the downlink signal to interference plus noise ratio fed back by the user n at the time k; n represents the number of users in the network cell in which the base station is located.
The training process of the reinforcement learning network adopted by each base station in the embodiment when performing interference optimization control includes:
s1, initializing reinforcement learning network parameters, wherein: the reinforcement learning network parameters comprise a learning rate, a discount factor, a V value and a Q value;
s2, for the base station interference control sample in the ultra-dense network with the known target state, determining the transmitting power set omega of the base station and the average user density rho corresponding to the current time k(k)Throughput f of the network cell in which the base station is located(k)Downlink signal to interference plus noise ratio for users served by a base stationAnd channel stateObtaining the node state of the base station at the time k
S3, node state S based on time k(k)Selecting downlink transmission power for users served by the base station according to a preset base station power control strategy
S4, acquiring base station simulation implementation X(k)Downlink signal to interference plus noise ratio of post-user feedback
S5, calculating the energy consumption and the inter-cell interference of the network cell where the base station is located, and determining that the base station implements the action X on the node k(k)Benefit u of(K);
S6, representing the long-term discount reward of BS taking action X at state S by Q function Q (S, X), based on node state S at time k(k)And behavior X(k)Updating the Q value of the reinforcement learning network;
s7, judging whether the base station reaches the preset target state on the current node: if so, stopping iteration and finishing the training of the reinforcement learning network; if notTo the target state, the process goes to steps S2-S7, based on the updated Q value and implementation X(k)Downlink signal to interference plus noise ratio of post-user feedbackAnd performing iterative training of the reinforcement learning network.
In order to reduce the time consumed by the exploration process of the reinforcement learning network on the random interference control at the beginning, the initialized Q value sequence can adopt an interference control experience sequence under a similar application scene obtained by using a transfer learning method.
The ultra-dense network parameters and the control target state of the interference control samples used for training are historical known data or comprise pre-calibrated signal-to-interference-and-noise ratio data and a channel state data range.
Given the maximum transmission power P of the base station BS, the transmission power of the BS to the user nThe wave quantization is L +1 magnitude order, and in the training process, the downlink transmission power distributed to the served user n by the base station in each iteration is selected from a transmission power set, namely:
the maximum transmission power of the base station should be changed according to the type of network cell, such as femto cell, micro cell, pico cell, etc., to meet the coverage and service requirements.
In step S3, downlink transmit power is calculated according to an epsilon greedy algorithm, specifically: the base station selects downlink transmit power for the served users according to the following equation:
in the formula, epsilon is random small positive number, epsilon is less than 1, and theta representsThe optimum transmission power of the radio frequency signal is,representing the power control strategy adopted by the base station when conducting the next action delta omega exploration,
in the formula (4), the optimal transmitting power is selected by using the probability of 1-epsilon, and the suboptimal transmitting power is selected by using the probability of epsilon, so that the network calculation can be prevented from staying at the local optimal solution.
Calculating the energy consumption E of the network cell where the base station is located and the inter-cell interference I, and performing the calculation according to the following formula:
bringing energy consumption of network cell where base station is located and inter-cell interference into base station to implement action X on node k(k)Benefit u of(K)To obtain a benefit u(K)The calculation formula of (2) is as follows:
in the formula (7), a first term represents the sum of downlink signal-to-interference-plus-noise ratios of users, a second term represents energy consumption of a base station, and a third term represents total inter-cell interference; in equations (5) to (7), B is the system bandwidth, σ is the noise power of the receiver, and CsIn terms of a unit of transmission cost,the number of users served by base station i, G represents the number of neighbouring cells of the cell in which the base station is located,represents the power of the interfering signal transmitted from the ith base station and received by user n,which represents the power gain of the channel and,the interference factor represents the inter-cell interference of the base station in the ith (i is more than or equal to 1 and less than or equal to G) adjacent cell to the user m of the cell in which the benefit base station to be calculated is located, and the expression is as follows:
in the formula (I), the compound is shown in the specification,representing the path loss from base station i to user m,represents the large-scale fading gain from base station i to user m, and η represents the number of network cells per unit area in the ultra-dense network.
In step S6, the Q function represented by Q (S, X) corresponds to a long-term discount reward for a base station taking action X at state S, where S ∈ S, S denotes the state space, and X ═ Xn]1≤n≤N,xnE.g. omega. This example updates the Q value using the following formula, Sarsa method:
and after the Q value is updated, if the node state reaches the target state at the moment, the iteration is stopped, and the training is stopped, otherwise, the next iteration training is carried out based on the updated Q value and the finally obtained signal to interference and noise ratio.
Example 2
Based on the same inventive concept as the embodiment, the embodiment introduces an inter-cell downlink interference control apparatus suitable for an ultra-dense network base station, including:
the data acquisition module is configured to acquire ultra-dense network parameter data of a network cell where the base station is located and current downlink signal to interference plus noise ratio data r of a user served by the base station;
the network analysis module is configured to obtain the average user density rho of the ultra-dense network, the current user channel state h of the base station and the network throughput f of the network cell where the base station is located based on the obtained data;
the control strategy calculation module is configured to input the obtained r, rho, h and f as node state parameters at the moment k to a pre-trained reinforcement learning network to obtain downlink transmission power data of a base station output by the reinforcement learning network to a served user;
and a control output module configured to implement a corresponding downlink transmit power for each user served by the base station according to the output of the reinforcement learning network.
The specific function implementation of each functional module refers to the relevant content in the method of embodiment 1.
Example 3
This embodiment introduces an inter-cell downlink interference control system suitable for an ultra-dense network, where the ultra-dense network has a structure as shown in fig. 3, and includes multiple network cells, each network cell is provided with a base station BS, each base station BS is equipped with multiple isotropic antennas, and a User served by each base station includes a User in a cell where the base station is located and a User in an adjacent cell.
In this embodiment, the base station in each cell respectively executes the inter-cell downlink interference control method in embodiment 1 in real time to obtain and implement the downlink transmission power for the served user, so that the transmission power for the served user can be optimized under the deployment of the ultra-dense network, the inter-cell interference of the downlink in the system is reduced, the benefit of the base station is optimized, and the overall energy consumption and cost of the ultra-dense network are reduced.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (10)
1. An inter-cell downlink interference control method suitable for an ultra-dense network base station is characterized by comprising the following steps:
acquiring ultra-dense network parameter data of a network cell where a base station is located and current downlink signal to interference plus noise ratio data r of a user served by the base station;
based on the acquired data, obtaining the average user density rho of the ultra-dense network, the current user channel state h of the base station and the network throughput f of the network cell where the base station is located;
inputting the obtained r, rho, h and f as current node state parameters into a pre-trained reinforcement learning network to obtain downlink transmission power data of a base station to a served user, wherein the downlink transmission power data is output by the reinforcement learning network;
and according to the obtained downlink transmission power data, implementing corresponding downlink transmission power for each user served by the base station.
2. The method of claim 1, wherein the ultra-dense network parameter data includes a current number of active users of a network cell in which the base station is located, and wherein the average user density is calculated according to the following formula:
3. The method of claim 1, wherein the network throughput f of the network cell in which the base station is located is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,downlink transmit power representing time instances when base station implements k-1 to user nThen, the downlink signal to interference plus noise ratio fed back by the user n at the time k; n represents the number of users in the network cell in which the base station is located.
4. The method according to any one of claims 1-3, wherein the training process of the reinforcement learning network comprises:
s1, initializing reinforcement learning network parameters, wherein: the reinforcement learning network parameters comprise a learning rate, a discount factor, a V value and a Q value;
s2, for the base station interference control sample in the ultra-dense network with the known target state, determining the transmitting power set omega of the base station and the average user density rho corresponding to the current time k(k)Throughput f of the network cell in which the base station is located(k)Downlink signal to interference plus noise ratio for users served by a base stationAnd channel stateObtaining the node state of the base station at the time k
S3, node state S based on time k(k)Selecting downlink transmission power for users served by the base station according to a preset base station power control strategy
S4, acquiring base station simulation implementation X(k)Downlink signal to interference plus noise ratio of post-user feedback
S5, calculating the energy consumption and the inter-cell interference of the network cell where the base station is located, and determining that the base station implements the action X on the node k(k)Benefit u of(K);
S6, representing the long-term discount reward of BS taking action X at state S by Q function Q (S, X), based on node state S at time k(k)And behavior X(k)Updating the Q value of the reinforcement learning network;
s7, judging whether the base station reaches the preset target state on the current node: if so, stopping iteration and finishing the training of the reinforcement learning network; if the target state is not reached, the process goes to steps S2-S7 to implement X based on the updated Q value(k)Downlink signal to interference plus noise ratio of post-user feedbackAnd performing iterative training of the reinforcement learning network.
5. The method of claim 4, wherein the target states include predetermined channel states of users served by the base station and user downlink signal to interference plus noise ratios.
6. The method of claim 4, wherein the downlink transmit power set of the base station for served user n is:
Ω={jP/L}0≤j≤L
where Ω represents the transmit power set, P represents the maximum transmit power for a given base station, and L represents the feasible transmit power level.
7. The method of claim 4, wherein the selecting the downlink transmit power for the user served by the base station is selected according to the following equation:
8. the method of claim 1, wherein the base station implements action X on node k(k)Benefit u of(K)Calculated according to the following formula:
wherein the first term represents the sum of the downlink signal-to-interference-plus-noise ratio of each user, and the second term represents the sum of the downlink signal-to-interference-plus-noise ratio of each userThe term represents the base station energy consumption, and the third term represents the overall inter-cell interference;respectively representing the downlink transmitting power of the base station to the user n at the moment k and the channel state of the user served by the base station at the moment k; b is the system bandwidth, σ is the noise power of the receiver, CsIn terms of a unit of transmission cost,the number of users served by base station i, G represents the number of neighbouring cells of the cell in which the base station is located,represents the power of the interfering signal transmitted from the ith base station and received by user n,which represents the power gain of the channel and,the interference factor represents the inter-cell interference of the base station in the ith (i is more than or equal to 1 and less than or equal to G) adjacent cell to the user m of the cell in which the benefit base station to be calculated is located, and the expression is as follows:
9. An inter-cell downlink interference control apparatus suitable for a super-dense network base station, comprising:
the data acquisition module is configured to acquire ultra-dense network parameter data of a network cell where the base station is located and current downlink signal to interference plus noise ratio data r of a user served by the base station;
the network analysis module is configured to obtain the average user density rho of the ultra-dense network, the current user channel state h of the base station and the network throughput f of the network cell where the base station is located based on the obtained data;
the control strategy calculation module is configured to input the obtained r, rho, h and f as node state parameters at the moment k to a pre-trained reinforcement learning network to obtain downlink transmission power data of a base station output by the reinforcement learning network to a served user;
and the control output module is configured to implement corresponding downlink transmission power for each user served by the base station according to the obtained downlink transmission power data.
10. An inter-cell downlink interference control system suitable for an ultra-dense network, wherein the ultra-dense network comprises a plurality of network cells, each network cell is respectively provided with a base station, and users served by each base station comprise users in the cell where the base station is located and users in adjacent cells; the method is characterized in that:
the base stations respectively execute the method for controlling downlink interference between cells in any one of claims 1 to 9, obtain downlink transmission power for served users and implement the method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110262914.5A CN113038583A (en) | 2021-03-11 | 2021-03-11 | Inter-cell downlink interference control method, device and system suitable for ultra-dense network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110262914.5A CN113038583A (en) | 2021-03-11 | 2021-03-11 | Inter-cell downlink interference control method, device and system suitable for ultra-dense network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113038583A true CN113038583A (en) | 2021-06-25 |
Family
ID=76469625
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110262914.5A Pending CN113038583A (en) | 2021-03-11 | 2021-03-11 | Inter-cell downlink interference control method, device and system suitable for ultra-dense network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113038583A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115087011A (en) * | 2022-06-20 | 2022-09-20 | 中国联合网络通信集团有限公司 | Downlink signal detection method and device of flexible frame structure simulation system |
WO2023067610A1 (en) * | 2021-10-20 | 2023-04-27 | Telefonaktiebolaget Lm Ericsson (Publ) | A method for network configuration in dense networks |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106358308A (en) * | 2015-07-14 | 2017-01-25 | 北京化工大学 | Resource allocation method for reinforcement learning in ultra-dense network |
US20190116560A1 (en) * | 2017-10-13 | 2019-04-18 | Intel Corporation | Interference mitigation in ultra-dense wireless networks |
CN109714786A (en) * | 2019-03-06 | 2019-05-03 | 重庆邮电大学 | Femto cell Poewr control method based on Q-learning |
CN109982434A (en) * | 2019-03-08 | 2019-07-05 | 西安电子科技大学 | Wireless resource scheduling integrated intelligent control system and method, wireless communication system |
CN110191489A (en) * | 2019-05-17 | 2019-08-30 | 北京科技大学 | Resource allocation methods and device based on intensified learning in a kind of super-intensive network |
CN110972309A (en) * | 2019-11-08 | 2020-04-07 | 厦门大学 | Ultra-dense wireless network power distribution method combining graph signals and reinforcement learning |
-
2021
- 2021-03-11 CN CN202110262914.5A patent/CN113038583A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106358308A (en) * | 2015-07-14 | 2017-01-25 | 北京化工大学 | Resource allocation method for reinforcement learning in ultra-dense network |
US20190116560A1 (en) * | 2017-10-13 | 2019-04-18 | Intel Corporation | Interference mitigation in ultra-dense wireless networks |
CN109714786A (en) * | 2019-03-06 | 2019-05-03 | 重庆邮电大学 | Femto cell Poewr control method based on Q-learning |
CN109982434A (en) * | 2019-03-08 | 2019-07-05 | 西安电子科技大学 | Wireless resource scheduling integrated intelligent control system and method, wireless communication system |
CN110191489A (en) * | 2019-05-17 | 2019-08-30 | 北京科技大学 | Resource allocation methods and device based on intensified learning in a kind of super-intensive network |
CN110972309A (en) * | 2019-11-08 | 2020-04-07 | 厦门大学 | Ultra-dense wireless network power distribution method combining graph signals and reinforcement learning |
Non-Patent Citations (1)
Title |
---|
郑冰原;孙彦赞;吴雅婷;王涛;: "基于深度强化学习的超密集网络资源分配", 电子测量技术, no. 09 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2023067610A1 (en) * | 2021-10-20 | 2023-04-27 | Telefonaktiebolaget Lm Ericsson (Publ) | A method for network configuration in dense networks |
CN115087011A (en) * | 2022-06-20 | 2022-09-20 | 中国联合网络通信集团有限公司 | Downlink signal detection method and device of flexible frame structure simulation system |
CN115087011B (en) * | 2022-06-20 | 2024-04-12 | 中国联合网络通信集团有限公司 | Method and device for detecting downlink signal of flexible frame structure simulation system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zappone et al. | User association and load balancing for massive MIMO through deep learning | |
CN110167176B (en) | Wireless network resource allocation method based on distributed machine learning | |
Zaher et al. | Learning-based downlink power allocation in cell-free massive MIMO systems | |
US9282568B2 (en) | Method and system for dynamic, joint assignment of power and scheduling of users for wireless systems | |
CN110492955B (en) | Spectrum prediction switching method based on transfer learning strategy | |
CN110769514A (en) | Heterogeneous cellular network D2D communication resource allocation method and system | |
CN113038583A (en) | Inter-cell downlink interference control method, device and system suitable for ultra-dense network | |
Zhang et al. | Deep learning based user association in heterogeneous wireless networks | |
Dinh et al. | Energy efficient resource allocation optimization in fog radio access networks with outdated channel knowledge | |
CN114337976A (en) | Transmission method combining AP selection and pilot frequency allocation | |
Kumaresan et al. | Efficient user clustering using a low-complexity artificial neural network (ANN) for 5G NOMA systems | |
CN111405596A (en) | Resource optimization method for large-scale antenna wireless energy-carrying communication system under Rice channel | |
CN111132333A (en) | Time distribution device and method based on time reversal in wireless power supply network | |
CN112788764A (en) | Method and system for task unloading and resource allocation of NOMA ultra-dense network | |
Tran et al. | First-order methods for energy-efficient power control in cell-free massive MIMO | |
Jiang et al. | Q-learning based task offloading and resource allocation scheme for internet of vehicles | |
Kim et al. | Completely distributed power allocation using deep neural network for device to device communication underlaying LTE | |
Gao et al. | Energy-efficient resource block assignment and power control for underlay device-to-device communications in multi-cell networks | |
Zhai et al. | A Q-learning-based resource allocation for downlink non-orthogonal multiple access systems considering QoS | |
Chen et al. | iPAS: A deep Monte Carlo Tree Search-based intelligent pilot-power allocation scheme for massive MIMO system | |
Ron et al. | Learning-based joint optimization of mode selection and transmit power control for D2D communication underlaid cellular networks | |
CN114710187B (en) | Power distribution method for multi-cell large-scale MIMO intelligent communication under user number dynamic transformation scene | |
Foroughi et al. | User association in hybrid UAV-cellular networks for massive real-time IoT applications | |
CN113873525A (en) | Task unloading method and terminal for ultra-dense edge computing network | |
Saeed et al. | Smart Pilot Assignment for IoT in Massive MIMO Systems: A Path Towards Scalable IoT Infrastructure |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |