WO2023273891A1 - 外环初始值的调整方法、设备及计算机可读存储介质 - Google Patents

外环初始值的调整方法、设备及计算机可读存储介质 Download PDF

Info

Publication number
WO2023273891A1
WO2023273891A1 PCT/CN2022/098970 CN2022098970W WO2023273891A1 WO 2023273891 A1 WO2023273891 A1 WO 2023273891A1 CN 2022098970 W CN2022098970 W CN 2022098970W WO 2023273891 A1 WO2023273891 A1 WO 2023273891A1
Authority
WO
WIPO (PCT)
Prior art keywords
outer ring
class
value
initial value
grid
Prior art date
Application number
PCT/CN2022/098970
Other languages
English (en)
French (fr)
Inventor
史珂
刘巧艳
毛凯
李建国
马泽鹏
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Priority to US18/260,514 priority Critical patent/US20240073893A1/en
Priority to CN202280006239.2A priority patent/CN116235436A/zh
Priority to EP22831710.3A priority patent/EP4358444A1/en
Publication of WO2023273891A1 publication Critical patent/WO2023273891A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/12—Wireless traffic scheduling
    • H04W72/1263—Mapping of traffic onto schedule, e.g. scheduled allocation or multiplexing of flows
    • H04W72/1273—Mapping of traffic onto schedule, e.g. scheduled allocation or multiplexing of flows of downlink data flows
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
    • H04L1/0002—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the transmission rate
    • H04L1/0003—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the transmission rate by switching between different modulation schemes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
    • H04L1/0009—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the channel coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
    • H04L1/0015—Systems modifying transmission characteristics according to link quality, e.g. power backoff characterised by the adaptation strategy
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0053—Allocation of signaling, i.e. of overhead other than pilot signals
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/50—Allocation or scheduling criteria for wireless resources
    • H04W72/54—Allocation or scheduling criteria for wireless resources based on quality criteria
    • H04W72/541—Allocation or scheduling criteria for wireless resources based on quality criteria using the level of interference
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/12—Arrangements for detecting or preventing errors in the information received by using return channel
    • H04L1/16—Arrangements for detecting or preventing errors in the information received by using return channel in which the return channel carries supervisory signals, e.g. repetition request signals
    • H04L1/18—Automatic repetition systems, e.g. Van Duuren systems

Definitions

  • the present application relates to the technical field of mobile communication, and in particular to a method, device and computer-readable storage medium for adjusting an initial value of an outer loop.
  • the wireless channel Compared with the wired channel, the wireless channel has a narrower coherence bandwidth and a shorter coherence time, so the wireless channel changes quickly, which is a remarkable feature of the wireless channel. If the wireless channel adopts a fixed modulation and coding method, it is difficult to make full use of spectrum resources. Therefore, Adaptive Modulation and Coding (AMC) technology is introduced in wireless communication technology to improve spectrum efficiency.
  • AMC Adaptive Modulation and Coding
  • Adaptive modulation and coding technology In order to realize the adaptation of the modulation and coding strategy and the user equipment (UE) channel conditions, it is necessary to perform ACK/NACK (Acknowledgment/Negative Acknowledgment, positive feedback/negative feedback) according to the scheduling feedback result of the UE. Adjust the outer loop to achieve the target reliability requirements set by the system. However, this convergence process often requires dozens or even hundreds of transmission opportunities, leading to the fact that some UEs have not yet converged and have completed their transmissions, or converged but did not use the best modulation and coding strategy for transmissions before convergence. Restricted, UE perception is also affected. Among them, convergence refers to the adjustment of the outer loop to make the BLER (Block Error Rate, block error rate) meet the target requirements.
  • BLER Block Error Rate, block error rate
  • Embodiments of the present application provide a method, device, and computer-readable storage medium for adjusting an initial value of an outer loop.
  • a method for adjusting an initial value of an outer loop includes: acquiring feature information of multiple user equipments scheduled by a cell; performing grid division on the feature information of the multiple user equipments, Obtain a plurality of grids; determine the outer ring of each grid, and the outer ring value of the grid is within the range of the target block error rate; according to the characteristic information of the target user equipment, determine that the target user equipment corresponds to a target grid, the target user equipment is one of multiple user equipments scheduled by the cell; and the outer ring value corresponding to the target grid is determined as the outer ring initial value of the target user equipment.
  • a device provided by an embodiment of the present application includes a memory, a processor, and a computer program stored in the memory and operable on the processor.
  • the processor executes the computer program, the above-mentioned first The method for adjusting the initial value of the outer ring described in the embodiment.
  • a computer-readable storage medium provided by the embodiments of the present application stores computer-executable instructions, and the computer-executable instructions are used to perform the adjustment of the initial value of the outer loop as described in the embodiment of the first aspect above. method.
  • Fig. 1 is a flow chart of the method for adjusting the initial value of the outer ring provided by an embodiment of the present application
  • FIG. 2 is a flow chart of dividing grids according to feature information provided by an embodiment of the present application
  • FIG. 3 is a flow chart of a method for adjusting the initial value of the outer ring provided by another embodiment of the present application.
  • FIG. 4 is a schematic diagram of constructing a two-dimensional table in chip type classification of a UE according to an embodiment of the present application
  • Fig. 5 is a schematic diagram of constructing an outer ring one-dimensional table for each grid in an embodiment of the present application
  • Fig. 6 is a flow chart of model revision in an embodiment of the present application.
  • FIG. 7 is a schematic diagram of a grid-class mapping relationship according to an embodiment of the present application.
  • AMC is an adaptive coding and modulation technology adopted on the wireless channel, which ensures the transmission quality of the link by adjusting the modulation mode and coding rate of the wireless link transmission.
  • the channel condition is poor, select a smaller modulation mode and coding rate.
  • the channel condition is good, a larger modulation mode is selected, thereby maximizing the transmission rate.
  • the coding technology and debugging method are modulated, the quality of the channel will also be improved accordingly.
  • the system always hopes that the transmitted data rate is consistent with the channel variation trend, so as to maximize the utilization of the transmission capacity of the wireless channel.
  • the selection of the coding and modulation mode is realized by two control loops, including an inner loop (Inner-Loop) and an outer loop (Outer-Loop).
  • the inner loop means that the base station determines the MCS (Modulation And Coding Scheme) of the uplink and downlink signals of the terminal according to the channel conditions of the UE.
  • the outer loop is a process in which the base station adjusts the MCS of the UE according to the BLER of the signal transmission, and tries to ensure that the actual BLER of the signal transmission is maintained near the target BLER.
  • the outer loop needs to be adjusted according to the ACK/NACK of the scheduling feedback result of the UE, so as to realize the target reliability requirement set by the system.
  • the process of making the BLER reach the target BLER requirement through the adjustment of the outer ring is called convergence, and the outer ring can also be understood as the outer ring value.
  • this convergence process often requires dozens or even hundreds of transmission opportunities, resulting in some UEs not yet converging and having completed transmission, or converging but converging before the transmission If the best modulation and coding strategy is not adopted, the spectrum efficiency is limited, and UE perception is also affected.
  • Small-packet refers to discontinuous burst services.
  • the AMC of small-packet UEs ends before the service converges, and the spectrum efficiency is limited.
  • this application proposes a method for adjusting the initial value of the outer ring of AMC.
  • UEs determine their own initial value of the outer ring according to the attribution grid, instead of giving all UEs a unified initial value of the outer ring;
  • the scheduling feedback results can adaptively adjust the outer ring value of each grid; it can also monitor the system performance in real time. If the grid division is found to be inconsistent with the current environment, the online grid division learning will be started according to the collected network data, which is more efficient. Adapting to the current actual wireless environment, it is more robust to environmental changes, and the initial value of the outer loop obtained is closer to the convergence value, which is conducive to accelerating the convergence speed of the outer loop and improving spectral efficiency.
  • Fig. 1 is a flowchart of a method for adjusting the initial value of the outer ring provided by an embodiment of the present application.
  • the method for adjusting includes but is not limited to the following steps:
  • Step S110 acquiring characteristic information of multiple UEs scheduled by the cell
  • Step S120 grid-dividing feature information of multiple user equipments to obtain multiple grids
  • Step S130 determining the outer ring of each grid, and the outer ring value of the grid is within the range of the target block error rate
  • Step S140 Determine a target grid corresponding to the target user equipment according to the feature information of the target user equipment, where the target user equipment is one of multiple user equipments scheduled by the cell;
  • Step S150 determining the outer circle value corresponding to the target grid as the initial outer circle value of the target user equipment.
  • the convergence process corresponding to different characteristic information of the UE has different adjustment values.
  • the UE may be a mobile phone, a tablet computer, a notebook computer, and the like.
  • the scheduling data collected in the cell is used to obtain characteristic information of multiple UEs through the scheduling data, and the characteristic information may include characteristics such as UE type and channel quality. Then divide the feature information of multiple UEs into grids, and each feature information corresponds to a grid, so that each feature information can be represented by the grid, and the relationship between the grid and the corresponding feature information is established. It can facilitate the management of the feature information of the UE.
  • the initial value of the outer loop determined according to the method of the embodiment will be used by the UE, so that the UE can have the initial value of the outer loop close to the convergence state.
  • each grid needs to determine the corresponding outer ring.
  • the AMC technology adjusts the outer ring according to the feedback of the UE, and the outer ring adjusts the MCS of the UE according to the BLER feedback to ensure that the actual BLER is maintained at the target. Block error rate range. Therefore, in the embodiment of the present application, the characteristic information of the UE is represented by a grid, each grid corresponds to an outer ring, and by determining the outer ring corresponding to each grid, the outer ring value of the grid is within the target block error rate range , so that the outer ring value of each grid can be obtained, and each grid has an independent outer ring value.
  • step S140 when it is necessary to determine the initial value of the outer circle of the target UE, select the corresponding target grid according to the characteristic information of the target UE, and then read the corresponding outer circle value according to the target grid as the initial value of the outer circle of the target UE.
  • Value that is, user equipment can determine their own initial value of the outer ring according to the attribution grid.
  • the initial value of the outer loop obtained is closer to the convergence value, which is beneficial to speed up the convergence speed of the outer loop and improve the spectral efficiency.
  • step S120 the characteristic information of multiple user equipments is divided into grids, including performing grid division on the chip types, as shown in FIG. 2 , further including the following step:
  • Step S121 determining the mapping curve of MCS and SINR (Signal to Interference plus Noise Ratio, Signal to Interference plus Noise Ratio) within the target block error rate range for each chip type;
  • Step S122 using the mapping curve to represent the chip type, clustering all the chip types, and obtaining the classification result of the chip type;
  • step S123 the chip types are divided into grids according to the classification results of the chip types.
  • the demodulation capability of each chip type is represented by the mapping curve of MCS and SINR, and then the slope and intercept of the mapping curve are used as a sample point to represent the mapping curve, and then the chip type is represented, and the sample points are clustered by clustering
  • the collection of objects is divided into multiple categories, and the chip types of the UE are divided into grids according to the clustering structure, so as to obtain the grids corresponding to the chip types.
  • different mobile phones have different chip types, such as baseband chips.
  • the mapping curve of MCS and SINR the chip type of the mobile phone is expressed in the form of a mapping curve, that is, the mapping curve can reflect the different mobile phone types. demodulation capability.
  • the step of determining the mapping curve of each chip type within the target block error rate range by constructing the MCS and SINR two-dimensional table, according to the MCS and SINR in the two-dimensional table, the corresponding delivery scheduling feedback result ACK/NACK, Then calculate the block error rate of each cell in the two-dimensional table; thus select the first cell whose block error rate meets the target block error rate range, each first cell contains the pair of information of MCS and SINR, according to all For the first cell that meets the requirements, the mapping curve between MCS and SINR is obtained, and the performance of the chip type is reflected through the mapping curve.
  • the characteristic information of the UE is not limited to the chip type shown in the above embodiment, it may also include characteristics such as transmission mode, channel quality and interference power, and different grids are divided according to different characteristic information, and different grids have independent the outer ring value of .
  • the chip type, transmission mode, channel quality and interference power of the UE are divided into grids, the corresponding number of grids can be obtained. According to these grid numbers, multi-dimensional grids can be generated to facilitate data management.
  • step S130 may specifically determine the outer ring of each grid through the following steps.
  • each grid construct an outer ring one-dimensional table, corresponding to the delivery scheduling feedback result ACK/NACK, then calculate the block error rate of each cell in the one-dimensional table, and select the block error rate that satisfies the target block error rate range
  • two cells each second cell includes the two information of the outer ring and the number of samples, and the outer ring value and the number of samples corresponding to each grid are obtained by weighted average method according to all the second cells. That is to say, each grid has a corresponding number of samples and a corresponding outer ring, and the corresponding outer ring value can be quickly found according to the grid.
  • the outer ring value of the class can also be corrected according to the actual situation. Therefore, it is also necessary to adjust the outer ring value of the corresponding class according to the scheduling feedback results of the UE, so that the outer ring value of the class is always kept at the target Within the block error rate range. According to the feedback results in the actual environment, the outer loop value is adaptively adjusted to effectively improve the spectrum efficiency.
  • system performance can also be monitored in real time. If the model effect is found to be inconsistent with expectations, indicating that the grid division at this time does not match the current environment, online grid division learning is started based on the collected network data.
  • the target UE after reading the outer ring value of the class as the initial value of the outer ring of the target UE, it also includes:
  • Each class adjusts the outer ring value of the corresponding class according to the scheduling feedback results, including:
  • the outer ring value of the class is increased according to the unit adjustment amount
  • the outer ring value of the class is reduced according to the unit adjustment amount.
  • each class adjusts the outer ring value of the corresponding class according to the scheduling feedback result, including:
  • the outer ring value of the class remains unchanged.
  • the adjustment method also includes:
  • the adjustment method of the initial value of the outer ring can be understood as mainly including five steps, which are online grid division, basic model determination, model application, model correction and model effect evaluation.
  • the online grid division is performed according to the network data, the basic model under the current grid division is determined, and then the model is applied when the initial value of the outer ring is required, and the model based on the idea of reinforcement learning is carried out according to the environmental feedback after the model is applied. Correct and save the latest model for subsequent application.
  • the system will periodically evaluate the model effect. If the model effect is normal and in line with expectations, continue to apply the model without any processing; if the model effect is not normal, based on the latest network data, online raster division, basic model determination, and model application will be re-performed. , model correction and other steps, the entire system continues to run, and adaptively adjusts the model to adapt to the environment.
  • the method for adjusting the initial value of the outer loop is described.
  • the latest scheduling data is used for online grid division.
  • the online grid division is divided into two parts: UE chip type classification and other feature grid division.
  • Steps S210 to S230 are UE chip type classification, and step S240 is grid division of other features.
  • step S210 the MCS and SINR mapping curves of each chip type within the target block error rate range are respectively obtained, and used as features to represent each chip type. For each chip type, the following steps S211 to S213 are specifically executed.
  • step S220 the slope and intercept of the mapping curve of each chip type are used as a sample point to represent the chip type, and multiple chip types have multiple sample points, and the multiple sample points are clustered.
  • the slope and intercept sample points determine the type of chip belonging to this class.
  • the clustering method selects the K-means clustering algorithm (K-means clustering algorithm), and the number of clusters k is set to 4.
  • K-means clustering algorithm divides the data into k clusters repeatedly according to a certain distance function according to a given set of data points and the required number of clusters k. This process will be repeated until a certain termination condition is met.
  • the clustering process It has the advantages of simplicity and efficiency.
  • step S210 the following steps are specifically included:
  • Step S211 constructing a two-dimensional table of scheduling MCS and inner loop SINR, each level of scheduling MCS is a row of the table, every X 1 dB of inner loop SINR is a column of the table, and X 1 is taken as 2.
  • Step S212 for each MCS, select inner ring SINR cells whose BLER is within the target block error rate range, wherein the target block error rate range of BLER is set to [7%, 13%]. If no cell meets the condition, record the MCS and the corresponding inner ring SINR point pair as (the MCS, NAN); if only one cell meets the condition, record the MCS and the corresponding inner ring SINR point pair as (the MCS, satisfy the condition The median value of the two boundaries of the cell); if more than one cell meets the condition (assuming there are Y), the SINR corresponding to the inner ring of the MCS is:
  • Each MCS gets a point pair (MCS, SINR), and deletes the point pair whose SINR is NAN.
  • the cells with filled colors are within the target block error rate range, BLER not within the target block error rate range or cells that cannot be counted on BLER, and are represented by no filled color .
  • Step S213 according to all valid (MCS, SINR) point pairs, obtain the mapping curve of MCS and SINR through linear fitting, and use the slope and intercept of the mapping curve as features to represent the chip type.
  • Step S240 perform grid division according to such characteristics as UE chip type, transmission mode, channel quality, and interference power.
  • step 250 for each grid, determine the outer ring corresponding to each grid, specifically including:
  • Step 252 select the outer ring cells within the target block error rate range, and set the target block error rate range to [7%, 13%]. If no cell meets the condition, record that the number of samples corresponding to the grid is 0, and the corresponding outer ring is the default value, where the default value is the initial value of the outer ring uniformly configured for all UEs by the traditional AMC strategy; if only one cell meets the condition, Record the number of samples corresponding to the grid as the number of samples of the cell, and the corresponding outer ring is the median value of the two boundaries of the cell that meets the condition); if more than one cell meets the condition (assuming there are Y), the grid corresponds to The number of samples is the sum of the sample numbers of Y cells that meet the conditions, and the corresponding outer ring is:
  • each grid has its corresponding number of samples and its corresponding outer ring.
  • the grid is represented by the corresponding outer ring in the grid, and the specific grid division is determined by using the idea of decision regression tree. Specific steps are as follows:
  • Step 261 traversing all the features and their splittable points, and selecting satisfying The optimal segmentation feature j and the optimal segmentation point s of this feature.
  • x i represents grid i
  • y i represents the corresponding outer ring in grid i
  • R 1 (j,s) ⁇ x
  • R 2 (j,s) ⁇ x
  • x j >s ⁇ means the two classes after being divided
  • x j ⁇ s means all grids with feature j ⁇ s
  • x j > s means all grids with feature j >s
  • num i indicates the number of samples corresponding to grid i.
  • Step 262 use the selected segmentation feature j and segmentation point s to divide the categories, and determine c 1 and c 2 as the outer ring values of the category.
  • Step 264 generate a regression tree, each leaf node is a class, the total number of classes is recorded as Ncluster, each class is numbered, and the class ID is recorded.
  • Step 270 according to the results of the regression tree, establish the mapping relationship between the grid ID and the class ID, and the outer ring value in the class is
  • Step 280 when it is necessary to determine the initial value of the outer loop of the transmission mode of the UE, determine the grid to which it belongs according to the UE chip type, transmission mode, channel quality, and interference power, find the class ID according to the mapping relationship between the grid ID and the class ID, and read The outer ring value in the class is used as the outer ring initial value of the transmission mode of the UE.
  • Step 290 mark the class ID queried in the above process for the UE in this scheduling.
  • Step 300 after receiving the scheduling feedback result marked with the class ID, counting in the corresponding class according to the scheduling feedback result. If the scheduling feedback result is ACK, add 1 to the ACK counter in the corresponding class; if the scheduling feedback result is NACK, add 1 to the NACK counter in the corresponding class.
  • Step 310 after the sum of ACK counters and NACK counters in a class exceeds a certain counting threshold ThrNum, model correction is performed as shown in FIG. 6, and only the latest model is always saved.
  • First calculate the ratio of NACKs in the class to the sum of ACKs and NACKs. If the ratio of NACKs in the class to the sum of ACKs and NACKs is lower than the first threshold BlerLow, the outer ring value in the class is adjusted to a larger adjustment amount AdjustUnit ; If the ratio of NACKs in the class to the sum of ACKs and NACKs exceeds the second threshold BlerHigh, the outer ring value in the class is adjusted to a smaller one unit adjustment amount AdjustUnit. After that, the ACK counter and NACK counter in this class are re-initialized to 0.
  • ThrNum 500, if ThrNum is too small, the calculated NACK ratio will not be statistically significant, if ThrNum is too large, it will be difficult to trigger model correction, and reinforcement learning will not be timely;
  • the latest scheduling data is used for online grid division.
  • the online grid division is divided into two parts: UE chip type classification and other feature grid division.
  • Steps S410 to S430 are UE chip type classification, and step S440 is grid division of other features.
  • step S410 the MCS and SINR mapping curves of each chip type within the target block error rate range are respectively obtained, and used as features to represent each chip type. For each chip type, the following steps S411 to S413 are specifically executed.
  • step S420 the slope and intercept of the mapping curve of each chip type are used as a sample point to represent the chip type, and multiple chip types have multiple sample points, and the multiple sample points are clustered.
  • the slope and intercept sample points determine the type of chip belonging to this class.
  • the clustering method selects the K-means clustering algorithm (K-means clustering algorithm), and the number of clusters k is set to 4.
  • step S410 the following steps are specifically included:
  • Step S411 constructing a two-dimensional table of scheduling MCS and inner loop SINR, each level of scheduling MCS is a row of the table, every X 1 dB of inner loop SINR is a column of the table, and X 1 is taken as 2.
  • Step S412 for each MCS, select inner ring SINR cells within the target block error rate range, where the target block error rate range is set to [7%, 13%]. If no cell meets the condition, record the MCS and the corresponding inner ring SINR point pair as (the MCS, NAN); if only one cell meets the condition, record the MCS and the corresponding inner ring SINR point pair as (the MCS, satisfy the condition The median value of the two boundaries of the cell); if more than one cell meets the condition (assuming there are Y), the SINR corresponding to the inner ring of the MCS is:
  • Each MCS gets a point pair (MCS, SINR), and deletes the point pair whose SINR is NAN.
  • Step S413 according to all valid (MCS, SINR) point pairs, obtain the mapping curve of MCS and SINR through linear fitting, and use the slope and intercept of the mapping curve as features to represent the chip type.
  • Step S440 perform grid division according to characteristics such as UE chip type, transmission mode, channel quality, and interference power.
  • characteristics such as UE chip type, transmission mode, channel quality, and interference power.
  • Step 450 for each grid, determine the outer ring corresponding to each grid, specifically including:
  • Step 452 select the outer ring cells within the target block error rate range, and set the target block error rate range to [7%, 13%]. If no cell meets the condition, record that the number of samples corresponding to the grid is 0, and the corresponding outer ring is the default value, where the default value is the initial value of the outer ring uniformly configured for all UEs by the traditional AMC strategy; if only one cell meets the condition, Record the number of samples corresponding to the grid as the number of samples of the cell, and the corresponding outer ring is the median value of the two boundaries of the cell that meets the condition); if more than one cell meets the condition (assuming there are Y), the grid corresponds to The number of samples is the sum of the sample numbers of Y cells that meet the conditions, and the corresponding outer ring is:
  • each grid has its corresponding number of samples and its corresponding outer ring.
  • the grid is represented by the corresponding outer ring in the grid, and the specific grid division is determined by using the idea of decision regression tree. Specific steps are as follows:
  • Step 461 traversing all the features and their splittable points, and selecting satisfying The optimal segmentation feature j and the optimal segmentation point s of this feature.
  • x i represents grid i
  • y i represents the corresponding outer ring in grid i
  • R 1 (j,s) ⁇ x
  • R 2 (j,s) ⁇ x
  • x j >s ⁇ means the two classes after being divided
  • x j ⁇ s means all grids with feature j ⁇ s
  • x j > s means all grids with feature j >s
  • num i indicates the number of samples corresponding to grid i.
  • Step 462 use the selected segmentation feature j and segmentation point s to divide the categories, and determine c 1 and c 2 as the outer ring values of the category.
  • Step 464 generate a regression tree, each leaf node is a class, the total number of classes is recorded as Ncluster, each class is numbered, and the class ID is recorded.
  • Step 470 according to the results of the regression tree, establish the mapping relationship between the grid ID and the class ID, and the outer ring value in the class is
  • Step 480 when it is necessary to determine the initial value of the outer loop of the transmission mode of the UE, determine the grid to which it belongs according to the UE chip type, transmission mode, channel quality and interference power, find the class ID according to the mapping relationship between the grid ID and the class ID, and read The outer ring value in the class is used as the outer ring initial value of the transmission mode of the UE.
  • Step 490 mark the class ID queried in the above process for the UE in this scheduling.
  • each class performs model correction based on the idea of reinforcement learning according to the scheduling feedback results, that is, adjusts the outer loop value saved by each class, so as to be more suitable for the current actual environment. If the proportion of NACK in the class to the sum of ACK and NACK is lower than the first set value, the outer ring value in the class will be adjusted to a larger value; if the proportion of NACK in the class to the sum of ACK and NACK exceeds the second set value When , the outer ring value in the class will be adjusted to the small one. Otherwise, the outer loop value of the class remains unchanged.
  • Step 500 after receiving the scheduling feedback result marked with the class ID, counting in the corresponding class according to the scheduling feedback result. If the scheduling feedback result is ACK, add 1 to the ACK counter in the corresponding class; if the scheduling feedback result is NACK, add 1 to the NACK counter in the corresponding class.
  • Step 510 after the sum of ACK counters and NACK counters in a certain class exceeds a certain counting threshold ThrNum, model correction is performed as shown in FIG. 6, and only the latest model is always saved.
  • First calculate the ratio of NACKs in the class to the sum of ACKs and NACKs. If the ratio of NACKs in the class to the sum of ACKs and NACKs is lower than the first threshold BlerLow, the outer ring value in the class is adjusted to a larger adjustment amount AdjustUnit ; If the ratio of NACKs in the class to the sum of ACKs and NACKs exceeds the second threshold BlerHigh, the outer ring value in the class is adjusted to a smaller one unit adjustment amount AdjustUnit. After that, the ACK counter and NACK counter in this class are re-initialized to 0.
  • ThrNum 500, if ThrNum is too small, the calculated NACK ratio will not be statistically significant, if ThrNum is too large, it will be difficult to trigger model correction, and reinforcement learning will not be timely;
  • Step 530 in the convergence judgment cycle, sequentially judge the relationship between the ThrWindow NACK ratio and BlerLow and BlerHigh in the statistics window, record the number CounterWin that is not lower than BlerLow and not more than BlerHigh, if CounterWin/ThrWindow>ThrRatio, it is determined that the model is converged , go to step 540, otherwise wait for the next convergence judgment cycle.
  • ThrRatio 60%
  • the convergence judging period is set to 3 hours.
  • the specific statistical judgment steps are as follows:
  • Step 541 initialize the total number of scheduling SchdNumSum counted at the cell level to 0, and the number of NACK counts NackNumSum counted at the cell level to be 0.
  • an embodiment of the present application also provides a device, which includes: a memory, a processor, and a computer program stored in the memory and operable on the processor.
  • the processor and memory can be connected by a bus or other means.
  • memory can be used to store non-transitory software programs and non-transitory computer-executable programs.
  • the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage devices.
  • the memory may include memory located remotely from the processor, which remote memory may be connected to the processor via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
  • the non-transitory software programs and instructions required to implement the method for adjusting the initial value of the outer loop in the above embodiment are stored in the memory, and when executed by the processor, the method for adjusting the initial value of the outer loop in the above embodiment is executed, for example, Execute the above-described method steps S110 to S140 in FIG. 1 , method steps S121 to S123 in FIG. 2 , method steps S210 to S310 in the above-mentioned embodiment, and method steps S410 to S543 in the above-mentioned embodiment.
  • the device embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • an embodiment of the present application also provides a computer-readable storage medium, the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are executed by a processor or a controller, for example, by the above-mentioned Execution by a processor in the terminal embodiment can cause the above-mentioned processor to execute the method for adjusting the initial value of the outer loop in the above-mentioned embodiment, for example, perform the above-described method steps S110 to S140 in FIG. 1 and the method in FIG. 2 Steps S121 to S123, method steps S210 to S310 in the above embodiment, and method steps S410 to S543 in the above embodiment.
  • the method for adjusting the initial value of the outer loop in the embodiment of the present application obtains the characteristic information of multiple user equipments scheduled by the cell, and then divides the characteristic information of the multiple user equipments into grids to obtain multiple grids, and each Determine the outer ring value of the grid respectively, and the outer ring value of the grid meets the target block error rate range; when it is necessary to determine the initial value of the outer ring of the target user equipment, select the target grid according to the characteristic information of the target user equipment, and according to the target
  • the grid reads the corresponding outer ring value as the initial value of the outer ring of the user equipment, that is, the user equipment can determine its own initial value of the outer ring according to the attribution grid, which is more convenient than the scheme of using a unified initial value of the outer ring for all user equipment.
  • Adapting to the current actual wireless environment it is more robust to environmental changes, and the initial value of the outer loop obtained is closer to the convergence value, which is conducive to accelerating the convergence speed of the outer loop and improving
  • Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cartridges, tape, magnetic disk storage or other magnetic storage devices, or can Any other medium used to store desired information and which can be accessed by a computer.
  • communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media .

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Quality & Reliability (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Read Only Memory (AREA)
  • Signal Processing For Digital Recording And Reproducing (AREA)
  • Image Generation (AREA)

Abstract

一种外环初始值的调整方法、设备及可读存储介质，其中，外环初始值的调整方法包括：获取小区调度的多个用户设备的特征信息(110)；对所述多个用户设备的特征信息进行栅格化划分，得到多个栅格(120)；确定每个所述栅格所对应的外环，所述栅格的外环值位于目标误块率范围内(130)；在需要确定目标用户设备的外环初始值时，根据特征信息选择所属的目标栅格(140)，根据目标栅格读取对应的外环值作为目标用户设备的外环初始值(150)。

Description

外环初始值的调整方法、设备及计算机可读存储介质
相关申请的交叉引用
本申请基于申请号为202110719864.9、申请日为2021年06月28日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本申请涉及移动通信技术领域，特别涉及一种外环初始值的调整方法、设备及计算机可读存储介质。
背景技术
无线信道相对有线信道具有更窄的相干带宽，更短的相干时间，因而无线信道变化快，是无线信道的一个显著特点。无线信道若采用固定的调制编码方式，很难充分的利用频谱资源。因此，在无线通信技术中引入了自适应编码调制(Adaptive Modulation and Coding，AMC)技术用来提升频谱效率。
自适应调制编码技术为了实现调制编码策略和用户设备(User Equipment，UE)信道条件的适配，需要根据UE的调度反馈结果ACK/NACK(Acknowledgement/Negative Acknowledgement，肯定的反馈/否定的反馈)来调整外环，以实现系统设置的目标可靠性要求。然而，这一收敛过程往往需要用到几十次甚至上百次传输机会，导致有些UE还没收敛已经传输结束，或者是收敛了但是收敛之前的传输没有采用最佳的调制编码策略，频谱效率受到了限制，UE感知也受到影响。其中，收敛指通过外环的调整使BLER(Block Error Rate,误块率)达到目标要求。
当前的4G(the 4th Generation Mobile Communication Technology,第四代移动通信技术)/5G(the 5th Generation Mobile Communication Technology,第五代移动通信技术)网络中，依然存在大量的小包UE，小包指非连续性突发业务，小包UE的AMC还未收敛业务就已结束，频谱效率受限。为了使小包UE尽快收敛，需要给其一个更接近于收敛状态的外环值作为外环初始值，降低收敛过程所需的传输次数。
发明内容
以下是对本文详细描述的主题的概述。本概述并非是为了限制权利要求的保护范围。
本申请实施例提出一种外环初始值的调整方法、设备及计算机可读存储介质。
第一方面，本申请实施例提供的一种外环初始值的调整方法，包括：获取小区调度的多个用户设备的特征信息；对所述多个用户设备的特征信息进行栅格化划分，得到多个栅格；确定每个所述栅格的外环，所述栅格的外环值位于目标误块率范围内；根据目标用户设备的所述特征信息，确定所述目标用户设备对应的目标栅格，所述目标用户设备为所述小区调度的多个用户设备之一；将所述目标栅格对应的所述外环值确定为所述目标用户设备的外环初 始值。
第二方面，本申请实施例提供的一种设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现如上述第一方面实施例所述的外环初始值的调整方法。
第三方面，本申请实施例提供的一种计算机可读存储介质，存储有计算机可执行指令，所述计算机可执行指令用于执行如上述第一方面实施例所述的外环初始值的调整方法。
本申请的其它特征和优点将在随后的说明书中阐述，并且，部分地从说明书中变得显而易见，或者通过实施本申请而了解。本申请的目的和其他优点可通过在说明书、权利要求书以及附图中所特别指出的结构来实现和获得。
附图说明
附图用来提供对本申请技术方案的进一步理解，并且构成说明书的一部分，与本申请的实施例一起用于解释本申请的技术方案，并不构成对本申请技术方案的限制。
图1是本申请一实施例提供的外环初始值的调整方法的流程图；
图2是本申请一实施例提供的根据特征信息划分栅格的流程图；
图3是本申请另一实施例提供的外环初始值的调整方法的流程图；
图4是本申请一实施例UE的芯片类型分类中构建二维表格的示意图；
图5是本申请一实施例每一栅格构建外环一维表格的示意图；
图6是本申请一实施例模型修正的流程图；
图7是本申请一实施例栅格与类映射关系的示意图。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。
需要说明的是，虽然在装置示意图中进行了功能模块划分，在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于装置中的模块划分，或流程图中的顺序执行所示出或描述的步骤。说明书、权利要求书或上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。
在一些情形下，AMC是无线信道上采用的自适应的编码调制技术，通过调整无线链路传输的调制方式与编码速率，来确保链路的传输质量。当信道条件较差时，选择较小的调制方式与编码速率。当信道条件较好时，选择较大的调制方式，从而最大化了传输速率。当编码技术和调试方式调制后，信道的质量也会得到相应的改善。在AMC的调整过程中，系统总是希望传输的数据速率与信道变化的趋势一致，从而最大化地利用无线信道的传输能力。
AMC技术中编码调制方式的选择采用两个控制环来实现，包括内环(Inner-Loop)和外环(Outer-Loop)。内环是指基站根据UE的信道条件确定该终端上下行信号的MCS(Modulation And Coding Scheme，调制与编码策略)。外环是基站根据信号传输的BLER来反馈调整UE的MCS，尽量确保将信号传输的实际BLER维持在目标BLER附近的过程。为了实现调制编码策略和UE信道条件的适配，需要根据UE的调度反馈结果ACK/NACK来调整外环，以实现系统设置的目标可靠性要求。
通过外环的调整使BLER达到目标BLER要求这一过程称为收敛，外环也可理解为外环值。然而，由于所有UE共用统一的外环初始值，这一收敛过程往往需要用到几十次甚至上百次传输机会，导致有些UE还没收敛已经传输结束，或者是收敛了但是收敛之前的传输没有采用最佳的调制编码策略，频谱效率受到了限制，UE感知也受到影响。
而且在当前的4G/5G网络中，依然存在大量的小包UE，小包指非连续性突发业务，小包UE的AMC还未收敛业务就已结束，频谱效率受限。为了使小包UE尽快收敛，需要给其一个更接近于收敛状态的外环值作为外环初始值，降低收敛过程所需的传输次数。
基于此，本申请提出一种AMC外环初始值的调整方法，UE根据归属栅格确定各自的外环初始值，而不是给所有UE统一的外环初始值；同时本申请能够根据实际环境中的调度反馈结果自适应调整各栅格的外环值；还可以对系统性能实时监控，如果发现栅格划分与当前环境不相符，则根据收集到的网络数据启动在线栅格划分的学习，更加适配当前实际无线环境，对环境变化的鲁棒性更强，得到的外环初始值更加接近于收敛值，有利于加快外环收敛速度，提高频谱效率。
下面将结合附图对本申请的技术方案进行清楚、完整的描述，显然，以下所描述的实施例是本申请一部分实施例，并非全部实施例。
参见图1所示，图1是本申请一个实施例提供的外环初始值的调整方法的流程图，该调整方法包括但不限于以下步骤：
步骤S110，获取小区调度的多个UE的特征信息；
步骤S120，对多个用户设备的特征信息进行栅格化划分，得到多个栅格；
步骤S130，确定每个栅格的外环，栅格的外环值位于目标误块率范围内；
步骤S140，根据目标用户设备的所述特征信息，确定所述目标用户设备对应的目标栅格，所述目标用户设备为所述小区调度的多个用户设备之一；
步骤S150，将所述目标栅格对应的所述外环值确定为所述目标用户设备的外环初始值。
可以理解的是，UE的不同特征信息对应的收敛过程具有不同的调整量，为了使UE尽快收敛，需要给予更接近于收敛状态的外环值作为外环初始值，从而能够降低收敛过程所需的传输次数。其中，UE可以是手机、平板电脑、笔记本电脑等。
在上述调整方法中，利用小区中收集到的调度数据，通过该调度数据获取到多个UE的特征信息，该特征信息可以包括UE的类型、信道质量等特征。然后对多个UE的特征信息进行栅格化划分，每个特征信息分别对应有栅格，这样通过栅格能够表征每个特征信息，建立栅格与相应特征信息之间的关系，通过栅格能够便于对UE的特征信息的管理。根据实施例的方法所确定的外环初始值会给UE使用，从而使UE能够具有接近于收敛状态的外环初始值。
在步骤S130中，每个栅格需要确定对应的外环，可理解到，AMC技术根据UE的反馈来调整外环，外环是根据BLER来反馈调整UE的MCS，确保将实际BLER维持在目标误块率范围。因此，本申请实施例中通过栅格表征UE的特征信息，每个栅格对应有外环，通过确定每个栅格对应的外环，使栅格的外环值位于目标误块率范围内，这样能够得到每个栅格的外环值，每个栅格具有独立的外环值。
在步骤S140中，在需要确定目标UE的外环初始值时，根据目标UE的特征信息选择所对应的目标栅格，然后根据目标栅格读取对应的外环值作为目标UE的外环初始值，即用户 设备根据归属栅格能够确定各自的外环初始值，相对于传统的所有UE共用统一的外环初始值的方案，更加适配当前实际无线环境，对环境变化的鲁棒性更强，得到的外环初始值更加接近于收敛值，有利于加快外环收敛速度，提高频谱效率。
其中，以UE的芯片类型作为特征信息，步骤S120中，对多个用户设备的特征信息进行栅格化划分，包括对所述芯片类型进行栅格化划分，参见图2所示，进一步包括以下步骤：
步骤S121，确定每个芯片类型在目标误块率范围内的MCS与SINR(Signal to Interference plus Noise Ratio，信号与干扰加噪声比)的映射曲线；
步骤S122，以映射曲线表征芯片类型，对所有芯片类型进行聚类，得到芯片类型的分类结果；
步骤S123，根据芯片类型的分类结果对芯片类型进行栅格化划分。
可理解到，利用MCS与SINR的映射曲线表示每个芯片类型的解调能力，然后根据映射曲线的斜率和截距作为一个样本点表示映射曲线，进而表示芯片类型，采用聚类方式将样本点对象的集合分成多个类别，将UE的芯片类型按照聚类的结构进行栅格划分，从而得到芯片类型所对应的栅格。以手机为示例，不同的手机具有不同的芯片类型，如基带芯片，采用MCS与SINR的映射曲线的方式，将手机的芯片类型以映射曲线的形式表示出来，即通过映射曲线能够反映不同手机的解调能力。
其中，在确定每个芯片类型在目标误块率范围内的映射曲线的步骤，通过构建MCS与SINR二维表格的方式，根据MCS与SINR在二维表格内对应投递调度反馈结果ACK/NACK，然后计算出二维表格中每个单元格的误块率；从而选取误块率满足目标误块率范围的第一单元格，每个第一单元格包含MCS、SINR这一对信息，根据所有满足要求的第一单元格，得到MCS与SINR的映射曲线，通过映射曲线反映芯片类型的性能。
考虑到，UE的特征信息不限于上述实施例所示的芯片类型，还可以包括传输模式、信道质量和干扰功率这些特征，根据不同的特征信息进行不同栅格的划分，不同的栅格具有独立的外环值。对UE的芯片类型、传输模式、信道质量和干扰功率划分栅格后得到相应的栅格数量，根据这些栅格数量可生成多维栅格，便于数据管理。
在一些实施例中，步骤S130可具体通过以下步骤确定每个栅格的外环。在每个栅格内，构建外环一维表格，对应投递调度反馈结果ACK/NACK，然后计算一维表格中每个单元格的误块率，选取误块率满足目标误块率范围的第二单元格，每个第二单元格包括外环和样本数两个信息，根据所有第二单元格通过加权平均法得到每个栅格对应的外环值和样本数。也就是说，每个栅格内具有对应的样本数和对应的外环，根据栅格可快速查找到相应的外环值。
需要说明的是，在后续在线调整外环值过程中，为了使相似的不同栅格可以混合计数，共同更新以提高效率和准确度，需要对栅格进行类划分，将相似栅格聚集为一类，建立栅格到类的映射关系，确定类的外环值。类划分可采用回归树(regression tree)技术进行，具体过程参见下面的示例。
需要确定UE的外环初始值时，根据UE的特征信息查找所属栅格，根据栅格到类的映射关系，读取类的外环值作为UE的外环初始值。
由于收集的网络数据会不断更新，类的外环值也可根据实际情景进行修正，因此，还需要根据UE的调度反馈结果调整对应类的外环值，使得类的外环值始终保持在目标误块率范围内。根据实际环境中的反馈结果自适应调整外环值，有效提升频谱效率。
可理解到，还可以对系统性能实时监控，如果发现模型效果与预期不相符，表示此时的栅格划分与当前环境不相符，则根据收集到的网络数据启动在线栅格划分的学习。
其中，读取类的外环值作为目标UE的外环初始值之后，还包括：
每个类根据调度反馈结果调整对应类的外环值，包括：
根据调度反馈结果在对应的类进行计数；
类内的NACK占ACK与NACK之和的比例少于第一设定值时，根据单位调整量增大类的外环值；
类内的NACK占ACK与NACK之和的比例大于第二设定值时，根据单位调整量减小类的外环值。
其中，每个类根据调度反馈结果调整对应类的外环值，还包括：
类内的NACK占ACK与NACK之和的比例大于或等于第一设定值且小于或等于第二设定值时，类的外环值保持不变。
其中，调整方法还包括：
根据所有UE的调度反馈结果计算误块率；
判断误块率是否满足预设值，若否，重新执行对所述多个用户设备的特征信息进行栅格化划分，并重新确定目标UE的外环初始值。
参见图3所示，外环初始值的调整方法可理解为主要包括五个步骤，分别是在线栅格划分、基础模型确定、模型应用、模型修正和模型效果评估。该调整方法启动后根据网络数据进行在线栅格划分，确定当前栅格划分下的基础模型，然后在需要外环初始值时应用模型，并根据模型应用后的环境反馈进行基于强化学习思想的模型修正，并保存最新模型以便后续的应用。
系统会周期性的进行模型效果评估，如果模型效果正常符合预期，不做任何处理，继续进行模型应用；如果模型效果不正常，基于最新网络数据重新进行在线栅格划分、基础模型确定、模型应用、模型修正等步骤，整个系统持续运行，自适应的调整模型适配环境。
以UE的芯片类型、传输模式、信道质量和干扰功率这些特征为示例，对外环初始值的调整方法进行描述。
示例一：
首先利用最新的调度数据进行在线栅格划分，在线栅格划分分为UE芯片类型分类和其它特征栅格划分两部分。其中步骤S210至步骤S230为UE芯片类型分类，步骤S240为其它特征的栅格划分。
步骤S210，分别获取各芯片类型在目标误块率范围内的MCS与SINR映射曲线，作为特征来表示各芯片类型。对于每一种芯片类型，具体执行下面的步骤S211至步骤S213。
步骤S220，将每个芯片类型的映射曲线的斜率、截距作为一个样本点表示该芯片类型，多种芯片类型则有多个样本点，对多个样本点进行聚类，根据每一类内的斜率、截距样本点确定属于该类的芯片类型。
其中，聚类方法选择k均值聚类算法(K-means clustering algorithm)，聚类数目k取值为4。k均值聚类算法根据给定数据点集合和需要的聚类数目k，根据某个距离函数反复把数据分入k个聚类中，这个过程将不断重复直到满足某个终止条件，聚类过程具有简洁和效率等优点。
步骤S230，异常处理。如果某个芯片类型有效的(MCS，SINR)点对小于ThrDot，或线性拟合得到的MCS与SINR的映射曲线的斜率大于HighThrSlope，或线性拟合得到的MCS与SINR的映射曲线的斜率小于LowThrSlope，该芯片类型不参与聚类，所有不参与聚类的芯片类型属于一类。其中ThrDot＝4、HighThrSlope＝2.5、LowThrSlope＝0.5。
参见图4所示，在步骤S210中，具体包括以下步骤：
步骤S211，构建调度MCS、内环SINR二维表格，调度MCS每一阶为表格的一行，内环SINR每X 1dB为表格的一列，X 1取2。该二维表格内投递对应调度MCS、内环SINR的调度反馈结果ACK/NACK，统计每个单元格内的样本数量并计算BLER，BLER＝NACK/(ACK+NACK)。
步骤S212，对于每个MCS，挑选BLER在目标误块率范围内的内环SINR单元格，其中，BLER的目标误块率范围设置为[7％,13％]。如果没有单元格满足条件，记录该MCS与对应内环SINR点对为(该MCS，NAN)；如果只有一个单元格满足条件，记录该MCS与对应内环SINR点对为(该MCS，满足条件的单元格两边界的中值)；如果有多于一个单元格满足条件(假设有Y个)，该MCS对应内环SINR为：
Figure PCTCN2022098970-appb-000001
每个MCS得到一个点对(MCS，SINR)，删除SINR为NAN的点对。
需要说明的是，图4所示的表格中，具有填充色的单元格为在目标误块率范围内，不在目标误块率范围内的BLER或无法统计BLER的单元格，以无填充色表示。
步骤S213，根据所有有效的(MCS，SINR)点对，通过线性拟合得到MCS与SINR的映射曲线，以映射曲线的斜率和截距作为特征表示该芯片类型。
步骤S240，根据UE芯片类型、传输模式、信道质量、干扰功率这些特征进行栅格划分。其中UE的芯片类型维度的栅格个数Nchiptype＝K+1＝5；上行传输时，传输模式有两种(单端口、闭环空分复用)，Ntrans＝2；下行传输时，传输模式有三种(最佳波束、闭环空分复用、波束赋形)，Ntrans＝3；信道质量等间隔划分为Nsinr＝10个栅格；干扰功率等间隔划分为Nni＝10个栅格；生成四维栅格共计Nmesh个，其中Nmesh＝Nchiptype×Ntrans×Nri×Nsinr×Nni＝2000(上行)或6000(下行)，并为每一个栅格编号，记录栅格ID。
参见图5所示，步骤250，对于每个栅格，确定每个栅格所对应的外环，具体包括：
步骤251，构建外环一维表格，外环每X 2dB为一个单元格，X 2＝2。一维表格内按照对应栅格特征和外环投递对应调度反馈结果ACK/NACK，统计每个单元格内的样本数量并计算BLER，BLER＝NACK/(ACK+NACK)。
步骤252，挑选在目标误块率范围内的外环单元格，目标误块率范围设置为[7％,13％]。如果没有单元格满足条件，记录该栅格对应样本数为0、对应外环为默认值，其中默认值为传统AMC策略给所有UE统一配置的外环初始值；如果只有一个单元格满足条件，记录该栅格对应样本数为该单元格样本数、对应外环为满足条件的单元格两边界的中值)；如果有多于一个单元格满足条件(假设有Y个)，该栅格对应样本数为满足条件的Y个单元格的样本数量之和、对应外环为：
Figure PCTCN2022098970-appb-000002
步骤260，每个栅格内有其对应的样本数和对应的外环。以栅格内对应的外环来表征栅格，利用决策回归树思想确定具体的栅格划分。具体步骤如下：
步骤261，遍历所有特征及其可切分点，选择满足
Figure PCTCN2022098970-appb-000003
的最优的切分特征j和该特征最优的切分点s。其中x i表示栅格i，y i表示栅格i内对应的外环；R 1(j,s)＝{x|x j≤s}、R 2(j,s)＝{x|x j＞s}表示被划分后的两个类，x j≤s表示特征j≤s的所有栅格，x j＞s表示特征j＞s的所有栅格；
Figure PCTCN2022098970-appb-000004
表示属于第一个类的所有栅格对应外环的加权平均值，
Figure PCTCN2022098970-appb-000005
表示属于第二个类的所有栅格对应外环的加权平均值，num i表示栅格i对应的样本数。
步骤262，用选定的切分特征j和切分点s划分类别，并确定c 1、c 2为该类的外环值。
步骤263，继续对两个子类调用步骤261和262，直至满足停止条件，其中停止条件为任意类内的均方误差
Figure PCTCN2022098970-appb-000006
都小于门限值ThrMse或类个数超过限制ThrNcluster，其中，ThrMse＝0.5、ThrNcluster＝100。
步骤264，生成回归树，每一个叶子节点为一类，类的总数记为Ncluster，为每个类进行编号，记录类ID。
对于基础模型的确定，包括以下步骤：
步骤270，按照回归树结果，建立栅格ID与类ID的映射关系，类内的外环值为
Figure PCTCN2022098970-appb-000007
参见图7所示，假设所有栅格平铺，相同填充的栅格为同一类，建立栅格到类的映射关系。
对于模型应用，包括以下步骤：
步骤280，需要确定UE的传输模式的外环初始值时，根据UE芯片类型、传输模式、信道质量和干扰功率确定所属栅格，根据栅格ID与类ID的映射关系找到类ID，读取类内的外环值作为该UE的传输模式的外环初始值。
步骤290，为UE在该次调度标记上述过程中所查询到的类ID。
对于模型修正，包括以下步骤：
步骤300，收到带有类ID标记的调度反馈结果后，根据调度反馈结果在对应类内计数。如果调度反馈结果为ACK，对应类内的ACK计数器加1；如果调度反馈结果为NACK，对应类内的NACK计数器加1。
步骤310，某个类内ACK计数器与NACK计数器之和超过某一计数门限ThrNum后， 参见图6所示进行模型修正，始终只保存最新模型。首先，计算该类内NACK占ACK与NACK之和的比例，如果类内NACK占ACK与NACK之和的比例低于第一门限BlerLow，该类内的外环值向大调整一个单位调整量AdjustUnit；如果类内NACK占ACK与NACK之和的比例超过第二门限BlerHigh，该类内的外环值向小调整一个单位调整量AdjustUnit。之后该类内的ACK计数器、NACK计数器重新初始化为0。
其中ThrNum＝500，ThrNum过小会导致计算的NACK占比不具有统计意义，ThrNum过大会导致难以触发模型修正，强化学习不及时；BlerLow和BlerHigh表示设定目标值的上下波动限制，基于目标BLER(NACK占比)通常为10％，允许上下波动4％的情况下，可设置BlerLow＝6％、BlerHigh＝14％；AdjustUnit过大会影响强化学习的稳定性，AdjustUnit过小会导致模型修正不及时，通常设置AdjustUnit＝0.7(当外环单位为dB时)或1(当外环单位为阶数时)。
示例二：
首先利用最新的调度数据进行在线栅格划分，在线栅格划分分为UE芯片类型分类和其它特征栅格划分两部分。其中步骤S410至步骤S430为UE芯片类型分类，步骤S440为其它特征的栅格划分。
步骤S410，分别获取各芯片类型在目标误块率范围内的MCS与SINR映射曲线，作为特征来表示各芯片类型。对于每一种芯片类型，具体执行下面的步骤S411至步骤S413。
步骤S420，将每个芯片类型的映射曲线的斜率、截距作为一个样本点表示该芯片类型，多种芯片类型则有多个样本点，对多个样本点进行聚类，根据每一类内的斜率、截距样本点确定属于该类的芯片类型。其中，聚类方法选择k均值聚类算法(K-means clustering algorithm)，聚类数目k取值为4。
步骤S430，异常处理。如果某个芯片类型有效的(MCS，SINR)点对小于ThrDot，或线性拟合得到的MCS与SINR的映射曲线的斜率大于HighThrSlope，或线性拟合得到的MCS与SINR的映射曲线的斜率小于LowThrSlope，该芯片类型不参与聚类，所有不参与聚类的芯片类型属于一类。其中ThrDot＝4、HighThrSlope＝2.5、LowThrSlope＝0.5。
在步骤S410中，具体包括以下步骤：
步骤S411，构建调度MCS、内环SINR二维表格，调度MCS每一阶为表格的一行，内环SINR每X 1dB为表格的一列，X 1取2。该二维表格内投递对应调度MCS、内环SINR的调度反馈结果ACK/NACK，统计每个单元格内的样本数量并计算BLER，BLER＝NACK/(ACK+NACK)。
步骤S412，对于每个MCS，挑选在目标误块率范围内的内环SINR单元格，其中，目标误块率范围设置为[7％,13％]。如果没有单元格满足条件，记录该MCS与对应内环SINR点对为(该MCS，NAN)；如果只有一个单元格满足条件，记录该MCS与对应内环SINR点对为(该MCS，满足条件的单元格两边界的中值)；如果有多于一个单元格满足条件(假设有Y个)，该MCS对应内环SINR为：
Figure PCTCN2022098970-appb-000008
每个MCS得到一个点对(MCS，SINR)，删除SINR为NAN的点对。
步骤S413，根据所有有效的(MCS，SINR)点对，通过线性拟合得到MCS与SINR的映射曲线，以映射曲线的斜率和截距作为特征表示该芯片类型。
步骤S440，根据UE芯片类型、传输模式、信道质量、干扰功率这些特征进行栅格划分。其中UE的芯片类型维度的栅格个数Nchiptype＝K+1＝5；上行传输时，传输模式有两种(单端口、闭环空分复用)，Ntrans＝2；下行传输时，传输模式有三种(最佳波束、闭环空分复用、波束赋形)，Ntrans＝3；信道质量等间隔划分为Nsinr＝10个栅格；干扰功率等间隔划分为Nni＝10个栅格；生成四维栅格共计Nmesh个，其中Nmesh＝Nchiptype×Ntrans×Nri×Nsinr×Nni
＝2000(上行)或6000(下行)，并为每一个栅格编号，记录栅格ID。
步骤450，对于每个栅格，确定每个栅格所对应的外环，具体包括：
步骤451，构建外环一维表格，外环每X 2dB为一个单元格，X 2＝2。一维表格内按照对应栅格特征和外环投递对应调度反馈结果ACK/NACK，统计每个单元格内的样本数量并计算BLER，BLER＝NACK/(ACK+NACK)。
步骤452，挑选在目标误块率范围内的外环单元格，目标误块率范围设置为[7％,13％]。如果没有单元格满足条件，记录该栅格对应样本数为0、对应外环为默认值，其中默认值为传统AMC策略给所有UE统一配置的外环初始值；如果只有一个单元格满足条件，记录该栅格对应样本数为该单元格样本数、对应外环为满足条件的单元格两边界的中值)；如果有多于一个单元格满足条件(假设有Y个)，该栅格对应样本数为满足条件的Y个单元格的样本数量之和、对应外环为：
Figure PCTCN2022098970-appb-000009
步骤460，每个栅格内有其对应的样本数和对应的外环。以栅格内对应的外环来表征栅格，利用决策回归树思想确定具体的栅格划分。具体步骤如下：
步骤461，遍历所有特征及其可切分点，选择满足
Figure PCTCN2022098970-appb-000010
的最优的切分特征j和该特征最优的切分点s。其中x i表示栅格i，y i表示栅格i内对应的外环；R 1(j,s)＝{x|x j≤s}、R 2(j,s)＝{x|x j＞s}表示被划分后的两个类，x j≤s表示特征j≤s的所有栅格，x j＞s表示特征j＞s的所有栅格；
Figure PCTCN2022098970-appb-000011
表示属于第一个类的所有栅格对应外环的加权平均值，
Figure PCTCN2022098970-appb-000012
表示属于第二个类的所有栅格对应外环的加权平均值，num i表示栅格i对应的样本数。
步骤462，用选定的切分特征j和切分点s划分类别，并确定c 1、c 2为该类的外环值。
步骤463，继续对两个子类调用步骤261和262，直至满足停止条件，其中停止条件为任 意类内的均方误差
Figure PCTCN2022098970-appb-000013
都小于门限值ThrMse或类个数超过限制ThrNcluster，其中，ThrMse＝0.5、ThrNcluster＝100。
步骤464，生成回归树，每一个叶子节点为一类，类的总数记为Ncluster，为每个类进行编号，记录类ID。
对于基础模型的确定，包括以下步骤：
步骤470，按照回归树结果，建立栅格ID与类ID的映射关系，类内的外环值为
Figure PCTCN2022098970-appb-000014
对于模型应用，包括以下步骤：
步骤480，需要确定UE的传输模式的外环初始值时，根据UE芯片类型、传输模式、信道质量和干扰功率确定所属栅格，根据栅格ID与类ID的映射关系找到类ID，读取类内的外环值作为该UE的传输模式的外环初始值。
步骤490，为UE在该次调度标记上述过程中所查询到的类ID。
对于模型修正，每个类根据调度反馈结果基于强化学习思想进行模型修正，即调整各自所保存的外环值，从而更加适配当前实际环境。如果类内NACK占ACK与NACK之和的比例低于第一设定值时，类内的外环值会向大调整；如果类内NACK占ACK与NACK之和的比例超过第二设定值时，类内的外环值会向小调整。否则，类的外环值保持不变。
具体来说，包括以下步骤：
步骤500，收到带有类ID标记的调度反馈结果后，根据调度反馈结果在对应类内计数。如果调度反馈结果为ACK，对应类内的ACK计数器加1；如果调度反馈结果为NACK，对应类内的NACK计数器加1。
步骤510，某个类内ACK计数器与NACK计数器之和超过某一计数门限ThrNum后，参见图6所示进行模型修正，始终只保存最新模型。首先，计算该类内NACK占ACK与NACK之和的比例，如果类内NACK占ACK与NACK之和的比例低于第一门限BlerLow，该类内的外环值向大调整一个单位调整量AdjustUnit；如果类内NACK占ACK与NACK之和的比例超过第二门限BlerHigh，该类内的外环值向小调整一个单位调整量AdjustUnit。之后该类内的ACK计数器、NACK计数器重新初始化为0。
其中ThrNum＝500，ThrNum过小会导致计算的NACK占比不具有统计意义，ThrNum过大会导致难以触发模型修正，强化学习不及时；BlerLow和BlerHigh表示设定目标值的上下波动限制，基于目标BLER(NACK占比)通常为10％，允许上下波动4％的情况下，可设置BlerLow＝6％、BlerHigh＝14％；AdjustUnit过大会影响强化学习的稳定性，AdjustUnit过小会导致模型修正不及时，通常设置AdjustUnit＝0.7(当外环单位为dB时)或1(当外环单位为阶数时)。
对于模型效果评估，包括以下步骤：
步骤520，模型修正中，某个类内ACK计数器与NACK计数器之和超过某一计数门限ThrNum后需要计算该类内NACK占ACK/NACK之和的比例，据此进行周期性的模型收敛性判断。不区分类ID，滑窗统计最新的ThrWindow个计算得到的NACK比例。其中 ThrWindow＝50。
步骤530，收敛性判断周期中，依次判断统计窗中ThrWindow个NACK比例与BlerLow、BlerHigh的关系，记录不低于BlerLow且不超过BlerHigh的个数CounterWin，如果CounterWin/ThrWindow>ThrRatio，判定为模型收敛，进入步骤540，否则等待下一收敛性判断周期。其中ThrRatio＝60％、收敛性判断周期设置为3小时。
步骤540，限定统计时长内，只统计每个UE的前NumThr次新传调度的调度结果(ACK/NACK)，所有UE的调度结果混合计算BLER(NACK比例)。其中NumThr＝50，统计时长限定为30分钟。根据统计得到的BLER判断模型效果是否正常，如果认为模型效果正常，继续进行模型应用、模型修正、模型效果评估；如果认为模型效果不正常，触发第一步在线栅格划分，然后基于新的划分结果确定基础模型，进行模型应用、模型修正、模型效果评估。具体统计判断步骤如下：
步骤541，初始化小区级统计的总调度次数SchdNumSum为0、小区级统计的NACK次数NackNumSum为0。
步骤542，限定统计时长内，对于所有新接入UE，UE实例建立后，初始化UE级计数器NumCounter为0。UE每次调度后，如果是新传调度，判断NumCounter是否不超过NumThr，如果满足，令SchdNumSum＝SchdNumSum+1，进一步判断该次调度的调度结果是否是NACK，如果是NACK，令NackNumSum＝NackNumSum+1。
步骤543，限定统计时长到，计算小区级BLER＝NackNumSum/SchdNumSum，如果小区级BLER低于门限BlerLow，认为模型效果不正常；如果小区级BLER超过门限BlerHigh，认为模型效果不正常；否则，认为模型效果正常。
另外，本申请的实施例还提供了一种设备，该设备包括：存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序。处理器和存储器可以通过总线或者其他方式连接。
存储器作为一种非暂态计算机可读存储介质，可用于存储非暂态软件程序以及非暂态性计算机可执行程序。此外，存储器可以包括高速随机存取存储器，还可以包括非暂态存储器，例如至少一个磁盘存储器件、闪存器件、或其他非暂态固态存储器件。在一些实施方式中，存储器可包括相对于处理器远程设置的存储器，这些远程存储器可以通过网络连接至该处理器。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
实现上述实施例的外环初始值的调整方法所需的非暂态软件程序以及指令存储在存储器中，当被处理器执行时，执行上述实施例中的外环初始值的调整方法，例如，执行以上描述的图1中的方法步骤S110至S140、图2中的方法步骤S121至S123、上述实施例中的方法步骤S210至S310、上述实施例中的方法步骤S410至步骤S543。
以上所描述的装置实施例仅仅是示意性的，其中作为分离部件说明的单元可以是或者也可以不是物理上分开的，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。
此外，本申请的一个实施例还提供了一种计算机可读存储介质，该计算机可读存储介质存储有计算机可执行指令，该计算机可执行指令被一个处理器或控制器执行，例如，被上述终端实施例中的一个处理器执行，可使得上述处理器执行上述实施例中的外环初始值的调整方法，例如，执行以上描述的图1中的方法步骤S110至S140、图2中的方法步骤S121至S123、上述实施例中的方法步骤S210至S310、上述实施例中的方法步骤S410至步骤S543。
本申请实施例的外环初始值的调整方法通过获取小区调度的多个用户设备的特征信息，然后对多个用户设备的特征信息进行栅格化划分，得到多个栅格，并对每个栅格分别确定外环值，栅格的外环值满足目标误块率范围；在需要确定目标用户设备的外环初始值时，根据目标用户设备的特征信息选择所属的目标栅格，根据目标栅格读取对应的外环值作为用户设备的外环初始值，即用户设备根据归属栅格能够确定各自的外环初始值，相对于所有用户设备用统一的外环初始值的方案，更加适配当前实际无线环境，对环境变化的鲁棒性更强，得到的外环初始值更加接近于收敛值，有利于加快外环收敛速度，提高频谱效率。
本领域普通技术人员可以理解，上文中所公开方法中的全部或某些步骤、系统可以被实施为软件、固件、硬件及其适当的组合。某些物理组件或所有物理组件可以被实施为由处理器，如中央处理器、数字信号处理器或微处理器执行的软件，或者被实施为硬件，或者被实施为集成电路，如专用集成电路。这样的软件可以分布在计算机可读介质上，计算机可读介质可以包括计算机存储介质(或非暂时性介质)和通信介质(或暂时性介质)。如本领域普通技术人员公知的，术语计算机存储介质包括在用于存储信息(诸如计算机可读指令、数据结构、程序模块或其他数据)的任何方法或技术中实施的易失性和非易失性、可移除和不可移除介质。计算机存储介质包括但不限于RAM、ROM、EEPROM、闪存或其他存储器技术、CD-ROM、数字多功能盘(DVD)或其他光盘存储、磁盒、磁带、磁盘存储或其他磁存储装置、或者可以用于存储期望的信息并且可以被计算机访问的任何其他的介质。此外，本领域普通技术人员公知的是，通信介质通常包含计算机可读指令、数据结构、程序模块或者诸如载波或其他传输机制之类的调制数据信号中的其他数据，并且可包括任何信息递送介质。
以上是对本申请的若干实施方式进行了具体说明，但本申请并不局限于上述实施方式，熟悉本领域的技术人员在不违背本申请精神的前提下还可作出种种的等同变形或替换，这些等同的变形或替换均包含在本申请权利要求所限定的范围内。

Claims (14)

  1. 一种外环初始值的调整方法，包括：
    获取小区调度的多个用户设备的特征信息；
    对所述多个用户设备的特征信息进行栅格化划分，得到多个栅格；
    确定每个所述栅格的外环，所述栅格的外环值位于目标误块率范围内；
    根据目标用户设备的所述特征信息，确定所述目标用户设备对应的目标栅格，所述目标用户设备为所述小区调度的多个用户设备之一；
    将所述目标栅格对应的所述外环值确定为所述目标用户设备的外环初始值。
  2. 根据权利要求1所述的外环初始值的调整方法，其中，所述特征信息包括用户设备的芯片类型，所述对所述多个用户设备的特征信息进行栅格化划分，包括：
    确定每个所述芯片类型在所述目标误块率范围内的调制与编码策略MCS和信号与干扰加噪声比SINR的映射曲线；
    以所述映射曲线表征所述芯片类型，对所有所述芯片类型进行聚类，得到芯片类型的分类结果；
    根据所述芯片类型的分类结果对所述芯片类型进行栅格化划分。
  3. 根据权利要求2所述的外环初始值的调整方法，其中，所述确定每个所述芯片类型在所述目标误块率范围内的调制与编码策略MCS和信号与干扰加噪声比SINR的映射曲线，包括：
    构建所述MCS与所述SINR的二维表格，根据所述MCS与所述SINR在所述二维表格内对应投递调度反馈结果，所述调度反馈结果包括肯定的反馈ACK/否定的反馈NACK；
    根据所述调度反馈结果ACK/NACK计算所述二维表格中每个单元格的误块率；
    选取所述误块率满足所述目标误块率范围的第一单元格，每个所述第一单元格包含MCS、SINR这一对信息，根据所有所述第一单元格，得到所述MCS与所述SINR的映射曲线。
  4. 根据权利要求2所述的外环初始值的调整方法，其中，所述特征信息还包括传输模式、信道质量和干扰功率，所述对所述多个用户设备的特征信息进行栅格化划分，还包括：
    对所述传输模式、所述信道质量和所述干扰功率分别进行栅格化划分，得到相应的多个栅格；
    根据所述芯片类型、所述传输模式、所述信道质量和所述干扰功率所对应的栅格数量，生成多维栅格。
  5. 根据权利要求1所述的外环初始值的调整方法，其中，所述确定每个所述栅格的外环，包括：
    构建外环一维表格，根据所述栅格与所述外环在所述外环一维表格内对应投递调度反馈结果ACK/NACK；
    根据所述调度反馈结果ACK/NACK计算所述外环一维表格中每个单元格的误块率；
    选取所述误块率满足所述目标误块率范围的第二单元格，每个所述第二单元格包括外环和样本数两个信息，根据所有所述第二单元格通过加权平均法得到每个所述栅格的外环值和样本数。
  6. 根据权利要求5所述的外环初始值的调整方法，其中，所述外环初始值的调整方法， 还包括：
    根据决策回归树方法对所有所述栅格进行聚类，得到多个类；
    对每个所述栅格进行编号，得到栅格身份识别号；
    对得到的每个所述类进行编号，得到类身份识别号；
    建立所述栅格身份识别号与所述类身份识别号的映射关系，所述类的外环值根据与所述类对应的所有所述栅格的外环值和样本数得到。
  7. 根据权利要求6所述的外环初始值的调整方法，其中，所述类的外环值根据与所述类对应的所有所述栅格对应的外环值和样本数得到，包括：
    根据所述栅格身份识别号与所述类身份识别号的映射关系，确定每个所述类包含的所有所述栅格，根据所述所有所述栅格的外环值和样本数通过加权平均法计算得到所述类的外环值。
  8. 根据权利要求7所述的外环初始值的调整方法，其中，所述将所述目标栅格对应的所述外环值确定为所述目标用户设备的外环初始值，包括：
    根据所述栅格身份识别号与所述类身份识别号的映射关系查找所述目标栅格对应的类，读取所述类的外环值作为所述目标用户设备的外环初始值。
  9. 根据权利要求8所述的外环初始值的调整方法，其中，所述读取所述类的外环值作为所述目标用户设备的外环初始值之后，包括：
    每个所述类根据所述调度反馈结果调整所述类的外环值。
  10. 根据权利要求9所述的外环初始值的调整方法，其中，所述每个所述类根据所述调度反馈结果调整所述类的外环值，包括：
    根据所述调度反馈结果在对应的所述类进行计数；
    所述类内的所述NACK占所述ACK与所述NACK之和的比例少于第一设定值时，增大所述类的外环值；
    所述类内的所述NACK占所述ACK与所述NACK之和的比例大于第二设定值时，减小所述类的外环值。
  11. 根据权利要求10所述的外环初始值的调整方法，其中，所述每个所述类根据所述调度反馈结果调整所述类的外环值，还包括：
    所述类内的所述NACK占所述ACK与所述NACK之和的比例大于或等于所述第一设定值且小于或等于所述第二设定值时，保持所述类的外环值不变。
  12. 根据权利要求1所述的外环初始值的调整方法，其中，所述调整方法还包括：
    根据所有所述目标用户设备的调度反馈结果计算误块率；
    判断所述误块率是否满足预设值，若否，重新执行所述对所述多个用户设备的特征信息进行栅格化划分，并重新确定所述目标用户设备的所述外环初始值。
  13. 一种设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，其中，所述处理器执行所述计算机程序时实现如权利要求1至12中任意一项所述的外环初始值的调整方法。
  14. 一种计算机可读存储介质，存储有计算机可执行指令，其中，所述计算机可执行指令用于执行如权利要求1至12中任意一项所述的外环初始值的调整方法。
PCT/CN2022/098970 2021-06-28 2022-06-15 外环初始值的调整方法、设备及计算机可读存储介质 WO2023273891A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
US18/260,514 US20240073893A1 (en) 2021-06-28 2022-06-15 Method and apparatus for adjusting initial values of outer-loop, and computer-readable storage medium
CN202280006239.2A CN116235436A (zh) 2021-06-28 2022-06-15 外环初始值的调整方法、设备及计算机可读存储介质
EP22831710.3A EP4358444A1 (en) 2021-06-28 2022-06-15 Outer loop initial value adjustment method, device, and computer readable storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110719864 2021-06-28
CN202110719864.9 2021-06-28

Publications (1)

Publication Number Publication Date
WO2023273891A1 true WO2023273891A1 (zh) 2023-01-05

Family

ID=84692539

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/098970 WO2023273891A1 (zh) 2021-06-28 2022-06-15 外环初始值的调整方法、设备及计算机可读存储介质

Country Status (4)

Country Link
US (1) US20240073893A1 (zh)
EP (1) EP4358444A1 (zh)
CN (1) CN116235436A (zh)
WO (1) WO2023273891A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023221803A1 (zh) * 2022-05-20 2023-11-23 中兴通讯股份有限公司 外环初始值的调整方法、设备及计算机可读存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012094874A1 (zh) * 2011-01-11 2012-07-19 中兴通讯股份有限公司 根据phr进行qos调度方法及服务器
CN108702769A (zh) * 2016-01-20 2018-10-23 华为技术有限公司 资源分配的方法、基站和信道质量分级设备
CN112929980A (zh) * 2019-12-06 2021-06-08 中兴通讯股份有限公司 初始mcs值确定方法、电子设备及存储介质

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012094874A1 (zh) * 2011-01-11 2012-07-19 中兴通讯股份有限公司 根据phr进行qos调度方法及服务器
CN108702769A (zh) * 2016-01-20 2018-10-23 华为技术有限公司 资源分配的方法、基站和信道质量分级设备
CN112929980A (zh) * 2019-12-06 2021-06-08 中兴通讯股份有限公司 初始mcs值确定方法、电子设备及存储介质

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023221803A1 (zh) * 2022-05-20 2023-11-23 中兴通讯股份有限公司 外环初始值的调整方法、设备及计算机可读存储介质

Also Published As

Publication number Publication date
EP4358444A1 (en) 2024-04-24
US20240073893A1 (en) 2024-02-29
CN116235436A (zh) 2023-06-06

Similar Documents

Publication Publication Date Title
US9240859B2 (en) Method and device for correcting channel quality indicator value
EP3841776A1 (en) Method, apparatus and computer program
US8565177B2 (en) Dynamic load control for downlink signaling channels
EP2615875A1 (en) Long term evolution (lte) downlink control channel resource allocation method and base station
WO2007055619A1 (en) Selection of radio resources in a radio communications network
WO2023273891A1 (zh) 外环初始值的调整方法、设备及计算机可读存储介质
WO2017020651A1 (zh) 分配上行功率的方法、设备和系统
EP2887572B1 (en) Open-loop link adaption adjusting method and device
CN106922030B (zh) 调度的处理方法及装置
WO2023221803A1 (zh) 外环初始值的调整方法、设备及计算机可读存储介质
CN113824543A (zh) 自适应调整pdcch聚合度的方法、基站及存储介质
CN112583519B (zh) 一种链路自适应调整方法、装置、服务器及存储介质
US20180035486A1 (en) Control of Radio Connections in a Cellular Network
CN113873562B (zh) 应用于双卡双通终端的编码控制方法、装置、系统和基站
CN116155469B (zh) 基站下行控制信道资源调度优化方法、装置和计算机设备
CN105472641B (zh) 基于3D-Markov链的IEEE802.11ad网络用户QoE优化方法
WO2023066057A1 (zh) 编码调制方法、网络设备及计算机可读存储介质
CN117394963A (zh) 信道质量指示确定方法、装置及设备
WO2024067100A1 (zh) 目标误块率控制方法及装置
CN106572474B (zh) 频带共享网络的pdcch调度及功率调整的方法及装置
CN115442820B (zh) 一种小区服务优化方法、装置及电子设备
CN111526518B (zh) 一种带宽分配方法及装置
CN110858793A (zh) 一种基站的数据处理方法和装置
CN114599026B (zh) 一种下载速率的优化方法、装置、存储介质和基站
WO2022194099A1 (zh) 模型训练方法、信道调整方法、电子设备和计算机可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22831710

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 18260514

Country of ref document: US

WWE Wipo information: entry into national phase

Ref document number: 2022831710

Country of ref document: EP

Ref document number: 22831710

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2022831710

Country of ref document: EP

Effective date: 20240118