KR101017426B1 - Method for reducing channel feedback based mode selection in ofdma system - Google Patents

Abstract

PURPOSE: A mode selection based channel feedback reduction method for an orthogonal frequency division multiple access system is provided to select a feedback mode according to the channel state of a user terminal. CONSTITUTION: An average total throughput of an SF(Selective Feedback) method and a BF(Bitmap Feedback) method is calculated through statistical channel properties of user terminals. One feedback method is selected as an RFM(Recommended Feedback Mode) by using the comparison of the calculated total yield. The selected RFM is broadcasted to a user terminal of the cell.

Description

Mode selection reducing channel feedback reduction method for orthogonal frequency division multiple access system {Method for reducing channel feedback based mode selection in OFDMA system}

The present invention relates to a mode selection based channel feedback reduction method for an orthogonal frequency division multiple access system.

More specifically, in a downlink communication situation in an orthogonal frequency division multiple access system, the BS performs a selective feedback (SF) technique and a bit-map based feedback (BF) based on a user's channel state. The present invention relates to a method of reducing channel feedback by selecting one feedback technique and allowing opportunistic scheduling to be performed through the selected feedback technique. In addition, in a downlink communication situation in an orthogonal frequency division multiple access system, each user selects one of the selective feedback (SF) technique and the bitmap feedback (BF) technique based on the channel condition and the network recommendation. The present invention relates to a method of reducing channel feedback by allowing opportunistic scheduling to be performed through a feedback technique.

The present invention is derived from the research conducted as part of the IT source technology development project of the Ministry of Knowledge Economy and the Ministry of Information and Communication Research and Development. [Task management number: 2008-F-007-02, Task name: Intelligent wireless communication in a three-dimensional environment system].

Recently, to improve the performance of wireless communication systems, opportunistic scheduling has been proposed to allocate resources to users with the best channel conditions at any given moment (R. Knopp and PA Humblet, “Information capacity and power control in single-cell multiuser). communications, ”in Proc. IEEE ICC, 1995.).

OFDMA is a technique of increasing the spectral efficiency by dividing the overall bandwidth into multiple subchannels orthogonal to each other. In order to use the opportunistic scheduling technique in the OFDMA system, each user needs to feed back his channel information to the base station for all subchannels. As a result, the required feedback amount increases in proportion to the number of subchannels. Therefore, many attempts have been made to reduce the amount of feedback in Orthogonal Frequency Division Multiple Access (OFDMA) systems.

Techniques for reducing the amount of feedback in opportunistic scheduling are largely classified into opportunistic feedback (OF) technique, selective feedback (SF) technique, and bit-map based feedback (BF) technique. Can be.

The opportunistic feedback technique opportunistically feeds back the MCS grade only for subchannels with a higher modulation and coding scheme (MCS) grade than a predetermined threshold (S. Patil and G. Veciana, "Reducing feedback for opportunistic scheduling in wireless systems). , "IEEE Trans. Wireless Commun., Vol. 6, pp. 4227-4231, Dec. 2007.), (T. Tang, RW Heath Jr., S. Cho, and S. Yun," Opportunistic feedback in multiuser MIMO systems with linear receivers, "IEEE Trans. Commun., vol. 55, pp. 1020-1032, May 2007.).

The selective feedback scheme feeds back a predetermined number of MCS grades of subchannels having good channel status among all subchannels (P. Svedman, SK Wilson, LJ Cimini, and B. Ottersten, "A simplified opportunistic feedback and scheduling scheme for OFDM , "in Proc. IEEE VTC spring, 2004.), (Z. Han and Y. Lee," Opportunistic scheduling with partial channel information in OFDMA / FDD systems, "in Proc. IEEE VTC fall, 2004.)

The bitmap feedback technique feeds back 1 bit whether the MCS class of each subchannel is larger or smaller than the threshold value based on a predetermined MCS threshold (S. Sanayei, A. Norstatina, and N. Alhdahir, "Opportunistic dynamic subchannel allocation in multiuser OFDM networks with limited feedback, "in Proc. Information Theory Workshop, 2004.), (C. Jieying, RA Berry, and ML Honig," Large system performance of downlink OFDMA with limited feedback, "in Proc. IEEE ISIT, 2006 .).

Among the opportunistic feedback schemes, since the bandwidth of the feedback channel changes with time, the base station must determine the bandwidth allocated to the feedback channel among the entire bandwidths in consideration of the worst case. Thus, the efficiency of using the entire bandwidth is also reduced. However, although some of the opportunistic feedback techniques provide contention-based feedback information transmission to fix the amount of feedback, the performance of these techniques is limited due to the inherent inefficiency of the contention-based approach. Have

On the other hand, the selective feedback method and the bitmap feedback method do not cause the above problem because the fixed bandwidth feedback channel is allocated for each user.

When trying to reduce the amount of feedback on the channel information, a trade-off relationship is established between the quantity of feedback and the quality of feedback. In this case, the amount of feedback is a viewpoint that focuses on the total number of subchannels to feed back, and the quality of feedback is a view that values accurate feedback on the actual MCS grade of the subchannels.

Selective feedback techniques focus on the quality of feedback rather than the amount of feedback. In contrast, the bitmap feedback technique places more emphasis on the amount of feedback than on the quality of feedback. However, improving overall system performance requires a good balance between the two points of view, depending on the channel state of the user at every moment.

The present invention reduces the channel feedback by selecting an appropriate feedback scheme among the selective feedback (SF) technique and the bitmap (BF) feedback technique based on the channel state of the users, and enabling the opportunistic scheduling to be performed through the selected feedback technique. It aims to provide a new way of reducing feedback that effectively trades off the amount of feedback and the quality of feedback .

The channel feedback reduction method according to the present invention is divided into a network mode selection feedback (NMF) method and a user mode selection feedback (UMF) method according to a subject for determining a feedback mode. In the case of the NMF method, the base station selects the most suitable feedback scheme by predicting and comparing the time-averaged overall yields of the SF mode and the BF mode based on the statistical channel characteristics of the given users. Choose the most appropriate feedback technique based on the status and network recommendations .

Mode selection based channel feedback reduction method according to an embodiment of the present invention relates to a mode selection based channel feedback reduction method in a downlink communication situation of an orthogonal frequency division multiple access system, the base station, a) statistical channel characteristics of the user terminal; Calculating time average overall yield of the selective feedback (SF) technique and the bitmap feedback (BF) technique; b) comparing the time average overall yield of the calculated selective feedback (SF) scheme with the time average overall yield of a bitmap feedback (BF) scheme and selecting either one of the selective feedback (SF) technique and the bitmap feedback (BF) technique. Selecting a feedback technique of the network recommended mode (RFM); And c) broadcasting the selected network recommended mode (RFM) to user terminals of the cell.

In particular, the statistical channel characteristic is characterized in that the cumulative mass function (CMF) for the actual MCS class of the user terminals of the cell.

The bitmap feedback (BF) scheme may further include: selecting, by the base station, a rank threshold n, and broadcasting the selected rank threshold n to the user terminals; And receiving bitmaps for MCS thresholds and subchannels from the user terminals.

The MCS threshold may be an actual MCS grade of an upper n-th subchannel among all subchannels based on the rank threshold n.

The base station may repeat steps a) to c) when a new user terminal enters the cell or the user terminal belonging to the cell exits the cell.

In addition, the base station estimates a cumulative mass function (CMF) of the actual MCS class of the user terminals using statistical reciprocal between the uplink communication channel and the downlink communication channel.

In addition, the base station receives feedback of the actual MCS grade for one predetermined subchannel from user terminals and estimates a cumulative mass function (CMF) of the actual MCS grade of the user terminals based on the received information. It features.

Meanwhile, a mode selection based channel feedback reduction method according to another embodiment of the present invention relates to a mode selection based channel feedback reduction method in a downlink communication situation of an orthogonal frequency division multiple access system. Receiving (RFM); Determining whether a change to another selectable feedback scheme in the network recommended mode (RFM) satisfies a set allowance condition; Calculating an expected data rate of each of the selectable feedback techniques when the acceptance condition is satisfied; Selecting a feedback scheme having the highest expected rate among the selectable feedback schemes; And feeding back feedback information to the base station through the selected feedback scheme.

In particular, the expected rate is characterized in that the expected value for the data rate of the user terminal according to a specific feedback scheme.

In addition, the network recommended mode (RFM) is characterized in that the feedback scheme of any one of a selective feedback (SF) technique, and bitmap feedback (BF) techniques according to the rank threshold determined by the base station. do.

In addition, when the network recommended mode (RFM) is any one of the bitmap feedback (BF) schemes, the allowance condition is that the actual MCS class of the entire subchannels is equal to the MCS threshold value according to the rank threshold value; If the number of channels is smaller than the allowable threshold A, it is possible to change to the other selectable feedback techniques, and if the number of channels is larger than the allowable threshold A, it is impossible to change to the other selectable feedback techniques. It features.

In addition, when the network recommended mode (RFM) is the selective feedback (SF) scheme, the allowance condition is that when the actual MCS class of the subchannel having the best characteristics among all subchannels is smaller than the allowable threshold (B). It is possible to change to other selectable feedback techniques, and if it is larger than the allowable threshold B, it is impossible to change to the other selectable feedback techniques.

According to the present invention has the following effects.

Compared with the existing feedback reduction methods, the proposed feedback reduction methods show good performance with similar feedback amount.

The network mode selection feedback (NMF) method achieves an efficient tradeoff between the amount and quality of feedback and obtains significant performance improvement over existing feedback reduction methods with similar feedback amounts. In addition, the user mode selection feedback (UMF) method considers statistical channel characteristics of all user terminals and instantaneous channel characteristics of a specific user terminal at the same time, and is a more effective trade between the quantity and quality of feedback than the network mode selection feedback (NMF) method. Enable off.

That is, the feedback reduction methods according to the present invention select the most desirable feedback scheme according to the channel state of the user terminals, so that an efficient tradeoff is made between the quantity and the quality of the feedback.

The present invention will now be described in detail with reference to the accompanying drawings. The same reference numerals are used for the same components, and repeated descriptions and detailed descriptions of well-known functions and configurations that may unnecessarily obscure the subject matter of the present invention will be omitted. Embodiments of the present invention are provided to more completely explain the present invention to those skilled in the art. Accordingly, the shape and size of elements in the drawings may be exaggerated for clarity.

First, in the present invention, a network mode selection-based feedback (NMF) method and user for reducing feedback required for opportunistic scheduling in a downlink communication situation of an orthogonal frequency division multiple access (OFDMA) system We present a user mode selection-based feedback (UMF) method.

The method proposed in the present invention is based on the channel state of the user terminal based on one of the feedback (selective feedback (SF) technique (mode) and bit-map based feedback (BF) technique (mode) It is designed to choose.

The selective feedback technique (hereinafter referred to as 'SF technique') accurately feeds back some of the entire channels in real MCS (modulation and coding scheme) class, and the bitmap feedback technique (hereinafter referred to as 'BF technique') The presence or absence of feedback on the channel is expressed using 1 bit, and the MCS level of the channels to be fed back is expressed as one representative value.

In the network mode selection feedback method (hereinafter, referred to as 'NMF method'), a base station (BS) predicts and compares the average total throughput of two feedback techniques based on statistical channel characteristics of channels of user terminals. Choose a feedback technique with superior value.

In the user mode selection feedback method (hereinafter, 'UMF method'), each user terminal determines the most suitable feedback method at every moment based on its channel realization and network recommendations.

According to the present invention, it is confirmed that the proposed methods show superior performance with similar feedback amounts when compared with the conventional feedback reduction methods.

In the present invention, an OFDMA system including K user terminals and M subchannels is considered in a downlink situation of a cellular system. In the present invention, the OFDMA system basically has a base station managing one cell, and one or more user terminals communicate with the base station of the cell.

The statistical channel characteristics of the M subchannels are independent and identically distributed (i.i.d). The state of the subchannels does not change within each time slot, but changes as the time slot changes.

으로 나타내고, 여기서 L은 MCS 등급의 개수를 나타낸다.The user terminal measures signal-to-noise-ratio (SNR) of subchannels used for downlink communication for each time slot, and determines the MCS class of each subchannel. The MCS rating for subchannel m of user terminal k is

Where L represents the number of MCS grades.

의 구간에 속한다면 이에 해당하는 MCS 등급과 데이터 전송률은 각각

,

이다. 여기서

와 같다. 도 1의 데이터 전송률은

로 정의된다.1 shows the relationship between the SNR threshold, the MCS class and the data rate for the case of L = 16. If the received SNR is

If it belongs to, the corresponding MCS class and data rate are respectively

,

to be. here

Same as The data rate of Figure 1

Is defined as

). The user terminals belonging to one cell determine the MCS class of the subchannels and then transmit the information to the base station through the error-free feedback dedicated channel. In this case, when the user terminals transmit the MCS class of the subchannels to the base station by using a feedback reduction scheme of the SF method or the BF method, the information on the subchannel obtained through the feedback from the user terminals may be different from the actual MCS class. have. For example, consider the case where the SF technique is used and the data rate of the lowest MCS level is 0 (ie,

).

과 무관하게 사용자 단말 k의 부채널 m에 해당하는 MCS 등급이 1인 것으로 간주한다. 이하에서는 설명의 편의를 위해

인 경우만 고려하도록 하겠다. 하지만 본 발명에서의 모든 논의는

의 경우에도 적용될 수 있다. 이 경우, 데이터 전송률

의 가상 MCS 등급 0이 MCS 등급 집합에 추가되고, 기지국은 피드백 정보가 전송되지 않은 부채널의 MCS 등급을 0으로 간주한다.In terms of scheduling, the case of not feeding back information on a subchannel and the case of feeding back a subchannel having an MCS class of 1 are the same. Therefore, if the user terminal k does not feed back information on the subchannel m, the base station is the actual MCS class

Regardless, the MCS level corresponding to the subchannel m of the user terminal k is assumed to be 1. Hereinafter, for convenience of explanation

Will be considered only. However, all the discussion in the present invention

This may also apply. In this case, the data rate

The virtual MCS class 0 of is added to the MCS class set, and the base station considers the MCS class of the subchannel where feedback information is not transmitted to be zero.

으로 나타낸다. 기지국은 피드백 된 정보 (즉,

)를 기반으로 다음 시간 슬롯에서 서비스 받을 사용자 단말을 결정하고, 그 사용자 단말에게 데이터를 전송한다.The feedback reduction method is used to determine the MCS rating for the subchannel m fed back by the user terminal k.

Represented by The base station is fed back information (i.e.

) Determines a user terminal to be serviced in the next time slot, and transmits data to the user terminal.

In order to achieve the above, the user terminal in the present invention, SNR measuring unit (not shown) for measuring the SNR of the sub-channels used for downlink communication for each time slot, the SNR of the sub-channels measured through the SNR measuring unit A grade determining unit (not shown) for determining MCS grades of respective subchannels based on each other, and a communication unit (not shown) for transmitting the MCS grade of each subchannel determined through the class determining unit to a base station or receiving data transmitted from the base station. It is provided.

On the other hand, the base station in the present invention determines the user terminal to be serviced in the next time slot based on the communication unit (not shown) for receiving information fed back from the user terminals, and the information fed back from the user terminals to the determined user terminal A scheduler (not shown) for processing data to be transmitted is provided.

의 누적 질량 함수 (cumulative mass function;CMF)를 이용하여 생성한다.

의 누적 질량 함수(CMF)는 수학식 1과 같이 표현된다. The scheduler allocates a resource to a user terminal having the best channel value relative to its statistical characteristics using a cumulative mass function-based scheduling (CS) algorithm. To this end, the scheduler determines the scheduling metric of user terminal k.

It is generated using the cumulative mass function (CMF) of.

The cumulative mass function (CMF) of is expressed by Equation (1).

은 동일하다. 따라서 이하에서는 채널 인덱스(index)를 생략하고,

으로 나타내도록 한다. Each subchannel is independent and identical (iid) to each subchannel.

Is the same. Therefore, the channel index is omitted below.

To indicate

은 시스템에서 사용되는 피드백 절감 기법의 종류와 사용자 단말 k의 실제 MCS 등급에 대한 누적 질량 함수(이하, 'CMF')인

값에 의존한다. 후술될 실시예 1의 A에서는 기지국이 모든 사용자 단말들의 실제 MCS 등급에 대한 CMF (즉,

)를 알고 있다는 가정하에

을 계산하는 방법에 대해서 살펴보고, 실시예 1의 B에서는 기지국이 사용자 단말들의 실제 MCS 등급에 대한 CMF를 사전에 알지 못하는 경우에 대하여 살펴보도록 하겠다.

Is the type of feedback reduction technique used in the system and the cumulative mass function ('CMF') for the actual MCS rating of the user terminal k.

Depends on the value. In A of embodiment 1 to be described later, the base station is a CMF for the actual MCS class of all user terminals (that is,

Assuming you know

Looking at the method of calculating the, will be described in the case of B of the first embodiment when the base station does not know the CMF for the actual MCS class of the user equipment in advance.

Next, a procedure of scheduling by a scheduler of a base station using a CS algorithm in an orthogonal frequency division multiple access system using a feedback reduction scheme will be described.

으로 기지국에 피드백한다.Iii) user terminal k is the MCS class for subchannel m.

To the base station.

으로 정의한다.Ii) the scheduler An equal probability variable is generated between the intervals of and the generated equal probability variable is used as a scheduling metric. For computational convenience

It is defined as

에게 부채널 m을 할당한다.Iii) the scheduler compares the equal probability variable values generated in step ii) and the user terminal with the largest value.

Assigns subchannel m to.

에게 부채널 m을 통하여

의 데이터 전송률로 데이터를 전송한다.Iii) the base station is a user terminal

Through subchannel m to

Transfer data at data rate.

으로 피드백 할 경우, 사용자 단말 k가 스케줄러에 의해 선택될 확률은 수학식 2로 계산된다.By the above algorithm, the user terminal k

In case of feedback, the probability that the user terminal k is selected by the scheduler is calculated by Equation 2.

와 시스템의 시간 평균적 전체 수율

는 수학식 3과 같이 표기할 수 있다.And average throughput of user terminal k.

Average yield of systems with

May be written as in Equation 3.

The time fraction of the user terminals is 1 / K, so temporal fairness is satisfied.

In the following description, an NMF method and a UMF method for reducing feedback required for opportunistic scheduling in a downlink communication situation of an OFDMA system will be described according to embodiments.

Example 1-NMF Method

The NMF method is a method in which a base station (BS) selects a feedback technique having a superior average total throughput value among SF and BF techniques based on statistical channel characteristics of channels of user terminals.

)을 이용하여 두 가지 피드백 기법의 시간 평균적 전체 수율을 예측하고, 성능이 더 우월한 피드백 기법을 선택한다. The base station is responsible for a given statistical channel characteristic of the user

) To predict the time-averaged overall yield of the two feedback techniques, and to select a feedback technique with better performance.

First, the SF technique will be described. In the SF scheme, each user terminal feeds back a predetermined number of information of subchannels having a good channel state among all subchannels to a base station. This is the same as the existing method.

에 속하는 부채널들의 인덱스(식별정보)와 실제 MCS 등급 값을 피드백 한다. 여기서 집합

는 전체 부채널 중 상위 W개의 부채널들로 이루어진 집합이며, 복수 개의 부채널들의 MCS 등급이 동일한 경우, 임의 선택(random selection)을 한다.Specifically, user terminal k is a set

It feeds back the index (identification information) of the subchannels belonging to and the actual MCS class value. Where set

Is a set consisting of upper W subchannels among all subchannels, and when a plurality of subchannels have the same MCS grade, random selection is performed.

은 수학식 4와 같이 표기할 수 있다.Therefore, when using the SF technique, the MCS class for the sub-channel m fed back by the user terminal k

May be written as in Equation 4.

2 (a) shows a feedback frame structure in the SF technique.

bits이고, 여기서

와

은 각각 MCS 등급과 부채널의 인덱스를 표현하는데 필요한 bit 개수를 나타낸다. 그리고

은 A 보다 크거나 같은 최소의 정수를 의미한다.The total length of the feedback frame

bits, where

Wow

Denotes the number of bits required to represent the MCS class and subchannel index, respectively. And

Is the smallest integer greater than or equal to A.

Next, we will look at the BF technique.

In the BF scheme proposed in the present invention, unlike the conventional BF scheme, each user terminal determines its MCS threshold value at every time slot.

을 선정하고, 그 값이 해당 셀의 사용자 단말들에게 브로드캐스트되도록 한다. 사용자 단말 k는 기지국으로부터 수신한 순위 임계값을 기반으로 매 시간 슬롯 마다 전체 부채널 중 상위 n번째 부채널의 실제 MCS 등급을 MCS 임계값

으로 지정한다. The scheduler of the base station has a rank threshold

Is selected, and its value is broadcasted to user terminals of the cell. User terminal k, based on the rank threshold received from the base station, the MCS threshold of the actual MCS class of the upper n-th subchannel of the total subchannels in every time slot

To be specified.

에 속하는 부채널의 bit는 1, 그 외의 경우는 0으로 표현한다. 여기서 집합

는 전체 부채널 중 실제 MCS 등급이

보다 크거나 같은 부채널들의 집합을 의미하며,

으로 표현된다. 최종적으로 사용자 단말 k는 MCS 임계값과 생성된 비트맵을 기지국으로 전송한다. After this process, the user terminal k generates bitmaps for M subchannels. At this time,

The subchannel bits belonging to are represented by 1, otherwise, 0. Where set

Is the actual MCS rating of all subchannels

Means a collection of sub-channels greater than or equal to,

. Finally, user terminal k transmits the MCS threshold and the generated bitmap to the base station.

은 수학식 5와 같이 표현된다.In case of using the BF technique of the rank threshold n, the MCS class for the subchannel m fed back by the user terminal k

Is expressed as in Equation 5.

bits이다.2 (b) shows the structure of a feedback frame of the BF scheme, the length of which is

bits.

으로 표기한다. BF 기법의 순위 임계값 n은 1에서 M까지의 값을 가질 수 있으므로, 총 M개의 BF 기법들이 존재한다. 따라서 순위 임계값 n의 BF 기법은

으로 표기한다. 기지국은 SF 기법과 M개의 BF 기법들로 구성된 (M+1)개의 후보 중에서 가장 적합한 피드백 기법을 선택한다. 이때 기지국이 정한 피드백 기법을 '네트워크 권장 모드(recommended feedback mode;RFM)'라고 칭하기로 한 다. SF technique

Mark as. Since the rank threshold n of the BF scheme may have a value from 1 to M, there are a total of M BF schemes. Therefore, the BF technique of rank threshold n is

Mark as. The base station selects the most suitable feedback scheme among (M + 1) candidates consisting of the SF scheme and the M BF schemes. In this case, the feedback scheme determined by the base station will be referred to as a 'recommended feedback mode (RFM)'.

3 shows a procedure for selecting a feedback technique of a base station in an NMF method.

는 피드백 기법의 시간 평균적 전체 수율을 의미하고,

는 기지국에 의해 결정된 순위 임계값을 표현한다. 기지국은 결정된 RFM을 사용자 단말들에게 브로드캐스트하며, 해당 셀에 새로운 사용자 단말이 들어오거나, 해당 셀에 속해 있던 사용자 단말이 해당 셀을 빠져나갈 때마다 위의 피드백 기법 선택 절차를 반복한다.here

Denotes the time-averaged overall yield of the feedback technique,

Represents the rank threshold determined by the base station. The base station broadcasts the determined RFM to user terminals, and repeats the above feedback technique selection procedure whenever a new user terminal enters the cell or a user terminal belonging to the cell exits the cell.

와

에 해당하는 피드백 프레임의 길이가 각각

bits,

bits이므로, 시간 슬롯 당 사용자 단말별 피드백 양은

bits이다.The structure of the feedback frame is determined by the broadcast network recommended mode (hereinafter, referred to as 'RFM'), and user terminals transmit feedback information according to a feedback frame structure corresponding to the RFM every time slot.

Wow

The length of the feedback frame corresponding to

bits,

bits, the amount of feedback per user terminal per time slot

bits.

에 대한 정보를 알고 있는 경우A. The base station of the user terminals

If you know information about

은 사용자 단말 k에 의해 측정될 수 있다. 만약 하향 통신과 상향 통신(uplink)에 사용되는 채널들이 서로 통계적 상호성(statistically reciprocal)을 가진다면, 기지국은 상향 통신을 통해서 사용자 단말들의 실제 MCS 등급에 대한 CMF를 추정할 수 있다. 따라서, 본 절에서는 기지 국이 모든 사용자 단말들의

에 대한 정보를 안다고 가정하기로 한다. 한편 '통계적 상호성'은 상호성(reciprocity)과 의미상 차이가 있다. 즉, 상향 통신 채널과 하향 통신 채널의 통계적 특성은 동일하지만 각 채널의 매 순간 값은 다를 수 있음을 의미한다. 예를 들어, 시분할 이중화(time division duplex;TDD) 시스템의 경우, 하향 링크와 상향 링크 구간으로 구성된 프레임의 길이(duration)가 가간섭성 시간(coherence time) 보다 길면 통계적 상호성은 성립하나 상호성이 성립되지 않을 수 있다. In a downlink communication situation

May be measured by user terminal k. If the channels used for the downlink and uplink have statistically reciprocal relations with each other, the base station may estimate the CMF for the actual MCS class of the user terminals through the uplink. Therefore, in this section, the base station is

Suppose you know information about. On the other hand, 'statistical reciprocity' differs in meaning from reciprocity. That is, the statistical characteristics of the uplink communication channel and the downlink communication channel are the same, but each instantaneous value of each channel may be different. For example, in a time division duplex (TDD) system, if the duration of a frame consisting of downlink and uplink intervals is longer than the coherence time, statistical interactivity is established but mutuality is established. It may not be.

의 계산 과정을 살펴보면 다음과 같다. 우선, 이를 위해서

의 CMF,

을 구해야 한다.

의 경우,

The calculation process of is as follows. First of all,

CMF,

Should be obtained.

In the case of,

를 의미하고,

은 집합

의 크기 (cardinality)를 나타낸다.Equal to, where

Means,

Silver set

Indicates the size of the cardinality.

에 대한 수식은 수학식 7과 같이 표현된다.The last equal sign of Equation 6 holds because all subchannels have the same distribution characteristic.

The equation for is expressed by Equation 7.

은 사용자 단말 k의 전체 부채널 중 실제 MCS 등급이

보다 크거나 같은 부채널들의 개수이고,

은 0으로 정의한다. 또한

이다. 피드백 기법이

일 때, 기지국은

을 이용하여 사용자 k의 스케줄링 메트릭을 생성하며,

의 관계식을 가진다. 따라서 수학식 3의

대신에

을 대입하면

이 구해진다. here

Is the actual MCS rating of all subchannels of

The number of subchannels greater than or equal to

Is defined as 0. Also

to be. Feedback techniques

When the base station is

To generate a scheduling metric for user k,

Has a relation of. Therefore, the equation

Instead of

If you substitute

Is obtained.

의 값을 구하는 과정을 살펴기로 한다. 모든

에 대하여

의 CMF,

은 수학식 8로 계산된다.to the next

Let's look at the process of finding the value of. all

about

CMF,

Is calculated by equation (8).

The third equal sign is true because all subchannels have the same distribution characteristics.

은 아래 수학식 9와 같다.In equation (8)

Is shown in Equation 9 below.

은 M개의 부채널 중 상위 w번째 부채널을 의미한다. here

Denotes the upper w-th subchannel of the M subchannels.

In Equation 9, the third and last equal signs are established by the same distribution characteristic of subchannels and the order statics theory.

을 구할 수 있다.

의 피드백 기법을 사용할 경우, 기지국은

을 이용하여 사용자 단말 k에 대한 스케줄링 메트릭을 생성한다.

은 수학식 3의

대신에

을 대입함으로써 계산된다. Substituting Equation 9 into Equation 8

Can be obtained.

Using the feedback technique of the

Generates a scheduling metric for the user terminal k using.

Of equation (3)

Instead of

Calculated by substituting

에 대한 정보를 알지 못하는 경우B. The base station of the user terminals

If you don't know information about

에 대한 정보를 기지국이 미리 알고 있다고 가정하였다.In A of Embodiment 1, the downlink communication channel and the uplink communication channel have statistical mutual relations,

It is assumed that the base station knows about.

However, hereinafter, the base station does not know the information about the CMF of the user terminal in advance will be described.

The procedure of the extended NMF technique (for B of Example 1) is similar to that discussed in A of Example 1 except for the following four items.

이다. 1) Each user terminal always feeds back the actual MCS class of one predetermined subchannel, for example, subchannel 1, and feeds back the remaining (M-1) subchannels in the same manner as described in A of Embodiment 1 Send the information. Therefore, the amount of feedback per user terminal per time slot

to be.

의 정보를 기반으로

을 아래 수학식 10과 같이 추정한다.2) the base station is fed back

Based on information from

Is estimated as in Equation 10 below.

＜1은 망각 인자(forgetting factor)를 의미하고,

은 A의 조건이 만족되면 1, 만족되지 않으면 0의 값을 가지는 지시함수이다. Where 0 <

<1 means forgetting factor,

Is an indication function that has a value of 1 if the condition of A is satisfied and 0 if not satisfied.

와

은 서로 동일하다.Because each user terminal always feeds back the actual MCS rating for subchannel 1

Wow

Are the same as each other.

은 대수의 법칙(law of large numbers)에 의해서 실제 확률 값으로 수렴하게 된다. 또한 각 부채널들이 동일한 분포 특성을 가지므로

로부터

의 값을 유도해 낼 수 있다. Thus estimated

Is converged to the actual probability value by the law of large numbers. In addition, since each subchannel has the same distribution characteristic,

from

We can derive the value of.

을, 나머지 (Ｍ-1)개의 부채널에 대해서는

을 이용한다. 추정된

을 이용하여

을 계산하는 과정은 실시예 1의 A에 기재된 방식과 유사하다.3) When the base station generates the scheduling metric for the user terminal k, for subchannel 1

For the remaining (M-1) subchannels

Use Estimated

Using

The process of calculating is similar to the method described in A of Example 1.

4) The base station selects RFM by substituting M instead of (M-1) in the procedure of FIG. The selected RFM is broadcast to user terminals, and the base station periodically updates the RFM.

Example 2-UMF Method

The UMF method is a method in which each user terminal selects the most suitable feedback technique at every moment based on its channel realization and network recommendations.

In the NMF method of Embodiment 1 described above, the base station selects an RFM based on statistical characteristics of channels of user terminals. At this time, in determining the feedback scheme, the base station cannot use information on channel conditions of the user terminals.

On the other hand, the user terminal can know information about all its subchannels in every time slot. Therefore, if each user terminal determines its feedback scheme based on the instantaneous channel state and RFM, the performance of the entire system will be improved. This is because a more suitable feedback scheme can be selected by considering both the instantaneous channel state of a specific user terminal and the statistical channel characteristics of all user terminals.

In consideration of this, the UMF technique selects its feedback scheme according to the following two rules.

으로 제한된다.1) Each user terminal basically follows the RFM broadcast by the base station, and the change from RFM to another feedback scheme is made only when the allowance condition is satisfied. A description of the permissible conditions will be given later, and the rate of change in the feedback technique is

Limited to.

2) When the change of the feedback scheme is allowed, each user terminal selects a feedback scheme having the highest expected data rate. Here, the expected data rate refers to an expected value for data rates of user terminals when following a specific feedback scheme.

The procedure of UMF method is as follows.

과

는 각각 사용자 단말 k가 선택한 피드백 기법과 특정 피드백 기법을 이용할 경우의 사용자 단말 k의 기대 전송률을 의미한다.The base station determines the RFM in the same manner as in the NMF method and broadcasts the selected feedback scheme. At this time, the total number of user terminals K is broadcast together. Each user terminal selects the most suitable feedback scheme based on its channel state and RFM at every time slot. 4 shows a selection procedure performed by user terminal k.

and

Denotes the expected transmission rate of the user terminal k when the feedback technique selected by the user terminal k and the specific feedback technique are used.

에 의해 결정되고, 각 사용자 단말들은 추가적인 1 bit 전송을 통해서 자신의 피드백 프레임 종류를 기지국으로 알려준다. 시간 슬롯 당 사용자 단말별 피드백 양은

bits이며, 이 값은 NMF 방법의 경우와 1 bit 차이가 난다.Feedback frame structure of user terminal k is

Determined by, each user terminal informs the base station of its feedback frame type through an additional 1 bit transmission. The amount of feedback per user terminal per time slot is

bits, which is one bit different from that of the NMF method.

와

일 때의 피드백 기법 변경에 대한 허용 조건과

,

의 계산 방법을 살펴보기로 한다.In A and B of Embodiment 2 to be described below, RFMs are respectively

Wow

Conditions for the change of the feedback technique when

,

Let's look at how to calculate.

인 경우A. The network recommended mode

If

의 값이 임계값

보다 작거나, 같거나 혹은 큰 경우에 대하여 각각 1,

, 0의 확률로 자신의 피드백 기법을 변경한다. 여기서

으로 정의되고, 전체 부채널 중 실제 MCS 등급이

인 부채널의 개수를 나타낸다. 여기서, 임계값

는 본 발명의 특허청구범위에 기재된 허용 임계치(A)에 해당한다.User terminal k is

Is the threshold

1 for each less than, equal to or greater than 1,

Change your feedback technique with a probability of 0. here

Is defined as the actual MCS rating of all subchannels

It indicates the number of sub-channels. Where threshold

Corresponds to the tolerance threshold (A) described in the claims of the present invention.

의 값이 작을 때, 사용자 단말 k가 자신의 피드백 기법을

에서 다른 피드백 기법으로 변경한다면 피드백 양에 대한 최소의 감소로 피드백의 질을 크게 증가 시킬 수 있거나, 질에 대한 최소의 감소로 양을 크게 증가 시킬 수 있다.This acceptance condition is based on the following evidence:

When the value of is small, the user terminal k uses its feedback scheme.

If you change from one feedback method to another, the quality of the feedback can be greatly increased by the minimum reduction in the amount of feedback, or the quantity can be greatly increased by the minimum reduction in the quality of the feedback.

와

의 값을 계산하기 위해서, 우선

에 대한

을 살펴보면,

Wow

To calculate the value of, first

For

Looking at it,

와

은

의 관계식을 만족시키는 값으로 정한다. 따라서 피드백 기법 변경의 허용 비율은

값으로 제한된다.Equation 11 Using Equation 11

Wow

silver

Set the value to satisfy the relation of. Therefore, the acceptable rate of change in the feedback technique is

Limited by value.

을 이용하여 사용자 단말 k를 위한 스케줄링 메트릭을 만든다. 이것은

의 값을 충분히 작게 설정한다면, 사용자 단말 k가 피드백 하는 MCS 등급의 CMF가

와 유사하기 때문이다. 따라서

은 The base station

Create a scheduling metric for the user terminal k using. this is

If the value of is set small enough, the CMF of the MCS grade fed back by the user terminal k is

Because it is similar to therefore

silver

대신에

을 대입함으로써 얻어진다. 수학식 12에서

의 표기는 기지국이 사용자 단말 k의 CMF를

으로 여기는 것을 강조하기 위해서이다. 이와 비슷하게

은 수학식 13과 같이 계산된다.Where the last equal sign is

Instead of

It is obtained by substituting. In equation (12)

Notation indicates that the base station uses the CMF of user terminal k.

This is to emphasize what is considered. Similarly

Is calculated as shown in Equation 13.

인 경우B. The network recommended mode

If

일 때, 피드백 기법 변경의 허용 조건을 살펴보기로 한다. 사용자 단말 k는 가장 상태가 좋은 부채널의 실제 MCS 등급 값이 임계값

보다 작거나, 같거나 혹은 큰 경우에 대하여 각각 1,

, 0의 확률로 피드백 기법을 변경한다. 여기서 임계값

은 본 발명의 특허청구범위에 기재된 허용 임계치(B)에 해당한다.RFM

In this case, the allowance conditions for the change of the feedback scheme will be described. The user terminal k has a threshold value of the actual MCS class value of the best state subchannel.

1 for each less than, equal to or greater than 1,

We change the feedback scheme with a probability of 0. Where threshold

Corresponds to the tolerance threshold (B) described in the claims of the present invention.

은 피드백의 양을 희생하여 피드백의 질을 최대화 시키는 것이 목적이다. 그러나 가장 좋 은 부채널의 실제 MCS 등급이 작은 경우, 피드백의 질을 최대화 시킴으로써 얻는 이득은 적어진다. 따라서 RFM을 계속 유지하는 것은 비효율적이다.

와

에 대한 계산을 위해,

을 살펴보면Such acceptance conditions are based on the following grounds: RFM, i.e.

The goal is to maximize the quality of the feedback at the expense of the amount of feedback. However, if the actual MCS rating of the best subchannel is small, the gain from maximizing the quality of the feedback is less. Therefore, maintaining RFM is inefficient.

Wow

For the calculation,

If you look at

Equation (14).

와

은 수학식 14를 이용하여

을 만족시키는 값으로 설정된다.

Wow

Using Equation 14

It is set to a value satisfying.

를 이용하여 사용자 단말 k의 스케줄링 메트릭을 생성한다. 따라서

와

은 수학식 12 내지 수학식 13의

대신에

을 대입함으로써 계산 할 수 있다. The base station has the same reason as mentioned in A of Embodiment 2

Generates the scheduling metric of the user terminal k using. therefore

Wow

Is represented by Equations 12 to 13

Instead of

Can be calculated by substituting

5 to 9 show experimental results in an OFDMA downlink communication system environment in which the total number of subchannels is M = 46 to demonstrate the performance of the mode selection based channel feedback reduction method according to the present invention.

는 [0,10]dB 구간에서의 균등 확률 변수 생성을 통해 임의로 선정하였다. 또한 사용자 단말 k의 채널은 평균

인 레일리(Rayleigh) 분포를 따른다. Average Signal-to-Noise Ratio of User Terminal k

Is chosen randomly by generating equal random variable in [0,10] dB. Also, the channel of user terminal k is average

Follow the Rayleigh distribution.

은 수학식 15로 표현된다. Using the relationship between the SNR threshold, the MCS class and the data rate of FIG. 1,

Is expressed by equation (15).

와

중의 하나로 고정시키는 것이며, 각각 'SF-SM', 'BF-SM'으로 명칭하기로 한다. For performance comparison, we consider single mode (SM) technique, Sanayei's technique, and full feedback technique. The single mode technique uses a feedback reduction technique.

Wow

It is fixed to one of them, and will be referred to as 'SF-SM' and 'BF-SM', respectively.

이다.Sanayei's technique is a type of BF technique, and the base station broadcasts a common MCS threshold value to the user terminals that maximizes the time-averaged overall yield in a given system environment. The feedback amount per user terminal per time slot of this technique is M bits. In the case of the full feedback scheme, the user terminals feed back the MCS grade for all subchannels to the base station, and are denoted as Full. The amount of feedback per user terminal per time slot is

to be.

Several types of feedback reduction techniques were set to Ｗ = 5 to have similar feedback amounts (FIG. 5). It can be seen that the feedback amount for each user terminal of the feedback reduction techniques is significantly smaller than the feedback amount of the entire feedback technique.

는 0.2로 설정하였다. The time-averaged overall yields of various feedback reduction techniques are simulated for 20,000 time slots. Figure 6 shows the results, and the performance of each technique is normalized to the performance of the overall feedback reduction technique. Where the UMF method

Was set to 0.2.

The experimental results showed five notable results.

1) The performance of BF-SM and Sanayei's technique, which focuses on quantity rather than quality of feedback, is superior to SF-SM in the area of few user terminals. The performance difference decreases as the number of user terminals increases. In the area where the number of user terminals increases, the performance of SF-SM, which focuses on the quality of feedback, is superior to BF-SM and Sanayei's technique. Through this, it can be seen that when the number of user terminals is small, the emphasis is placed on the amount of feedback, and when the number of user terminals is large, the emphasis should be placed on the quality of feedback.

를 RFM의 후보로 사용하는 것이 합리적임을 알 수 있다. 또한 사용자 단말들의 통계적 채널 특성이 서로 다른(heterogeneous) 경우, 모든 사용자 단말들에게 동일한 MCS 임계값을 적용하는 것은 비효율적임을 알 수 있다.2) BF-SM outperforms Sanayei's scheme for any number of user terminals. From this NMF technique

It can be seen that it is rational to use as a candidate for RFM. In addition, when the statistical channel characteristics of the user terminals are different (heterogeneous), it can be seen that applying the same MCS threshold to all the user terminals is inefficient.

와

에 대한 예측한 값이 모의 실험 결과값 완전히 일치한다.3) in section 3-A

Wow

The predicted values for match the simulation results completely.

와

의 값을 서로 비교하여 우월한 값을 가지는 피드백 기법을 RFM으로 선택하기 때문이다. 4) The performance of NMF scheme is better than that of SF-SM and BF-SM for any number of user terminals. The result is that the base station

Wow

This is because RFM selects a feedback technique with superior values by comparing the values of.

In FIG. 7, selected RFMs according to changes in the total number of user terminals are shown. This suggests that the NMF technique places more emphasis on quality than the amount of feedback as the total number of user terminals increases, and maintains an effective tradeoff between the amount and quality of feedback.

5) The performance of the UMF technique is superior to that of the NMF technique. This is because each user terminal efficiently changes a feedback scheme from RFM to another feedback scheme based on its instantaneous channel state.

One important characteristic of the CS algorithm is that temporal fairness is maintained between user terminals. Therefore, a simulation experiment was conducted to confirm whether the proposed techniques achieve temporal fairness and Jain's fairness index was used to evaluate fairness.

8 shows the result of temporal fairness according to the change of the total number of user terminals, and it can be seen that the temporal fairness of the proposed techniques is well maintained.

=0.999으로 설정하였다. 도 9는 임의로 선택된 세 명의 사용자 단말들에 대한 시간 평균적 수율 추적(trajectory) 그래프이다. 각 사용자 단말들의 시간 평균적 수율은 수학식 3의 예측 값으로 수렴함을 알 수 있다. Finally, the base station proves the convergence of the extended NMF scheme when the base station does not have information on the CMF of the user terminals in advance. The total number of user terminals is 15 and the RFM is updated every 100 time slots. Also

It was set to = 0.999. 9 is a time average yield trace graph for three randomly selected user terminals. It can be seen that the time average yield of each user terminal converges to the predicted value of Equation 3.

In the present invention, an NMF scheme and a UMF scheme are proposed to reduce the amount of feedback required for opportunistic scheduling in a downlink communication situation of an OFDMA system. The proposed schemes select the most desirable feedback scheme among various feedback schemes according to the channel state of user terminals. This makes for an efficient tradeoff between the amount and quality of feedback.

In the NMF scheme, the base station selects the most superior feedback scheme in terms of time average using statistical channel characteristics of user terminals. Simulation results show that the NMF technique places more emphasis on quality than the amount of feedback as the total number of user terminals increases. This tendency makes an efficient tradeoff between the amount and quality of feedback, and provides a significant performance improvement over existing feedback reduction techniques with similar feedback amounts.

In addition, the UMF scheme selects the most appropriate feedback scheme based on the instantaneous channel state and RFM of each user terminal. This allows for a more effective tradeoff between the amount and quality of feedback than the NMF in a manner that simultaneously considers the statistical channel characteristics of all user terminals and the instantaneous channel characteristics of a particular user terminal. Therefore, it can be seen that the performance of the UMF technique is better than that of the NMF technique.

The present invention can be embodied as computer readable codes on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media may include ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical data storage. Also included are those implemented in the form of carrier waves (eg, transmission over the Internet). The computer readable recording medium can also be distributed over network coupled systems so that the computer readable code is stored and executed in a distributed fashion.

As mentioned above, although the preferred embodiment of this invention was shown and described, this invention is not limited to the specific embodiment mentioned above, Usually, in the technical field to which this invention pertains without deviating from the summary of this invention claimed in a claim. Various modifications may be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or the prospect of the present invention.

FIG. 1 is a diagram illustrating a relationship between an SNR threshold value, an MCS class, and a data rate when the number of MCS classes is 16. FIG.

2 (a) is a diagram illustrating a feedback frame structure in the SF technique.

2B is a diagram illustrating a feedback frame structure in the BF scheme.

3 is a diagram for explaining an NMF method algorithm.

4 is a graph illustrating a selection procedure performed by user terminal k in a UMF method.

5 is a diagram illustrating the amount of feedback for each user terminal of feedback reduction techniques.

6 is a diagram illustrating normalized time average overall yield according to a change in the number of user terminals.

7 is a diagram illustrating RFM determined by the NMF method.

FIG. 8 illustrates a jain's fairness index of temporal fairness of user terminals, and illustrates temporal fairness according to the change of the total number of user terminals.

9 is a time average yield tracking graph for three randomly selected users.

Claims (12)

A method for reducing channel feedback based on mode selection in a downlink communication situation of an orthogonal frequency division multiple access system, a) calculating time average overall yield of the selective feedback (SF) technique and the bitmap feedback (BF) technique using statistical channel characteristics of the user terminals; b) comparing the time average overall yield of the calculated selective feedback (SF) scheme with the time average overall yield of a bitmap feedback (BF) scheme and selecting either one of the selective feedback (SF) technique and the bitmap feedback (BF) technique. Selecting a feedback technique of the network recommended mode (RFM); And c) broadcasting the selected network recommended mode (RFM) to user terminals of the cell. The method according to claim 1, The statistical channel characteristic is a mode selection based channel feedback reduction method, characterized in that the cumulative mass function (CMF) for the actual MCS class of the user terminals of the cell. The method according to claim 1, The bitmap feedback (BF) technique, The base station selecting a rank threshold n and broadcasting the selected rank threshold n to the user terminals; And And receiving a bitmap for MCS threshold and subchannels from the user terminals. The method of claim 3, The MCS threshold is, And a mode selection based channel feedback reduction method according to the ranking threshold n, which is an actual MCS class of the upper n-th subchannel among all subchannels. The method according to claim 1, The base station comprises: And repeating the steps a) to c) when a new user terminal enters the cell or the user terminal belonging to the cell exits the cell. The method according to claim 2, The base station comprises: And estimating a cumulative mass function (CMF) for the actual MCS class of the user terminals using statistical reciprocal between uplink and downlink channels. The method according to claim 2, The base station comprises: Based on the mode selection, characterized in that the feedback from the user terminal to the actual MCS class for a predetermined subchannel, and estimates the cumulative mass function (CMF) for the actual MCS class of the user terminal based on the feedback information How to reduce channel feedback. A method for reducing channel feedback based on mode selection in a downlink communication situation of an orthogonal frequency division multiple access system, comprising: Receiving a network recommended mode (RFM) from a base station; Determining whether a change to another selectable feedback scheme in the network recommended mode (RFM) satisfies a set allowance condition; Calculating an expected data rate of each of the selectable feedback techniques when the acceptance condition is satisfied; Selecting a feedback scheme having the highest expected rate among the selectable feedback schemes; And And feeding back feedback information to the base station through the selected feedback scheme. The method according to claim 8, The expected transfer rate is, Mode selection based channel feedback reduction method characterized in that the expected value for the data rate of the user terminal according to a specific feedback scheme. The method according to claim 8, Network Recommended Mode (RFM) is And a feedback scheme of any one of a selective feedback (SF) scheme and a bitmap feedback (BF) scheme according to a rank threshold determined by the base station. The method according to claim 10, If the network recommended mode (RFM) is any one of the bitmap feedback (BF) techniques, The above acceptance condition is, If the number of subchannels whose actual MCS class is equal to the MCS threshold according to the rank threshold among the subchannels is smaller than the allowable threshold A, it is possible to change to the selectable other feedback techniques, and the allowable threshold (A) If larger than the mode selection based channel feedback reduction method characterized in that it is impossible to change to the selectable other feedback techniques. The method according to claim 10, If the network recommended mode (RFM) is the selective feedback (SF), The above acceptance condition is, If the actual MCS class of the subchannel having the best characteristics among all the subchannels is smaller than the allowable threshold B, it is possible to change to the other selectable feedback schemes. Mode selection based channel feedback reduction method characterized in that it is impossible to change to other feedback techniques.

Images