KR102358381B1 - Method and apparatus for channel estimation in wireless communication system - Google Patents

Abstract

The present disclosure relates to a ^5th generation ( ^5G ) or pre-5G communication system for supporting a higher data rate after a 4th generation (4G) communication system such as Long Term Evolution (LTE). The present disclosure provides a method and apparatus for estimating a channel at a receiving end. The receiving apparatus includes a receiving unit and a channel estimating unit. The receiver receives the reference signal. The channel estimator determines first channel estimates for regions in which reference signals are located. The channel estimator determines second channel estimates of the remaining regions by interpolating the first channel estimates. The method of interpolation is determined based on at least one of a Doppler frequency of a channel and a channel quality.

Description

Method and apparatus for channel estimation in a wireless communication system

The present disclosure relates to channel estimation in a wireless communication system.

Efforts are being made to develop an improved ^5th generation ( ^5G ) communication system or a pre-5G communication system in order to meet the increasing demand for wireless data traffic after commercialization of the 4G (4th generation) communication system. For this reason, the 5G communication system or the pre-5G communication system is called a 4G network after (Beyond 4G Network) communication system or LTE (Long Term Evolution) after (Post LTE) system.

In order to achieve a high data rate, the 5G communication system is being considered for implementation in a very high frequency (mmWave) band (eg, such as a 60 gigabyte (60 GHz) band). In order to alleviate the path loss of radio waves and increase the propagation distance of radio waves in the ultra-high frequency band, in the 5G communication system, beamforming, massive MIMO, and Full Dimensional MIMO (FD-MIMO) are used. ), array antenna, analog beam-forming, and large scale antenna technologies are being discussed.

In addition, to improve the network of the system, in the 5G communication system, an evolved small cell, an advanced small cell, a cloud radio access network (cloud radio access network, cloud RAN), and an ultra-dense network (ultra-dense network) , Device to Device communication (D2D), wireless backhaul, moving network, cooperative communication, Coordinated Multi-Points (CoMP), and reception interference cancellation Technology development is underway.

In addition, in the 5G system, FQAM (Hybrid Frequency Shift Keying and Quadrature Amplitude M Modulation) and SWSC (Sliding Window Superposition Coding), which are advanced coding modulation (Advanced Coding Modulation, ACM) methods, and FBMC (Filter Bank Multi), an advanced access technology, Carrier), Non Orthogonal Multiple Access (NOMA), and Sparse Code Multiple Access (SCMA) are being developed.

In a wireless communication system, a receiving end needs to estimate a channel of a received signal in order to perform demodulation and decoding of the received signal. Channel estimation for a downlink signal of LTE using an orthogonal frequency division multiplexing (OFDM) scheme is performed using a cell-specific reference signal (CRS).

In LTE, resource elements (REs) through which CRS is received are located in the system bandwidth, and are arranged at intervals of 6 subcarriers in the frequency domain. Such CRS-based channel estimation is used for coherent demodulation of a physical downlink control channel (PDCCH) or a physical downlink shared channel (PDSCH). It plays an important role in synchronous demodulation of PDCCH/PDSCH.

Channel estimation is treated as an important technique because most scenarios assume that channel state information is already known in the current wireless communication system. A commonly used training signal-based channel estimation method is mainly based on linear reconstruction using a reference signal, such as least square (LS) and minimum mean square error (MMSE) methods. Although this linear restoration has optimal performance when the number of taps of the channel impulse response in a multipath channel is large, recent research results show that the channel impulse response (CIR) is poor when a very wide bandwidth is used. It has been found to have sparse properties. Based on this fact, in the case of a wireless communication system using a high-dimensional signal space, CIR has a sparse characteristic, and a compressed sensing-based channel estimation method using a nonlinear reconstruction algorithm uses a linear reconstruction algorithm such as LS. It is shown to be superior to the used channel estimation method in various performance aspects.

An embodiment provides a method and apparatus for estimating a channel using a reference signal in a wireless communication system.

A method for estimating a channel of a receiving end according to an embodiment includes a process of receiving reference signals, a process of determining first channel estimates for regions in which the reference signals are located, and performing interpolation on the first channel estimates, and determining second channel estimates of the remaining regions, wherein the interpolation method may be determined based on at least one of a Doppler frequency and a channel quality of the channel.

A receiver according to an embodiment includes a receiver and a channel estimator. The receiver receives the reference signal. The channel estimator determines first channel estimates for regions in which reference signals are located. The channel estimator determines second channel estimates of the remaining regions by interpolating the first channel estimates. The method of interpolation is determined based on at least one of a Doppler frequency of a channel and a channel quality.

According to various embodiments, performance close to the optimal estimation technique may be achieved in a channel that changes rapidly with time.

For a more complete understanding of the present disclosure, the following detailed description is made with reference to the accompanying drawings. In the drawings, like reference numbers indicate like elements.
1 schematically illustrates a wireless communication system according to an embodiment of the present disclosure.
2 illustrates a block configuration of a receiving end according to an embodiment of the present disclosure.
3 illustrates a block configuration of a communication unit of a receiving end according to an embodiment of the present disclosure.
4 illustrates reference signals of a wireless communication system according to an embodiment of the present disclosure.
5 illustrates a CRS pattern used for channel estimation according to an embodiment of the present disclosure.
6A illustrates a reference signal-based channel estimation method according to an embodiment of the present disclosure.
6B illustrates an implementation of a time-varying channel and a channel estimate by linear interpolation according to an embodiment of the present disclosure.
7 is a flowchart illustrating a channel estimation method using a reference signal according to an embodiment of the present disclosure.
8 is a flowchart illustrating a channel estimation method using a basis expansion model (BEM) according to an embodiment of the present disclosure.
9 is a flowchart in which the number of BEMs is adaptively selected according to a moving speed according to an embodiment of the present disclosure.
10 is a flowchart in which the number of BEMs is adaptively selected according to an embodiment of the present disclosure.
11 illustrates a stagewise orthogonal matching pursuit (StOMP) algorithm according to various embodiments of the present disclosure.
12 illustrates a StOMP algorithm using BEM according to various embodiments of the present disclosure.
13 is a flowchart of an operation of a receiving end performing a StOMP algorithm using BEM according to various embodiments of the present disclosure.
14 illustrates a block StOMP algorithm using BEM according to various embodiments of the present disclosure.
15 is a flowchart illustrating an operation of a receiving end performing block stOMP using BEM according to various embodiments of the present disclosure.
16A and 16B show graphs of BLER performance according to an IDFT-based channel estimation and a channel estimation technique using a StOMP algorithm.
17A and 17B show graphs of BLER performance according to a channel estimation technique using StOMP combining IDFT, StOMP algorithm, and LP BEM.
18A and 18B show graphs of BLER performance according to a channel estimation technique using block StOMP combining IDFT, StOMP algorithm, and LP BEM.
19A to 19C show reference signal patterns according to the number of transmit antennas in a wireless communication system.
20A to 20C show reference signal patterns used for each OFDM symbol area when the number of transmit antennas is 4 in a MIMO system.
21A shows the operation of the StOMP algorithm for channel estimation in a MIMO system.
21B to 21C show graphs of BLER performance according to various channel estimation techniques in a MIMO system.
22A to 22D show reference signal patterns used for each OFDM symbol area when the number of transmit antennas is 4 in a MIMO system.
23 illustrates the operation of the StOMP algorithm according to an embodiment of the present disclosure in a MIMO system.
24 is a flowchart illustrating a channel estimation according to an embodiment of the present disclosure in a MIMO system.
25 illustrates an operation of a block StOMP algorithm according to an embodiment of the present disclosure in a MIMO system.
26A to 26B show graphs of BLER performance according to various channel estimation techniques in a MIMO system.

The operating principle of the present disclosure will be described in detail with reference to the accompanying drawings below. In the following description of the present disclosure, if it is determined that a detailed description of a related well-known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted. And, the terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to intentions or customs of users and operators. Therefore, the definition should be made based on the content throughout this specification.

Hereinafter, the present disclosure describes a technique for channel estimation in a wireless communication system.

Terms that refer to a reference signal, a term that refers to a channel estimate, a term that refers to operations of an apparatus, and a term that refer to components of an apparatus used in the following description are exemplified for convenience of description. Accordingly, the present invention is not limited to the terms described below, and other terms having equivalent technical meanings may be used.

For convenience of description below, some terms and names defined in the 3rd Generation Partnership Project Long Term Evolution (3GPP LTE) standard may be used. However, the present invention is not limited by the terms and names, and may be equally applied to systems conforming to other standards.

1 schematically illustrates a wireless communication system 100 according to an embodiment of the present disclosure. Referring to FIG. 1 , a wireless communication system 100 includes a transmitter 110 and a receiver 120 . Although the transmitter 110 and the receiver 120 are depicted as separate entities in FIG. 1, both the transmitter 110 and the receiver 120 may be configured as a transmitter/receiver to perform both transmission and reception operations. Also, although the wireless communication system 100 is depicted as including only one transmitting end 110 and one receiving end 120 in FIG. 1 , the wireless communication system 100 may include a plurality of transmitting end 110 and receiving end 120 .

The transmitter 110 and the receiver 120 of FIG. 1 refer to devices capable of communicating by transmitting or receiving a signal through a wireless network. The transmitter 110 and the receiver 120 may be referred to by terms such as a mobile station, user equipment, subscriber station, remote terminal, and wireless terminal. In addition, the transmitter 110 and the receiver 120 may be referred to by terms such as a base station, an enhanced NodeB (eNB), and an access point (AP). The transmitting end 110 and the receiving end 120 may be a mobile device such as a mobile phone or a smart phone, or a stationary device such as a desktop computer.

The protocol of wireless communication of FIG. 1 includes cellular communication such as long term evolution (LTE), LTE-Advanced (LTE-A), Wireless Broadband (WiBro), and Global System for Mobile Communication (GSM). and may also include short-distance communication such as WiFi (Wireless Fidelity), Bluetooth (Bluetooth), and NFC (Near Field Communication).

The transmitter 110 and the receiver 120 of FIG. 1 may transmit or receive a signal using an orthogonal frequency division multiplexing (OFDM) scheme in the multiplexing scheme. In addition, the transmitting end 110 and the receiving end 120 using code division multiple access (CDMA), frequency division multiple access (FDMA), or time division multiple access (TDMA) using It can transmit or receive signals.

The transmitting terminal 110 and the receiving terminal 120 of FIG. 1 may include a single or a plurality of antennas, and accordingly, multiple-input multiple-output (MIMO), multiple-input single-output (multi-input single-output) , MISO), single-input multiple-output (SIMO), or single-input single-output (SISO) transmission or reception techniques may be applied. Hereinafter, the present disclosure will describe a SISO system and a MIMO system having single input and output antennas at the transmitting end 110 and the receiving end 120 for convenience of description. However, the present invention is not limited to the SISO system and the MIMO system, and may be implemented in various wireless communication systems.

The wireless channel of FIG. 1 is a path through which a signal is transmitted or received, and the channel is delayed in the time domain by scattering, reflecting, or refracting a signal by a scatterer, a reflector, etc. or may be transmitted through multiple paths. . In addition, due to the Doppler effect by the movement of the transmitter 110 or the receiver 120, a Doppler shift phenomenon, that is, the frequency of the signal received at the receiver 120 may be different from the frequency when transmitted from the transmitter 110 have. A channel in which the above-described channel change occurs rapidly over time may be referred to as a fast fading channel.

Various embodiments of the present disclosure provide a method for adaptively selecting a channel estimation method for a channel that is rapidly changing with time or is stopped as described above.

2 illustrates a block configuration of a receiving terminal 120 according to various embodiments of the present disclosure. Terms such as '~ unit' and '~ group' used below mean a unit that processes at least one function or operation, which may be implemented as hardware or software, or a combination of hardware and software.

Referring to FIG. 2 , the receiving terminal 120 includes a communication unit 210 , a control unit 220 , and a storage unit 230 .

The communication unit 210 performs functions for receiving a signal through a wireless channel. For example, the communication unit 210 may be configured to receive a radio frequency (RF) signal, frequency transform, demodulate, decode, remove a cyclic prefix (CP), fast Fourier transform (FFT), channel estimation, and equalizing ) and so on. In particular, the communication unit 210 may include a channel estimator 212 for estimating a channel for a radio channel of the wireless communication system 100 . The communication unit 210 may additionally perform a function of transmitting the signal processed by the control unit 220 to another node.

The controller 220 controls overall operations of the receiver 120. For example, the control unit 220 receives a signal through the communication unit 210 . In addition, the control unit 220 writes and reads data in the storage unit 230 . To this end, the controller 220 may include at least one processor, a microprocessor, or a microcontroller, or may be a part of the processor.

The storage unit 230 stores data such as a basic program, an application program, and setting information for the operation of the receiving terminal 120 . For example, the storage unit 230 performs functions for storing data processed by the control unit 220 . The storage unit 230 may be configured as a volatile memory, a non-volatile memory, or a combination of a volatile memory and a non-volatile memory. For example, the storage unit 230 may include a random access memory (RAM), a flash memory, or the like.

In FIG. 2 , the receiving terminal 120 is illustrated to include a communication unit 210 , a control unit 220 , and a storage unit 230 . According to another embodiment, the receiving terminal 120 may further include an additional configuration in addition to the aforementioned configuration.

3 illustrates a block configuration of the communication unit 210 of the receiving terminal 120 according to various embodiments of the present disclosure. Referring to FIG. 3 , the communication unit 210 may include an RF receiver 312 , an OFDM demodulator 314 , a signal demapper 316 , an equalizer 318 , and a channel estimator 212 .

The RF receiver 312 may perform a function of receiving an RF signal from a wireless channel. The RF receiver 312 may receive an RF signal through an antenna, and the antenna may be a single or multiple antennas.

The OFDM demodulator 314 may remove the CP inserted in the transmitter 110 and perform FFT. CP is a signal added to prevent interference between signals of orthogonal frequencies and loss of frequency orthogonality due to time spread in a radio channel in the OFDM system.

The signal demapper 316 may perform a function of de-mapping the signal in the frequency domain output through the OFDM demodulator 314 to an OFDM symbol and a resource element (RE).

The equalizer 318 may perform a function of compensating for an error caused by inter symbol interference (ISI) or channel noise.

The channel estimator 212 may extract a reference signal from the signal demapper 316, perform channel estimation using the reference signal, and transmit output data to the equalizer 318 . The channel estimator may include an IFFT processor or an FFT processor for channel estimation in a time domain or a frequency domain. The channel estimator may additionally include a module for extracting a reference signal or a pilot signal.

Although the channel estimator 212 and the equalizer 318 are illustrated as separate components in FIG. 3 , the channel estimator 212 may be included in the equalizer 318 or perform the above-described functions as the same component according to various embodiments.

4 illustrates a location of an RE occupied by a physical resource block (PRB) 400 and a cell-reference signal (CRS) 0 of LTE according to various embodiments of the present disclosure. In the following description, for simplicity of explanation, it is assumed that the transmitter 110 is the transmitter of the base station and the receiver 120 is the receiver of a user equipment (UE), and a signal transmitted from the transmitter is a downlink (DL) signal. Explain. However, the present disclosure is not limited to this case. For example, the present disclosure may also be applied to an LTE uplink sounding signal.

Clusters of a plurality of scatterers or reflectors exist on the radio channel. As a result, the receiving end receives the signal transmitted from the transmitting end through multiple paths on a radio channel. Since the LTE DL OFDM symbol has a CP in front of the effective (net) OFDM symbol, when the CP is removed at the receiving end and a sample equal to the size of the FFT is taken, inter symbol interference does not occur. . In LTE, it is defined as one subframe for a time of 1 ms length, and one subframe includes a plurality of physical resource blocks (PRBs). One PRB is composed of 14 OFDM symbols in the time domain and 12 resource elements (REs) in the frequency domain when the CP type is a general CP. Thus, one PRB includes 168 REs.

In order for the receiving end to perform coherent demodulation on the received signal, the base station uses a cell-specific reference signal (CRS), and each transmit antenna port is disjointly associated with the CRS. . The LTE standard supports 1, 2 and 4 CRS ports. A user equipment (UE) or a receiving end may know the number of ports of a base station in a process of demodulating a physical broadcast channel (PBCH). If the receiving end removes the CP and performs FFT on N FFT samples in the taken time domain, it is possible to obtain a reception signal in a specific RE to which CRS is allocated in the frequency domain. In FIG. 4, the location of the RE occupied by LTE CRS 0 is shown.

The length of the CP is set to be larger than the maximum delay length of multipath experienced by a signal in the radio channel. In the case of a general CP, the CP length of OFDM symbols 0 and 7 among 14 OFDM symbols is 5.208us, and the remaining OFDM The CP length of the symbol is 4.6865us. In the case of a system bandwidth of 10 MHz, the chip (FFT sample) interval is 65.104 ns, and the CP length of the above-described OFDM symbol corresponds to 80 chips and 72 chips, respectively. There is a channel defined as an enhanced typical urban (ETU) among several channels experienced by the UE in a general outdoor environment or a radio channel environment. According to the ETU, the delay spread value corresponds to 5 us, respectively, and it can be seen that the delay spread value is shorter than the length of the CPs of OFDM symbols 0 and 7 and longer than the length of the CPs of the remaining OFDM symbols. ETU is a channel consisting of 9 multipaths. That is, in the radio channel, the delay spread value comes within approximately the OFDM CP length and the channel taps are sparsely generated.

A channel experienced by an actual reception signal is generated by a combination of a transmission filter of a base station, a sparse radio channel, and a reception filter, and in this document, it is assumed that the channel changes with time within one subframe.

Hereinafter, a channel estimation technique in a single antenna system or a SISO system will be described.

에서

번째 채널 탭 값인

은 이하의 수학식 1과 같이 표현된다.Timing of the sampled received signal from an analog-to-digital converter (ADC)

at

The second channel tap value

is expressed as in Equation 1 below.

은 ADC에서 샘플링된 수신 신호의 타이밍

에서

번째 채널 탭 값,

는 희소한 무선 채널의 다중경로의 개수,

는 칩 길이,

는 수신 신호의 타이밍

에서

번째 경로의 계수이며,

는

번째 경로의 딜레이(delay)이고,

라고 할 때,

및

는 각각

및

으로 정의된다.

는 시간

에서 송신 필터 및 수신 필터에 의한 합성 필터(composite filter)이며,

는

길이를 가진다고 가정한다.

은 지연 확산 값으로

으로 표현된다.

은 크로네커 델타(Kronecker delta)이다. ADC 샘플 공간(ADC sample space)에서 바라본 타이밍

에서

번째채널 탭

에

번째 다중경로가 기여하는 성분은

중에서

일 때,

이다.

는

인 경우, 이하의 수학식 2를 만족하게 하는 상수이다.here

is the timing of the received signal sampled by the ADC

at

second channel tap value,

is the number of multipaths in the sparse radio channel,

is the chip length,

is the timing of the received signal

at

is the coefficient of the th path,

Is

is the delay of the th path,

when said,

and

is each

and

is defined as

is the time

It is a composite filter by a transmit filter and a receive filter in

Is

Assume that it has a length.

is the delay spread value.

is expressed as

is the Kronecker delta. Timing as seen from ADC sample space

at

second channel tab

to

The component contributed by the second multipath is

Between

when,

to be.

Is

In the case of , it is a constant that satisfies Equation 2 below.

는 기대값(expectation),

은 ADC에서 샘플링된 수신 신호의 타이밍

에서

번째 채널 탭 값,

은 지연 확산값이다.here

is the expected value,

is the timing of the received signal sampled by the ADC

at

second channel tap value,

is the delay spread value.

은

의 선형 중첩으로 표현되므로, 벡터

및

을

및

으로 정의하면,

을 이하의 수학식 3과 같이 표현할 수 있다.in Equation 1

silver

Since it is expressed as a linear superposition of

and

second

and

If defined as

can be expressed as in Equation 3 below.

은 누출 행렬(leakage matrix)로 불리며, 수학식 1을 만족한다. 여기서 위 첨자

는 전치(transpose)를 의미한다.here

is called a leakage matrix, and satisfies Equation (1). superscript here

means transpose.

의 CRS가 위치한 OFDM 심볼

,

의 RE

에서의 수신신호를 생각할 수 있다. CRS 값은 기지국과 UE가 서로 알고 있으므로, 수신신호에서 CRS 값을 나눈 신호

는 이하의 수학식 4와 같이 표현될 수 있다.A subframe transmitted from the transmit antenna of the base station to the receive antenna of the user equipment

OFDM symbol where CRS of

,

RE of

We can think of the received signal from Since the CRS value is known to the base station and the UE, a signal obtained by dividing the CRS value from the received signal

can be expressed as in Equation 4 below.

는 수신 신호에서 CRS 값을 나눈 신호,

는 CRS가 위치한

번째 OFDM 신호에서 RE

에서의 잡음 신호이다. 잡음 신호

의 분산(variance)을

으로 정의한다.

은 지연 확산 값이다. 한 개의 OFDM 심볼 구간 동안 채널이 변화하지 않는다고 가정하면,

은 OFDM 심볼

마다 채널 탭 값

을 샘플링한 값이다. 도 4에서

이 CRS가 위치한 OFDM 심볼이다. OFDM 심볼

에서 CRS가 차지하는 전체 RE의 개수를

라고 하고,

을 수신 안테나의 OFDM 심볼

에서 CRS가 위치한 주파수 영역 잡음을 오름차순으로 나열한 벡터라 하고,

을 수신단에서 수신된 OFDM 심볼

에서 CRS가 위치한 주파수 영역 수신 신호를 오름차순으로 나열한 벡터라 하자. 여기에서 수신단은 정확한

을 모르므로, 통상적인 채널의 지연 확산 값이 최대 CP 길이

라 가정하고, 송신 및 수신 필터에 의한 추가 지연 확산 값을 고려해서

을 이하의 수학식 5와 같이 정한다.here

is the signal obtained by dividing the CRS value from the received signal,

is where the CRS is located

RE in the second OFDM signal

is the noise signal in noise signal

the variance of

to be defined as

is the delay spread value. Assuming that the channel does not change during one OFDM symbol period,

is the OFDM symbol

per channel tap value

is the sampled value. 4 in

This is the OFDM symbol in which the CRS is located. OFDM symbol

The total number of REs occupied by CRS in

say,

is the OFDM symbol of the receiving antenna

Let the vector listing the frequency domain noise where the CRS is located in ascending order,

OFDM symbol received at the receiving end

Let it be a vector in which the frequency domain reception signal in which the CRS is located is arranged in ascending order. Here, the receiving end is

Since we do not know, the delay spread value of a typical channel is the maximum CP length.

, and considering the value of the additional delay spread by the transmit and receive filters,

is determined as in Equation 5 below.

은 지연 확산 값,

는 최대 CP길이,

는 송신 및 수신 필터에 의한 추가 지연 확산값이다.

is the delay spread value,

is the maximum CP length,

is the value of the additional delay spread by the transmit and receive filters.

로 정의하면, 수신 신호 벡터는 이하의 수학식 6과 같이 표현될 수 있다.Channel impulse response (CIR) vector

If defined as , the received signal vector can be expressed as in Equation 6 below.

은 수신 신호 벡터이고,

은 CIR 벡터,

은 잡음 벡터이며,

은

번째 행(row)과

번째 열(column) (

)의 엔트리(entry)를

로 갖는 행렬을

라 할 때, OFDM 심볼

에서 CRS가 위치한 RE의 부반송파 인덱스(subcarrier index)에 해당하는

의 행 벡터들 및 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이며,

의 크기는

이다. 도 4에 따르면, CRS 0이 존재하는 OFDM 심볼은 4개이고, CRS 0이 점유하는 RE의 부반송파 인덱스를 고려하면 이하의 수학식 7 및 8과 같다.here

is the received signal vector,

is the CIR vector,

is the noise vector,

silver

the second row and

the second column (

) of the entry

a matrix with

When , OFDM symbol

Corresponding to the subcarrier index of the RE where the CRS is located in

row vectors and column vectors of 0-to-column vectors

is a submatrix of column vectors up to

the size of

to be. According to FIG. 4, there are 4 OFDM symbols in which CRS 0 exists, and the following Equations 7 and 8 are expressed in consideration of the subcarrier index of the RE occupied by CRS 0.

에 위치해 있다고 하면, CRS가 위치하는 빗금 표시된 수신 RE가 채널 추정을 위해 이용되며, 서브프레임 및 심볼 인덱스는

이다. 도 5의 (b)에서 음영 표시된 영역의 3개의 OFDM 심볼들이 서브프레임

에 위치해 있다고 하면, CRS가 위치하는 빗금 표시된 수신 RE가 채널 추정을 위해 이용되며, 서브프레임 및 심볼 인덱스는

이다. 도 5의 (c)에서 음영 표시된 영역의 4개의 OFDM 심볼들이 서브프레임

에 위치해 있다고 하면, CRS가 위치하는 빗금 표시된 수신 RE가 채널 추정을 위해 이용되며, 서브프레임 및 심볼 인덱스는

이다. 도 5의 (d)에서 음영 표시된 영역의 3개의 OFDM 심볼들이 서브프레임

에 위치해 있다고 하면, CRS가 위치하는 빗금 표시된 수신 RE가 채널 추정을 위해 이용되며, 서브프레임 및 심볼 인덱스는

이다. 도 5에서는 음영 표시된 영역 내의 첫번째 OFDM 심볼 기준으로 최대 7 OFDM 심볼 지연까지 허용하여 채널 추정을 위해 CRS가 위치한 RE를 관측하였으나, 도 5는 단지 예시에 불과하다. 따라서, 다양한 실시 예들에 따라 더 큰 심볼 지연이 허용될 수 있으며, 이러한 경우에도 본 개시의 내용이 적용될 수 있다.5 illustrates a CRS pattern used for channel estimation according to an embodiment of the present disclosure. In FIG. 5 , one subframe may be divided into four sections. For example, one subframe may be divided as shown in shaded portions in FIGS. 5A to 5D . A line protruding from the top in each area means a subframe boundary. 4 OFDM symbols in the shaded area in (a) of FIG. 5 are subframes

If it is located in , the hatched received RE in which the CRS is located is used for channel estimation, and the subframe and symbol index are

to be. In FIG. 5B, three OFDM symbols in the shaded area are subframes.

If it is located in , the hatched received RE in which the CRS is located is used for channel estimation, and the subframe and symbol index are

to be. 4 OFDM symbols in the shaded area in (c) of FIG. 5 are subframes

If it is located in , the hatched received RE in which the CRS is located is used for channel estimation, and the subframe and symbol index are

to be. 3 OFDM symbols in the shaded area in (d) of FIG. 5 are subframes

If it is located in , the hatched received RE in which the CRS is located is used for channel estimation, and the subframe and symbol index are

to be. In FIG. 5, the RE in which the CRS is located was observed for channel estimation by allowing up to 7 OFDM symbol delays based on the first OFDM symbol in the shaded area, but FIG. 5 is only an example. Accordingly, a larger symbol delay may be tolerated according to various embodiments, and even in this case, the contents of the present disclosure may be applied.

의 표시된 영역의 채널 추정을 위해 CRS가 위치한 RE에서의 수신 신호에서 CRS 값을 나눈 수신 신호 벡터

는 이하의 수학식 9와 같이 정의될 수 있다.subframe

Received signal vector obtained by dividing the CRS value from the received signal at the RE where the CRS is located for channel estimation of the indicated area of

may be defined as in Equation 9 below.

는 수신 신호 벡터,

는 시스템 행렬,

는 채널 벡터,

는 잡음 벡터이다.here

is the received signal vector,

is the system matrix,

is the channel vector,

is the noise vector.

는 이하의 수학식 10과 같이 표현할 수 있다.In the case of (a) and (c) of Figure 5,

can be expressed as in Equation 10 below.

는 시스템 행렬,

은

번째 행(row)와

번째 열(column) (

)의 엔트리를

로 갖는 행렬을

라 할 때, OFDM 심볼

에서 CRS가 위치한 RE의 부반송파 인덱스에 해당하는

의 행 벡터들 및 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이다.here

is the system matrix,

silver

the second row and

the second column (

) of the entry

a matrix with

When , OFDM symbol

Corresponding to the subcarrier index of the RE where the CRS is located in

row vectors and column vectors of 0-to-column vectors

Up to is a submatrix of column vectors.

는 이하의 수학식 11과 같이 표현할 수 있다.In the case of (b) and (d) of Figure 5,

can be expressed as in Equation 11 below.

는 시스템 행렬,

은

번째 행과

번째 열 (

)의 엔트리를

로 갖는 행렬을

라 할 때, OFDM 심볼

에서 CRS가 위치한 RE의 부반송파 인덱스에 해당하는

의 행 벡터들 및 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이다.here

is the system matrix,

silver

second row and

second column (

) of the entry

a matrix with

When , OFDM symbol

Corresponding to the subcarrier index of the RE where the CRS is located in

row vectors and column vectors of 0-to-column vectors

Up to is a submatrix of column vectors.

, 채널 응답 벡터

, 및 잡음 벡터

는 이하의 수학식 12 내지 14와 같이 정의될 수 있다.In the case of the shaded area in Fig. 5 (a), the signal vector

, the channel response vector

, and the noise vector

may be defined as in Equations 12 to 14 below.

, 채널 응답 벡터

, 및 잡음 벡터

는 이하의 수학식 15 내지 17과 같이 정의될 수 있다.In the case of the shaded area in Fig. 5 (b), the signal vector

, the channel response vector

, and the noise vector

may be defined as in Equations 15 to 17 below.

, 채널 응답 벡터

, 및 잡음 벡터

는 이하의 수학식 18 내지 20과 같이 정의될 수 있다.In the case of the shaded area in Fig. 5(c), the signal vector

, the channel response vector

, and the noise vector

may be defined as in Equations 18 to 20 below.

, 채널 응답 벡터

, 및 잡음 벡터

는 이하의 수학식 21 내지 23과 같이 정의될 수 있다.In the case of the shaded area in Fig. 5(d), the signal vector

, the channel response vector

, and the noise vector

may be defined as in Equations 21 to 23 below.

의 RE

에서의 채널 주파수 응답(channel frequency response, CFR)은 이하의 수학식 24와 같이 정의될 수 있다.OFDM symbols for shaded areas in grids 510, 520, 530 and 540 of FIG. 5

RE of

A channel frequency response (CFR) in can be defined as in Equation 24 below.

는 OFDM 심볼

의 RE

에서의 CFR,

은 OFDM 심볼

마다 채널 탭 값

을 샘플링한 값이다.

is the OFDM symbol

RE of

CFR in

is the OFDM symbol

per channel tap value

is the sampled value.

의 추정치는 선형 최소 평균 제곱 오차(linear minimum mean square error, LMMSE) 방식을 이용해서 구할 수 있으며, 이하의 수학식 25와 같다.According to various embodiments of the present disclosure, CFR

The estimated value of can be obtained using a linear minimum mean square error (LMMSE) method, and is expressed in Equation 25 below.

은 CFR,

은 CFR의 추정 값,

는 기대 값,

는 신호 벡터,

는 시스템 행렬,

는 크기가 4N_CRS인 단위 행렬, 윗 첨자 행렬의 H는 복소 공액 및 전치를 의미한다.

Silver CFR,

is an estimate of the CFR,

is the expected value,

is the signal vector,

is the system matrix,

is the identity matrix of size 4N _CRS , and H in the superscript matrix denotes complex conjugates and transposes.

는 채널 탭이 희소하게 형성된 희소 행렬이다. 수학식 25의 LMMSE 방식을 이용한 CFR의 추정치는 채널 응답 벡터

의 0이 아닌 엔트리의 위치를 수신기가 알고 있는 것을 내포하고 있다. 채널 응답 벡터

의 0이 아닌 엔트리는

의 서포트(support)로 지칭될 수 있다.Channel response vector as described above

is a sparse matrix in which channel taps are sparsely formed. Estimation of CFR using the LMMSE method of Equation 25 is a channel response vector

It implies that the receiver knows the location of the non-zero entry in . Channel Response Vector

A non-zero entry in

may be referred to as support of

The actual receiving end does not know how the base station's transmit filter is made or configured. In addition, it is difficult for the actual receiver to know exactly the position of the channel tap support in a noisy environment, and in the case of the LMMSE estimator, the second moment value of the channel must be known to obtain the above estimate. Therefore, a channel estimator that is easy to implement in the modem of the actual receiving end may be required.

6A illustrates channel estimation based on inverse discrete Fourier transform (IDFT) using reference signal time interpolation (RSTI) according to various embodiments of the present disclosure. In FIG. 6A , the case of FIG. 5C is illustrated as an example, but may be applied to the case of the extended subframe as well as FIGS. 5A to 5D . Information on the channel of the RE indicated in black in FIG. 6A can be obtained by linearly interpolating two descrambled signals located in adjacent REs on the time axis. The received signals located in the REs arranged at intervals of three subcarriers obtained in this way are transformed into the time domain and channel estimation is performed. After denoising, an estimate of the CFR is obtained by transforming the CIR estimate into the frequency domain. An estimate of the CFR of an OFDM symbol in which the CRS is not located is obtained by linearly interpolating the estimates of two adjacent CFRs.

6B illustrates an implementation of a time-varying channel and a channel estimate by linear interpolation according to various embodiments of the present disclosure. 6B follows Jakes' model and assumes a Doppler frequency of 300 Hz. The Doppler frequency of 300 Hz corresponds to the environment moving at 120.4 km/h in 2.69 GHz band communication. In FIG. 6B , the horizontal axis represents time, and one sample means a chip period during one subframe. In FIG. 6B , the vertical axis means power on a dB scale. The horizontal solid line indicates the OFDM symbol section including the CRS or reference signal, and there is no CRS in the OFDM symbol between the horizontal solid lines. The solid line of the curve represents the power over time according to the channel model described above. A dotted line indicates a result of linear interpolation between OFDM symbols in which CRSs exist by using an IFFT-based channel estimation method using RSTI. That is, the dotted line is the result of estimating the channel by assuming that the channel changes linearly, and it can be seen that the difference in power between the dotted line and the solid line is up to 2 dB. Therefore, in the case of a channel that changes rapidly with time, a method for sophisticated channel estimation is required.

First, a system equation assuming a stop channel is as shown in Equation 26 below.

는 수신 신호 벡터,

는 시스템 행렬,

는 채널 벡터,

는 잡음 벡터이다.here

is the received signal vector,

is the system matrix,

is the channel vector,

is the noise vector.

는 도 5의 (a) 및 (c)의 경우, 이하의 수학식 27과 같다.where the system matrix

In the case of (a) and (c) of Figure 5, is the same as the following Equation 27.

는 시스템 행렬,

은

번째 행(row)와

번째 열 (

)의 엔트리를

로 갖는 행렬을

라 할 때,OFDM 심볼

에서 CRS가 위치한 RE의 부반송파 인덱스에 해당하는

의 행 벡터들 및 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이다.here

is the system matrix,

silver

the second row and

second column (

) of the entry

a matrix with

When , OFDM symbol

Corresponding to the subcarrier index of the RE where the CRS is located in

row vectors and column vectors of 0-to-column vectors

Up to is a submatrix of column vectors.

는 도 5의 (b) 및 (d)의 경우, 이하의 수학식 28과 같다.Also, the system matrix

In the case of (b) and (d) of Figure 5, is the same as the following Equation 28.

는 시스템 행렬,

은

번째 행과

번째 열 (

)의 엔트리를

로 갖는 행렬을

라 할 때,OFDM 심볼

에서 CRS가 위치한 RE의 부반송파 인덱스에 해당하는

의 행 벡터들 및 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이다.here

is the system matrix,

silver

second row and

second column (

) of the entry

a matrix with

When , OFDM symbol

Corresponding to the subcarrier index of the RE where the CRS is located in

row vectors and column vectors of 0-to-column vectors

Up to is a submatrix of column vectors.

는 다음의 수학식 29와 같다.channel vector

is the following Equation 29.

및

는 각각 수학식 12 및 수학식 14, 도 5의 (b)의 경우

및

는 각각 수학식 15 및 수학식 17, 도 5의 (c)의 경우

및

는 각각 수학식 18 및 수학식 20, 도 5의 (d)의 경우 및

는 각각 수학식 21 및 수학식 23을 따를 수 있다.In the case of Figure 5 (a),

and

is Equation 12 and Equation 14, respectively, in the case of (b) of FIG.

and

are Equations 15 and 17, respectively, in the case of (c) of FIG.

and

In the case of Equations 18 and 20, FIG. 5 (d), respectively, and

may follow Equations 21 and 23, respectively.

7 is a flowchart 700 illustrating a channel estimation method using a reference signal according to various embodiments of the present disclosure. 7 is an example of an operation method of the receiving terminal 120 .

Referring to FIG. 7 , the receiving terminal 120 receives a reference signal in step 701 . Here, the reference signal is a signal shared by the transmitter 110 and the receiver 120 for channel estimation, and means a signal transmitted from the transmitter 110 to the receiver. can

In step 703, the receiver 120 determines a first channel estimate by performing channel estimation on the section occupied by the reference signal. The first channel estimate may be calculated by dividing a reference signal received at the receiving end by a reference signal value (eg, CRS value) shared between the transmitting end and the receiving end. Channel estimate means herein the estimated value of a channel vector or entry of a channel vector.

In step 705 , the receiver 120 determines the second channel estimate by interpolating between the first channel estimates on the time axis for the remaining section in which the reference signal does not exist. That is, the receiving terminal 120 calculates a second channel estimate for regions other than the regions including the reference signal by interpolating the first channel estimates. The channel estimation performed in step 703 or step 705 may be performed by an orthogonal matching pursuit (OMP) algorithm, a stagewise OMP (StOMP) algorithm, a compressive sampling matching pursuit (CoSaMP) algorithm, etc. can The interpolation method between the first channel estimates includes a method of interpolating with an average value between two adjacent estimates, a method of interpolating a straight line between two adjacent estimates, and a method of interpolating between two adjacent estimates in a straight line or a curve using a multidimensional function. How to interpolate.

In operation 705, the receiving end 120 may determine a channel estimate by interpolating between the first estimates using a basis expansion model (BEM). The BEM represents a channel varying in the time axis as a product of at least one basis and at least one coefficient. For BEM, Legendre Polynomial (LP) having a polynomial form may be used.

The interpolation method used in step 705 is a Doppler frequency according to the moving speed of the receiving end 110 or the transmitting end 120, or the quality of the channel (eg, signal to noise ratio (SNR) or carrier to noise ratio, CNR)). The interpolation method may vary according to selection of a basis used for estimating the channel change pattern. In other words, the method of interpolation may vary depending on which basis from among a plurality of basis is selected. For example, the multiple basis may include a first basis representing a constant value, a second basis representing a linear change, i.e., a first-order equation, and a third basis representing a curve-shaped change, i.e., a second-order or higher-order equation. have. In the case of using the LP BEM, the number of BEM basis may be selected according to the moving speed or the SNR.

8 is a flowchart 731 illustrating a channel estimation method using BEM according to various embodiments of the present disclosure. The flowchart 731 of FIG. 8 may be included as a part in steps 701 to 705 of FIG. 7 or may be further included in steps 701 to 705 .

Referring to FIG. 8 , in step 801, the receiving end reconstructs and expresses a system matrix used for channel estimation using a vector composed of BEM coefficients. In detail, a channel vector may be reconstructed as a product of a BEM basis vector and a coefficient of each basis, and a new system matrix may be expressed as a product of a previous system matrix and a basis vector. In step 803, the receiving end estimates the BEM coefficients used to newly express the system matrix in step 801. Estimation of the BEM coefficients may be estimated through the StOMP algorithm, or block StOMP. In step 805, the receiver performs channel estimation for an area in which a reference signal is not received by using the BEM coefficient estimated in step 803. Channel estimation may be performed for the remaining section by interpolating between regions in which the reference signal is received using the estimated BEM coefficients and a basis corresponding to each coefficient. Here, the number of basis may be determined by the moving speed of the receiver or the instantaneous SNR.

9 is a flowchart 900 in which the number of BEMs is adaptively selected according to a moving speed of a receiver according to various embodiments of the present disclosure. The moving speed of the receiver may be measured using a Doppler detector.

The instantaneous SNR may be defined as an average value of SNRs of OFDM symbols occupied by CRS for which channel estimation is used within one subframe, and infinite impulse response (IIR) of SNRs of OFDM symbols received over several subframes and in which CRS exists. ) filter or a finite impulse response (FIR) filter output.

Depending on the Doppler frequency due to the movement of the receiver, a method of estimating a channel for a region in which a reference signal does not exist may vary. In an embodiment, when interpolating between channel estimates for a region in which a reference signal exists, an average value between channel estimates, linear interpolation, curved interpolation, etc. may be determined according to a moving speed. When channel estimation is performed by applying the BEM method, the number of BEM bases or the order of a polynomial in a polynomial-type BEM may be selected.

를 이용하도록 결정한다. 도플러 주파수가 낮은 주파수가 아니면, 903 단계에서 수신단 102는 도플러 주파수가 중간의 주파수인지 여부를 판단한다. 도플러 주파수가 중간의 주파수인 경우, 909 단계에서 수신단 102는 두 개의 BEM 베이시스들

및

를 이용하도록 결정한다. 도플러 주파수가 중간의 주파수가 아닌 경우, 905 단계에서 수신단 102는 도플러 주파수가 높은 주파수로 결정하고, 911 단계에서 수신단 102는 세 개의 BEM 베이시스들

,

, 및

를 이용하도록 결정한다. According to the embodiment of FIG. 9 using a Legendre polynomial (LP) BEM, which will be described in detail below, in step 901 , the receiver 102 determines whether the Doppler frequency is a low frequency. If the Doppler frequency is a low frequency, in step 907, the receiving end 102 receives one BEM basis.

decide to use If the Doppler frequency is not a low frequency, in step 903, the receiver 102 determines whether the Doppler frequency is an intermediate frequency. If the Doppler frequency is an intermediate frequency, in step 909, the receiving end 102 performs two BEM basis values.

and

decide to use If the Doppler frequency is not an intermediate frequency, in step 905, the receiving end 102 determines that the Doppler frequency is a high frequency, and in step 911, the receiving end 102 determines the three BEM bases.

,

, and

decide to use

Although it is illustrated in FIG. 9 that a maximum of three BEM basis is used, the number of BEM basis of three or more may be used according to an embodiment.

10 is a flowchart 1000 in which the number of BEMs is adaptively selected according to a moving speed and an instantaneous SNR according to various embodiments of the present disclosure. The receiving end 120 may use a Doppler detector and an SNR meter to measure Doppler frequency and SNR.

Depending on the Doppler frequency and the instantaneous SNR, a method of estimating a channel for a region in which a reference signal does not exist may vary. According to an embodiment, when interpolating between channel estimates for a region in which a reference signal exists, an average value between channel estimates, linear interpolation, curved interpolation, etc. may be determined according to the moving speed and the instantaneous SNR. When channel estimation is performed by applying the BEM method, the number of BEM bases or the order of a polynomial in a polynomial-type BEM may be selected.

를 이용하도록 결정한다. 수신단 120이 도플러 주파수가 낮은 주파수가 아니라고 판단하면, 1003 단계에서, 수신단 120은 도플러 주파수가 중간의 주파수인지 여부를 판단한다. 수신단 120이 도플러 주파수가 중간의 주파수라고 판단하면, 1005 단계에서, 수신단 120은 슬롯(slot) 또는 서브프레임 레이트로 측정된 순시 SNR 과 특정 기준 값 T_m,1을 비교한다. 수신단 120이 순시 SNR이 특정값 T_m,1보다 크다고 판단하면, 1017 단계에서, 수신단 120은 두 개의 BEM 베이시스

및

를 이용하도록 결정한다. 수신단 120이 순시 SNR이 특정값 T_m,1보다 같거나 작다고 판단하면 한 개의 BEM 베이시스

를 이용하도록 결정한다. 1003 단계에서 수신단 120이 도플러 주파수가 중간의 주파수가 아니라고 판단하면, 수신단 120은 도플러 주파수가 높은 주파수라고 판단하고, 1009 단계에서, 수신단 120은 순시 SNR이 특정 기준 값 T_h,1 을 비교한다. 수신단 120이 순시 SNR이 T_h,1보다 작거나 같다고 판단하면, 1019 단계에서, 수신단 120은 한 개의 BEM 베이시스

를 이용하도록 결정한다. 1009 단계에서 수신단 120이 순시 SNR이 T_h,1보다 크다고 판단하면, 1011 단계에서, 수신단 120은 순시 SNR과 특정 기준 값 T_h,2를 비교한다. 수신단 120이 순시 SNR이 특정 기준 값 T_h,2보다 작거나 같다고 판단하면, 1021 단계에서 수신단은 120은 두 개의 BEM 베이시스들

및

를 이용하도록 결정한다. 수신단 120이 순시 SNR이 T_h,2보다 크다고 판단하면, 1023 단계에서, 수신단 120은 세 개의 BEM 베이시스들

,

, 및

를 이용하도록 결정한다.According to the embodiment of FIG. 10 using an LP BEM, which will be described in detail below, the receiving terminal 120 determines whether the Doppler frequency is a low frequency in step 1001 . If the receiving end 120 determines that the Doppler frequency is a low frequency, in step 1013, the receiving end 120 is one BEM basis.

decide to use If the receiving end 120 determines that the Doppler frequency is not a low frequency, in step 1003, the receiving end 120 determines whether the Doppler frequency is an intermediate frequency. When the receiving end 120 determines that the Doppler frequency is an intermediate frequency, in step 1005, the receiving end 120 compares the instantaneous SNR measured at a slot or subframe rate with a specific reference value T _m,1 . If the receiving end 120 determines that the instantaneous SNR is greater than the specific value T _m,1 , in step 1017, the receiving end 120 has two BEM basis

and

decide to use If the receiver 120 determines that the instantaneous SNR is less than or equal to the specific value T _m,1, one BEM basis

decide to use If the receiving end 120 determines that the Doppler frequency is not an intermediate frequency in step 1003, the receiving end 120 determines that the Doppler frequency is a high frequency, and in step 1009, the receiving end 120 compares the instantaneous SNR with a specific reference value T _h,1 . If the receiving end 120 determines that the instantaneous SNR is less than or equal to T _h,1 , in step 1019, the receiving end 120 sets one BEM basis

decide to use If the receiver 120 determines that the instantaneous SNR is greater than T _h,1 in step 1009 , in step 1011 , the receiver 120 compares the instantaneous SNR with a specific reference value T _h,2 . If the receiving end 120 determines that the instantaneous SNR is less than or equal to the specific reference value T _h,2 , in step 1021, the receiving end 120 determines the two BEM basis

and

decide to use If the receiving end 120 determines that the instantaneous SNR is greater than T _h,2 , in step 1023, the receiving end 120 performs three BEM basis values.

,

, and

decide to use

Although in FIG. 10 a maximum of three BEM basis is shown to be used. However, according to various embodiments, four or more BEM basis may be used.

가 희소 벡터(sparse vector)이므로 압축 센싱(compressing sensing, CS) 기법을 활용하여 0이 아닌 엔트리, 즉 서포트에 대한 복구를 수행하면, 최적 추정기에 근접한 성능을 얻을 수 있다. CS 기법 중 대표적으로 알려진 것이 OMP 알고리즘이고, OMP 알고리즘을 개량한 StOMP 알고리즘 및 CoSaMP 알고리즘이 있다. 본원에서는 강인하게(robust) 동작하는 StOMP 알고리즘을 기준으로 설명한다.11 illustrates a StOMP algorithm according to various embodiments of the present disclosure. As described above, the channel vector

Since is a sparse vector, performance close to the optimal estimator can be obtained if non-zero entries, ie, supports, are restored using a compression sensing (CS) technique. Among the CS techniques, the OMP algorithm is known as representative, and there are the StOMP algorithm and the CoSaMP algorithm that are improved OMP algorithms. Herein, the description will be made based on the robustly operating StOMP algorithm.

은 s번째 단계에서의 잔여 벡터(residual vector)이며, s=1에서

이고, 집합

는

이다. 여기서,

는 기준 신호들의 수신 값들을 포함하는 벡터이다. Referring to Figure 11,

is the residual vector in the s-th step, where s = 1

and set

Is

to be. here,

is a vector including received values of reference signals.

에 정합 필터(matched filter)를 적용한 벡터

를 계산한다.

의 정합 필터(matched filter) 출력

의 j번째 엔트리는

와

의 j번째 열 벡터의 매칭이 어느 정도 되어 있는지 알려주는 엔트리이다. In step 1110, the receiving end 120 receives the residual vector of the s (≥1)-th iteration stage.

vector with matched filter applied to

to calculate

matched filter output of

The jth entry of

Wow

This is an entry that indicates how well the j-th column vector of is matched.

의 엔트리 중에 에너지값 또는 절대값이 기준 값

보다 큰 엔트리들의 집합

를 출력하여 1130 단계에서 이전 단계에서 저장되어 있던 서포트 집합

에 추가한다. 즉, 수신단 120은 이전 반복 스테이지에서 저장되어 있던

과

의 합집합을 구하고 그 결과를

로 정의한다.

는 오름차순으로 나열될 수 있다. s+1번째 반복 스테이지에서는 s번째 반복 스테이지에서 구한

가 합집합의 입력이 될 수 있으며, 이를 위해 딜레이 유닛(delay unit)을 D로 표시하였다. 1120 단계에서의 기준 값

는 이하의 수학식 30과 같이 계산될 수 있다.In step 1120, the receiving end 120

The energy value or absolute value is the reference value during the entry of

a larger set of entries

output the support set stored in the previous step in step 1130

add to That is, the receiver 120 receives the stored data in the previous iteration stage.

class

Find the union of

is defined as

may be listed in ascending order. In the s+1th iteration stage, the value obtained in the sth iteration stage is

may be an input of the union, and for this purpose, a delay unit is denoted as D. Reference value at step 1120

can be calculated as in Equation 30 below.

은 미리 정의된 계수,

는 s 단계에서의 잔여 벡터

의 2 norm,

는 OFDM 심볼에서 CRS가 차지하는 전체 RE의 개수이다.here

is the predefined coefficient,

is the residual vector in step s

2 norm of

is the total number of REs occupied by the CRS in the OFDM symbol.

의 열 벡터 중에

의 엔트리에 해당하는 열 벡터 만을 모은 부분행렬이

로 정의되며, s번째 반복 스테이지에서의 시스템 행렬

및 수신 신호 벡터

를 입력 벡터로 하여 제로 포싱(zero forcing) 수신기를 통하여 추정된 CIR 벡터

가 출력 벡터로서 출력된다.In step 1140,

Among the column vectors of

A submatrix that collects only the column vectors corresponding to the entries of

is defined as, the system matrix at the s-th iteration stage

and received signal vector

CIR vector estimated through a zero forcing receiver with

is output as an output vector.

로부터 추정된 CIR 벡터

가 기여하는 값을 빼기 위해

를 간섭 벡터로 정의하여 간섭 벡터를 형성한다. 수신 신호

에서

를 빼준 값을

로 정의하며, 이를 위해 딜레이 유닛이 이용될 수 있다.In step 1150, the received signal

CIR vector estimated from

to subtract the value contributed by

is defined as an interference vector to form an interference vector. receive signal

at

minus the value

, and for this purpose, a delay unit may be used.

의 2 norm

이 기준값보다 작아질 때, 또는

의 엔트리 중 절대값이 가장 큰 엔트리가 기준 값보다 작아질 때 멈출 수 있다.This repetition operation determines the maximum number of repetitions and when the number of repetitions is reached,

of 2 norm

When it becomes less than this reference value, or

It can be stopped when the entry with the largest absolute value among the entries in is smaller than the reference value.

는 서포트 집합

이 갖는 인덱스에서는 0이 아닌 값을 갖게되며, 서포트 집합

의 여집합이 갖는 인덱스에서는 0인 값을 갖게된다. 추정된 CIR 벡터

로부터 CFR을 추정할 수 있으며, 이하의 수학식 31과 같다.CIR vector estimated using StOMP algorithm

is the support set

In this index, it has a non-zero value, and the support set

The index of the complement set has a value of 0. Estimated CIR vector

CFR can be estimated from

는 CFR의 추정치,

은 CIR의 추정치이다.

is the estimate of the CFR,

is an estimate of the CIR.

12 illustrates a StOMP algorithm 1200 using BEM according to various embodiments of the present disclosure.

Before introducing the algorithm, the BEM will be described first, and the basis of the Legendre polynomial (LP) for M channel samples having a polynomial form among BEMs well known to those skilled in the art is represented by Equations 32 to 34 below. can

은 0차 다항식을 담당하는 베이시스이며,

는 0차 및 1차 다항식을 담당하는 베이시스,

은 0차, 1차, 및 2차 다항식을 담당하는 베이시스이다. 더 높은 차수의 LP는 이 분야의 당업자에게 잘 알려져 있으므로 생략한다. 이때, 수학식 13, 16, 19, 22의 CIR 벡터

는 이하의 수학식 35와 같이 표현될 수 있다.here

is the basis responsible for the zero-order polynomial,

is the basis responsible for the 0th and 1st polynomials,

is the basis responsible for 0th, 1st, and 2nd order polynomials. Higher order LPs are well known to those skilled in the art and are therefore omitted. At this time, the CIR vectors of Equations 13, 16, 19, and 22

can be expressed as in Equation 35 below.

는 채널 벡터,

는 베이시스 행렬,

는 베이시스 계수 벡터이다.

is the channel vector,

is the basis matrix,

is the basis coefficient vector.

는 이하의 수학식 36 및 37과 같이 표현될 수 있다.where the coefficient vector

can be expressed as Equations 36 and 37 below.

은 이하의 수학식 38 내지 42와 같이 표현될 수 있다.In the case of (a) and (c) of FIG. 5, the basis matrix in the OFDM symbol with CRS

can be expressed as Equations 38 to 42 below.

는 베이이스 행렬,

은

인 단위 행렬(identity matrix)이고,

은 크로네커 곱셈(Kronecket product)이다.here

is the base matrix,

silver

is an identity matrix,

is the Kronecket product.

은 이하의 수학식 43 내지 47로 표현될 수 있다.In the case of (b) and (d) of FIG. 5, the basis matrix in the OFDM symbol with CRS

can be expressed by Equations 43 to 47 below.

From Equations 9 and 35, the following Equation 48 can be obtained.

는 수신 신호 벡터,

는 새롭게 정의된 시스템 행렬,

는 베이시스 계수 벡터, 는

는 잡음 벡터이다.

is the received signal vector,

is the newly defined system matrix,

is the basis coefficient vector, and is

is the noise vector.

는 도 5의 (a) 및 (c)의 경우 이하의 수학식 49와 같이 표현될 수 있다.Here, the newly defined system matrix

In the case of (a) and (c) of Figure 5 can be expressed as in Equation 49 below.

은

번째 행과

번째 열 (

)의 엔트리를

로 갖는 행렬을

라 할 때, OFDM 심볼

에서 CRS가 위치한 RE의 부반송파 인덱스에 해당하는

의 행 벡터들 및 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이고,

는 베이시스 행렬이다.

silver

second row and

second column (

) of the entry

a matrix with

When , OFDM symbol

Corresponding to the subcarrier index of the RE where the CRS is located in

row vectors and column vectors of 0-to-column vectors

is a submatrix of column vectors up to

is the basis matrix.

는 도 5의 (b) 및 (d)의 경우 이하의 수학식 50과 같이 표현될 수 있다.Newly defined system matrix

In the case of (b) and (d) of Figure 5 can be expressed as in Equation 50 below.

은

번째 행과

번째 열(column) (

)의 엔트리(entry)를

로 갖는 행렬을

라 할 때, OFDM 심볼

에서 CRS가 위치한 RE의 부반송파 인덱스(subcarrier index)에 해당하는

의 행 벡터들 및 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이고,

는 베이시스 행렬이다.

silver

second row and

the second column (

) of the entry

a matrix with

When , OFDM symbol

Corresponding to the subcarrier index of the RE where the CRS is located in

row vectors and column vectors of 0-to-column vectors

is a submatrix of column vectors up to

is the basis matrix.

의 j번째 열의 정규화된 절대값에 의해 하드 임계처리 및 서브셋 선택이 수행된다.

은

의 j 번째 열의 2-놈 이다. StOMP 알고리즘의

는 이하의 수학식 51과 같이 계산될 수 있다.12 shows the operation of the StOMP algorithm for the newly defined Equation (48). 12 , in block 1220

Hard thresholding and subset selection are performed by the normalized absolute value of the j-th column of .

silver

is the 2-norm of the j-th column of of the StOMP algorithm.

can be calculated as in Equation 51 below.

는 미리 정의된 계수,

는 s 단계에서의 잔여 벡터

의 2-놈,

는 OFDM 심볼에서 CRS가 차지하는 전체 RE의 개수이다.here

is the predefined coefficient,

is the residual vector in step s

of the 2-nome,

is the total number of REs occupied by the CRS in the OFDM symbol.

는 서포트 집합

이 갖는 인덱스에서는 0이 아닌 값을 가지며,

의 여집합이 갖는 인덱스에서는 0인 값을 갖게 된다. 추정된 베이시스 계수 벡터

로부터 이하의 수학식 52와 같이 원래의 추정된 CIR 벡터를 구할 수 있다.Basis coefficient vector estimated using StOMP algorithm

is the support set

In this index, it has a non-zero value,

At the index of the complement of , it has a value of 0. Estimated basis coefficient vector

From Equation 52 below, the original estimated CIR vector can be obtained.

는 추정된 채널 벡터,

는 베이시스 행렬,

는 추정된 베이시스 계수 벡터이다.

is the estimated channel vector,

is the basis matrix,

is the estimated basis coefficient vector.

는 이하의 수학식 53 내지 57과 같이 표현될 수 있다.In the case of (a) and (c) of Fig. 5, the basis matrix

can be expressed as in Equations 53 to 57 below.

는 이하의 수학식 58과 같이 표현될 수 있다.In the case of (b) and (d) of Fig. 5, the basis matrix

can be expressed as in Equation 58 below.

,

, 및

는 수학식 54 내지 56을 따른다.here

,

, and

follows Equations 54 to 56.

In the case of (a) of FIG. 5 , the estimated CIR vector may be expressed as Equation 59 below.

In the case of (b) of FIG. 5 , the estimated CIR vector may be expressed as Equation 60 below.

In the case of (c) of FIG. 5 , the estimated CIR vector may be expressed as Equation 61 below.

In the case of (d) of FIG. 5, the estimated CIR vector may be expressed as Equation 62 below.

으로부터 이하의 수학식 63과 같이 CFR을 추정할 수 있다.Estimated CIR vector

From Equation 63 below, the CFR can be estimated.

는 CFR의 추정치,

은 CIR의 추정치이다.

is the estimate of the CFR,

is an estimate of the CIR.

13 is a flowchart 1300 of an operation of the receiving terminal 120 performing the StOMP algorithm using the BEM according to various embodiments of the present disclosure.

Referring to FIG. 13 , in step 1301, the receiver 120 calculates a vector to which a matched filter is applied to the residual vector. Each entry of the vector to which the matched filter is applied corresponds to information on how much the residual signal matches the system matrix.

In step 1303, the receiver 120 adds a set of entries having an energy value greater than the reference value among the vector entries calculated in step 1301 to the support set. The newly calculated support index can be added to the support set by unioning the set consisting of the support index, which is a non-zero entry in the channel vector, which is a sparse vector, with the support set in the previous iteration stage.

In step 1305, the receiver 120 estimates the BEM coefficients by using a partial matrix of a system matrix consisting only of column vectors corresponding to the entry of the support set in step 1303. Estimation of the BEM coefficient may be performed by zero forcing. Here, the BEM coefficient means the coefficient of the basis of the channel vector reconstructed with the BEM basis.

In step 1307, the receiving end 120 updates the residual vector. The receiver 120 forms an interference vector to subtract a value contributed by the estimated BEM coefficient, defines a value obtained by subtracting the interference vector from the received signal as a new residual vector, and updates the residual vector.

14 illustrates a block StOMP algorithm 1400 using BEM according to various embodiments of the present disclosure.

의 엔트리 개수는

인데,

개의 엔트리들을 하나의 블록으로 정의하고 각 블록마다의 엔트리들의 값은 모두 0이거나 0이 아닌 엔트리로 구성된다.Here, a block is defined as entries in which all entries have zero or non-zero values. in other words,

The number of entries in

is,

n entries are defined as one block, and the values of entries for each block are all 0 or non-zero entries.

은 s번째 단계에서의 잔여 벡터이며, s=1에서

이고, 집합

는

이다. s(≥1)번째 반복 스테이지의 1410단계에서 잔여 벡터

에 정합 필터를 적용한 벡터

를 계산한다.

의 정합 필터 출력

의 j번째 엔트리는

와 의 j번째 열 벡터의 매칭이 어느 정도 되어 있는지 알려주는 엔트리이다.

의 엔트리 개수는

인데,

개의 엔트리들을 하나의 블록으로 정의한다. 그리하여

는

개의 블록을 갖게 된다. 1420단계에서

의

(

)번째 블록 내의 엔트리들의 파워를

으로 나눈 값의 합의 제곱근이 기준 값

보다 큰 블록들의 집합

를 출력한다. 여기서 기준 값과 비교되는 블록 에너지 값은 블록 내부의 엔트리들의 1 norm, 2 norm, 3 norm등의 값이 될 수도 있다. 기준 값

는 반복 스테이지마다 바뀔 수 있다. 1430 단계에서 이전 반복 스테이지에서 저장되어 있던 서포트 집합

에 추가한다. 즉, 이전 반복 스테이지에서 저장되어 있던

과

의 합집합을 구하고 그 결과를

로 정의한다.

는 오름차순으로 나열될 수 있다. s+1번째 반복 스테이지에서는 s번째 반복 스테이지에서 구한

가 합집합의 입력이 될 수 있으며, 이를 위해 딜레이 유닛(delay unit)을 D로 표시하였다. 1420단계에서의 기준 값

는

과 같이 계산될 수 있으며, 여기서

은 미리 정의된 계수이다.Referring to Figure 14,

is the residual vector at the s-th step, at s=1

and set

Is

to be. Residual vector at step 1410 of the s(≥1)th iteration stage

vector with matched filter applied to

to calculate

matched filter output of

The jth entry of

This is an entry indicating how well the j-th column vector of and is matched.

The number of entries in

is,

n entries are defined as one block. therefore

Is

have blocks. At step 1420

of

(

) the power of the entries in the block

The square root of the sum of the values divided by

set of larger blocks

to output Here, the block energy value compared with the reference value may be a value of 1 norm, 2 norm, 3 norm, etc. of entries in the block. reference value

may change for each iteration stage. At step 1430, the set of supports stored in the previous iteration stage

add to i.e. the previous iteration stage

class

Find the union of

is defined as

may be listed in ascending order. In the s+1th iteration stage, the value obtained in the sth iteration stage is

may be an input of the union, and for this purpose, a delay unit is denoted as D. Reference value in step 1420

Is

can be calculated as, where

is a predefined coefficient.

의 열 벡터 중에

의 블록 엔트리에 해당하는 열 벡터 만을 모은 부분행렬이

로 정의되며, s번째 반복 스테이지에서의 시스템 행렬

및 수신 신호 벡터

를 입력 벡터로 하여 제로 포싱 수신기를 통하여 추정된 베이시스 계수 벡터

가 출력 벡터로서 출력된다.At step 1440

Among the column vectors of

A submatrix that collects only the column vectors corresponding to the block entries of

is defined as, the system matrix at the s-th iteration stage

and received signal vector

The basis coefficient vector estimated through the zero-forcing receiver with

is output as an output vector.

로부터 추정된 베이시스 계수 벡터

가 기여하는 값을 빼기 위해

를 간섭 벡터로 정의하여 간섭 벡터를 형성한다. 수신 신호

에서

를 빼준 값을

로 정의하며, 이를 위해 딜레이 유닛이 이용될 수 있다.In step 1450, the received signal

Basis coefficient vector estimated from

to subtract the value contributed by

is defined as an interference vector to form an interference vector. receive signal

at

minus the value

, and for this purpose, a delay unit may be used.

의 2 norm

이 기준값보다 작아질 때, 또는

의 블록 엔트리 중 절대값이 가장 큰 블록 엔트리가 기준 값보다 작아질 때 멈출 수 있다.This repetition operation determines the maximum number of repetitions and when the number of repetitions is reached,

of 2 norm

When it becomes less than this reference value, or

It can be stopped when the block entry with the largest absolute value among the block entries of is smaller than the reference value.

는 서포트 집합

이 갖는 인덱스에서는 0이 아닌 값을 갖게되며, 서포트 집합

의 여집합이 갖는 인덱스에서는 0인 값을 갖게된다.Basis coefficient vector estimated using block StOMP algorithm

is the support set

In this index, it has a non-zero value, and the support set

The index of the complement set has a value of 0.

로부터 CIR 추정값과 CFR 추정 값을 구할 수 있다.Estimated using Equations 52 to 63

CIR estimates and CFR estimates can be obtained from

15 is a flowchart 1500 illustrating an operation of the receiving terminal 120 performing block stOMP to which BEM is applied according to various embodiments of the present disclosure.

Referring to FIG. 15 , in step 1501, the receiver 120 calculates a vector to which a matched filter is applied to the residual vector. Each entry of the vector to which the matched filter is applied corresponds to information on how much the residual signal matches the system matrix.

In step 1503, the receiver 120 adds a set of entries having a block energy value greater than a reference value among the vector entry blocks calculated in step 1501 to the support set. Here, a block is defined as entries in which all entries have zero or non-zero values. The newly calculated block index of support can be added to the support set by unioning the set consisting of the index of the support block, which is a non-zero entry in the channel vector, which is a sparse vector, with the support set of the previous iteration stage.

In step 1505, the receiver 120 estimates the BEM coefficients by using the partial matrix of the system matrix consisting only of column vectors corresponding to the entries of the support set determined in step 1503. Estimation of the BEM coefficient may be performed by zero forcing. Here, the BEM coefficient means the coefficient of the basis of the channel vector reconstructed with the BEM basis.

In step 1507, the receiving end 120 updates the residual vector. The receiver 120 forms an interference vector to subtract a value contributed by the estimated BEM coefficient, defines a value obtained by subtracting the interference vector from the received signal as a new residual vector, and updates the residual vector.

In the present disclosure, although LP having a polynomial form is used in BEM, it is not limited thereto, and Taylor polynomial, Prorate spheroidal sequence, complex exponential, oversample StOMP and block StOMP can be combined using various BEMs such as oversampled complex exponential.

16A and 16B show a graph 1600 for BLER performance according to an IDFT-based channel estimation according to a modulation and coding scheme (MCS) and a channel estimation scheme using a StOMP algorithm.

및

이며, 채널의 다중경로는 각각 독립이며, 영-평균 정규(zero-mean normal) 분포를 따르고, 도플러 주파수는 300Hz이다. 위 분포를 따르도록 구현된 채널은 총 앙상블 파워가 1이 되도록 스케일된다. 시스템 대역폭은 10MHz이고, 변조방식은 QPSK(quadrature phase shift keying)이며, 수신단에게 스케쥴된 PRB 개수는 50개이며, SISO이다. 전송 블록 크기는 1384 비트이며, 50개의 PRB에 포함될 수 있는 비트는 15000비트이므로 유효 채널 코드 레이트(effective channel code rate)는 1384/15000 = 0.0923이며, 채널 디코더는 맥스 로그-맵(max log-map) 방식으로 동작한다. 터보 디코더(turbo decoder)의 반복 동작 회수는 8회이며, 4000 서브프레임 구간 동안의 채널을 가지고 실험을 하였다. 도 16a의 BLER 곡선을 살펴보면 StOMP 방식이 BLER = 0.1에서 최적 채널 추정기인 LMMSE 방식에 비해 0.65dB의 성능 열화를 보이는 것을 알 수 있다. Genie는 채널을 알고 있을 때의 BLER 곡선을 의미한다. 즉, 낮은 반송파 대 잡음비(CNR)에서 동작하는 MCS인 경우에는 채널을 일정한(static) 것으로 가정하고 StOMP를 이용해도 최적 성능에 근접함을 알 수 있다.16A shows a block error rate (BLER) performance simulation result 1600 when MCS is 0. FIG. The horizontal axis represents an average carrier to noise ratio (CNR) in units of decibels (dB), and the vertical axis represents BLER. The delay and power profiles of the simulated channels are each

and

, and the channel multipath is independent, follows a zero-mean normal distribution, and the Doppler frequency is 300 Hz. A channel implemented to follow the above distribution is scaled so that the total ensemble power becomes 1. The system bandwidth is 10 MHz, the modulation method is QPSK (quadrature phase shift keying), the number of PRBs scheduled to the receiving end is 50, and SISO. The transport block size is 1384 bits, and the number of bits that can be included in 50 PRBs is 15000 bits, so the effective channel code rate is 1384/15000 = 0.0923, and the channel decoder is a max log-map ) works in this way. The number of repetitions of the turbo decoder is 8, and an experiment was conducted with channels for 4000 subframes. Looking at the BLER curve of FIG. 16A , it can be seen that the StOMP method exhibits performance degradation of 0.65 dB compared to the LMMSE method, which is an optimal channel estimator, at BLER = 0.1. Genie means the BLER curve when the channel is known. That is, in the case of an MCS operating at a low carrier-to-noise ratio (CNR), it can be seen that the optimal performance is close to the optimal performance even if the channel is assumed to be static and StOMP is used.

16B shows a BLER performance simulation result 1610 when MCS is 21. FIG. The horizontal axis represents the average CNR in dB, and the vertical axis represents BLER. The experimental environment is the same as in the case of FIG. 16A except for the MCS. Since the transport block size is 21384 bits and the number of bits that can be included in 50 PRBs is 45000 bits, the effective channel code rate is 21384/45000 = 0.4752. Looking at the BLER curve of FIG. 16B , it can be seen that BLER < 0.01 is possible in the LMMSE method and the Genie method, which are the optimal channel estimators. In contrast, IDFT-based channel estimation and channel estimation using the StOMP algorithm show an error floor in a high CNR region. can

17A and 17B show graphs for BLER performance according to a channel estimation technique using StOMP combining IDFT, StOMP algorithm, and LP BEM.

FIG. 17A shows a BLER evaluation simulation result 1700 of channel estimation using StOMP combining IDFT, StOMP algorithm, and LP BEM when MCS is 0. FIG. The horizontal axis represents the average CNR in dB, and the vertical axis represents BLER. The experimental environment is the same as in the case of FIG. 16A. Looking at the BLER curve, it can be seen that the StOMP method, StOMP using the basis of 2 LPs, and StOMP using the basis of 3 LPs all show 0.65dB performance degradation at BLER = 0.1 compared to the LMMSE method, which is the optimal estimator. That is, in the case of MCS 0 operating at low CNR, it can be seen that the StOMP combined with the LP BEM exhibits the same performance as the non-StOMP.

FIG. 17B shows a BLER evaluation simulation result 1710 of channel estimation using StOMP combining IDFT, StOMP algorithm, and LP BEM when MCS is 21. FIG. The horizontal axis represents the average CNR in dB, and the vertical axis represents BLER. The experimental environment is the same as in the case of FIG. 16B. The channel estimation method using StOMP shows an error floor phenomenon in the high CNR region, whereas the channel estimation method combining the basis of LP and StOMP increases the number of basis to 2 and 3, and BLER performance approaches the optimal estimator. can be known Adaptive BEM selection can be applied to select the optimal number of BEM basis according to the Doppler frequency and the instantaneous SNR.

18A and 18B show graphs of BLER performance according to a channel estimation technique using block StOMP combining IDFT, StOMP algorithm, and LP BEM. 18A shows a BLER evaluation simulation result 1800 of channel estimation using block StOMP combining IDFT, StOMP algorithm, and LP BEM when MCS is 0. FIG. The horizontal axis represents the average CNR in dB, and the vertical axis represents BLER. The experimental environment is the same as in the case of FIGS. 16A and 17A. Looking at the BLER curve, both StOMP methods show 0.65 dB performance degradation compared to the LLMSE method, which is the optimal channel estimator, at BLER = 0.1. it can be seen that Adaptive BEM selection can be applied to select the optimal number of BEM basis according to the Doppler frequency and the instantaneous SNR.

18B shows a BLER evaluation simulation result 1810 of channel estimation using block StOMP combining IDFT, StOMP algorithm, and LP BEM when MCS is 21. FIG. The horizontal axis represents the average CNR in dB, and the vertical axis represents BLER. The experimental environment is the same as in the case of FIGS. 16B and 17B. The channel estimation method using StOMP shows an error floor phenomenon in the high CNR region, whereas the channel estimation method using the basis and block StOMP of LP increases the number of basis to two or three, and the BLER performance approaches the optimal channel estimator and It can be seen that there is Adaptive BEM selection can be applied to select the optimal number of BEM basis according to the Doppler frequency and the instantaneous SNR.

Hereinafter, a channel estimation technique in a MIMO system or a multi-antenna system will be described. The following descriptions will be made on the assumption that the number of antennas at the transmitting end is four, but this is only an assumption for description. The present disclosure does not exclude embodiments for different numbers of antennas.

19A to 19C show reference signal patterns according to the number of transmit antennas in a wireless communication system. Specifically, FIG. 19A shows a CRS pattern when the number of transmit antennas is one, FIG. 19B shows a CRS pattern when the number of transmit antennas is two, and FIG. 19C shows a CRS pattern when the number of transmit antennas is four. .

A grid 1910 of FIG. 19A shows a CRS pattern when the number of transmit antennas is one. CRS R0 located at RE indicated by hatching in FIG. 19A is used for channel estimation.

19B shows a CRS pattern when the number of transmit antennas is two. The grid 1920 of FIG. 19B shows the CRS pattern for antenna 0, and CRS R ₀ located in the hatched RE is used for channel estimation. In addition, the grid 1922 of FIG. 19B shows the CRS pattern for antenna No. 1, and CRS R ₁ located in the hatched RE is used for channel estimation.

19C shows a CRS pattern when the number of transmit antennas is four. The grid 1930 of FIG. 19C represents the CRS pattern for antenna 0, and the CRS R ₀ located at the RE indicated by hatching in the grid 1930 is used for channel estimation. In addition, the grid 1932 of FIG. 19C shows the CRS pattern for antenna No. 1, and CRS R ₁ located in the hatched RE is used for channel estimation. The grid 1934 of FIG. 19C shows the CRS pattern for antenna #2, and CRS R ₂ located in the RE indicated by hatching in 1934 is used for channel estimation. In addition, grid 1936 of FIG. 19C shows the CRS pattern for antenna 3, and CRS R ₃ located in the hatched RE is used for channel estimation.

에서

번째 송신 안테나(

)로부터

번째 수신 안테나(

) 사이의

번째 채널 탭 값인

은 이하의 수학식 64와 같이 표현되면,

는 송신 안테나의 총 개수,

은 수신 안테나의 총 개수이다.Timing of the Sampled Receive Signal at the ADC

at

second transmit antenna (

)from

the second receiving antenna (

) between

The second channel tap value

is expressed as Equation 64 below,

is the total number of transmit antennas,

is the total number of receive antennas.

는 송신 안테나의 번호,

는 수신 안테나의 번호,

는 희소한 무선 채널의 다중 경로(multipath)의 개수,

는 칩 길이를 나타낸다.

는 수신 신호의 타이밍

에서

번째 송신 안테나와

번째 수신 안테나 사이의

번째 경로의 계수를 나타내며,

가

번째 경로의 딜레이를 나타내고,

인 경우,

와

는 각각

과

으로 정의된다.here,

is the number of the transmitting antenna,

is the number of the receiving antenna,

is the number of multipaths of the sparse radio channel,

represents the chip length.

is the timing of the received signal

at

second transmit antenna

between the second receiving antenna

represents the coefficient of the th path,

go

represents the delay of the th path,

If ,

Wow

is each

class

is defined as

에서 송신 필터 및 수신 필터에 의한 합성 필터를

라고 할 때,

는

의 길이를 갖는다고 가정한다.

은 지연 확산 값으로서,

으로 표현된다.

는 크로네커 델타이다. ADC 샘플 공간에서 바라본 타이밍

에서

번째 송신 안테나로부터

번째 수신 안테나 사이의

번째 채널 탭

에

번째 경로가 기여하는 성분은

중에서

일 때,

이다. 수학식 64에서

는

인 경우, 이하의 수학식 65를 만족하게 하는 상수이다.hour

Synthesis filter by transmit filter and receive filter in

when said,

Is

Assume that it has a length of

is the delay spread value,

is expressed as

is the Kronecker delta. Timing from ADC Sample Space

at

from the second transmit antenna

between the second receiving antenna

second channel tab

to

The component that the second pathway contributes is

Between

when,

to be. in Equation 64

Is

In the case of , it is a constant that satisfies Equation 65 below.

는 평균 또는 기대 값이고,

는 ADC 샘플 공간에서 바라본 타이밍

에서

번째 송신 안테나로부터

번째 수신 안테나 사이의

번째 채널 탭,

은 지연 확산 값이다.From Equation 65

is the mean or expected value,

is the timing as seen in ADC sample space

at

from the second transmit antenna

between the second receiving antenna

second channel tab,

is the delay spread value.

은

의 선형 중첩으로 표현되므로, 벡터

와

을

과

로 정의할 때, 위 첨자

는 전치를 의미하고,

와

의 관계는 이하의 수학식 66과 같이 표현할 수 있다.in Equation 64

silver

Since it is expressed as a linear superposition of

Wow

second

class

When defined as , superscript

means transpose,

Wow

The relationship between can be expressed as in Equation 66 below.

와

는

과

로 정의되며, 위 첨자

는 전치를 의미하고,

는 ADC 샘플 공간에서 바라본 타이밍

에서

번째 송신 안테나로부터

번째 수신 안테나 사이의

번째 채널 탭,

는 수신 신호의 타이밍

에서

번째 송신 안테나로부터

번째 수신 안테나 사이의

번째 경로의 계수,

은 누출 행렬로 지칭되며, 수학식 66을 만족하게 하는 행렬이다.

Wow

Is

class

is defined as a superscript

means transpose,

is the timing as seen in ADC sample space

at

from the second transmit antenna

between the second receiving antenna

second channel tab,

is the timing of the received signal

at

from the second transmit antenna

between the second receiving antenna

coefficient of the second path,

is referred to as a leak matrix, and is a matrix satisfying Equation 66.

번째 송신 안테나로부터

번째 수신 안테나로 전송된 서브프레임

의 CRS가 위치한 OFDM 심볼

(

)의 RE

에서의 수신 신호를 생각할 수 있다. 송신 안테나 별로 직교하게 배치된 CRS의 값과 위치는 기지국과 UE가 서로 알고 있으므로,

번째 수신 안테나의 수신 신호에서

번째 송신 안테나와 연관된 CRS 값을 나눈 신호

는 이하의 수학식 67과 같이 표현될 수 있다.

from the second transmit antenna

Subframe transmitted to the second receive antenna

OFDM symbol where CRS of

(

) of RE

We can think of the received signal from Since the base station and the UE know the value and location of the CRS orthogonally arranged for each transmit antenna,

In the received signal of the second receiving antenna

A signal obtained by dividing the CRS value associated with the second transmit antenna

can be expressed as in Equation 67 below.

는

번째 수신 안테나의 수신 신호에서

번째 송신 안테나와 연관된 CRS 값을 나눈 신호,

는

번째 수신 안테나의 신호에서 CRS가 위치한

번째 OFDM 심볼의 RE

에서의 잡음 신호이다. 잡음 신호

의 분산을

으로 정의한다. 한 개의 OFDM 심볼 구간 동안 채널은 변화하지 않는다고 가정하면,

은 OFDM 심볼

마다

을 샘플링 한 값이다. 도 19a 내지 도 19c에서

은 0번 송신 안테나와 1번 송신 안테나를 위한 CRS가 위치한 OFDM 심볼이며,

은 2번 송신 안테나와 3번 송신 안테나를 위한 CRS가 위치한 OFDM 심볼이다. OFDM 심볼

에서 CRS가 차지하는 전체 RE의 개수를

라고 하고,

을 수신 안테나의 OFDM 심볼

에서

번째 송신 안테나와 연관된 CRS가 위치한 주파수 영역 잡음을 오름차순으로 나열한 벡터라 하고,

을

번째 수신 안테나의 OFDM 심볼

에서

번째 송신 안테나와 연관된 CRS가 위치한 주파수 영역 수신 신호를 오름차순으로 나열한 벡터로 정의한다. 여기에서 수신기는 정확한

을 알 수 없으므로, 통상 채널 지연 확산이 최대 CP 길이

라 가정하고, 송신 필터 및 수신 필터에 의한 추가 지연 확산 값을 고려해서, 지역 확산 값

은 아래의 수학식 68과 같이 표현될 수 있다.here,

Is

In the received signal of the second receiving antenna

A signal obtained by dividing the CRS value associated with the second transmit antenna,

Is

In the signal of the second receiving antenna, the CRS is located

RE of the th OFDM symbol

is the noise signal in noise signal

the dispersion of

to be defined as Assuming that the channel does not change during one OFDM symbol period,

is the OFDM symbol

every

is the sampled value. 19a to 19c

is an OFDM symbol in which CRSs for transmit antenna 0 and transmit antenna 1 are located,

is an OFDM symbol in which CRSs for the second transmit antenna and the third transmit antenna are located. OFDM symbol

The total number of REs occupied by CRS in

say,

is the OFDM symbol of the receiving antenna

at

Let the vector listing the frequency domain noise in which the CRS associated with the th transmit antenna is located in ascending order,

second

OFDM symbol of the second receive antenna

at

It is defined as a vector in which the frequency domain reception signal in which the CRS associated with the th transmit antenna is located is arranged in ascending order. Here the receiver is

is unknown, so usually the channel delay spread is the maximum CP length.

, and considering the additional delay spread by the transmit filter and the receive filter, the local spread value

can be expressed as in Equation 68 below.

은 지연 확산 값,

은 최대 CP 길이,

는 송신 및 수신 필터에 의한 추가 지연 확산값이다.

is the delay spread value,

is the maximum CP length,

is the value of the additional delay spread by the transmit and receive filters.

라고 할 때, 수신 신호 벡터는 아래의 수학식 69와 같이 정의된다.CIR vector

, the received signal vector is defined as in Equation 69 below.

번째 행과

번째 열

의 엔트리를

로 갖는 행렬을

라 할 때,

은 OFDM 심볼

에서 에서

번째 송신 안테나와 연관된 CRS가 위치한 RE의 서브캐리어 인덱스에 해당하는

의 행 벡터들과 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이다. 그러므로

의 크기는

이다. here,

second row and

second column

entry of

a matrix with

When you say

is the OFDM symbol

in from

Corresponding to the subcarrier index of the RE in which the CRS associated with the th transmit antenna is located

row vectors and column vectors 0 to column vectors of

Up to is a submatrix of column vectors. therefore

the size of

to be.

은 다음의 수학식 70과 수학식 71과 같이 표현될 수 있다.19A to 19C, CRSs for transmit antenna 0 and transmit antenna 1 are located on four OFDM symbols, and the subcarrier index of RE occupied by the CRS is considered. In addition, when the CRS for the second transmit antenna and the third transmit antenna are located on two OFDM symbols, and the index of the RE occupied by the CRS is considered, a partial matrix

can be expressed as in Equations 70 and 71 below.

번째 행과

번째 열

의 엔트리를

로 갖는 행렬을

라 할 때,

은 OFDM 심볼

에서 에서

번째 송신 안테나와 연관된 CRS가 위치한 RE의 서브캐리어 인덱스에 해당하는

의 행 벡터들과 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이다.here,

second row and

second column

entry of

a matrix with

When you say

is the OFDM symbol

in from

Corresponding to the subcarrier index of the RE in which the CRS associated with the th transmit antenna is located

row vectors and column vectors 0 to column vectors of

Up to is a submatrix of column vectors.

번째 행과

번째 열

의 엔트리를

로 갖는 행렬을

라 할 때,

은 OFDM 심볼

에서 에서

번째 송신 안테나와 연관된 CRS가 위치한 RE의 서브캐리어 인덱스에 해당하는

의 행 벡터들과 열 벡터 0부터 열 벡터

까지 열 벡터들로 이루어진 부분행렬이다.here,

second row and

second column

entry of

a matrix with

When you say

is the OFDM symbol

in from

Corresponding to the subcarrier index of the RE in which the CRS associated with the th transmit antenna is located

row vectors and column vectors 0 to column vectors of

Up to is a submatrix of column vectors.

20A to 20C show reference signal patterns used for each OFDM symbol area when the number of transmit antennas is four. A line protruding at the top in FIGS. 20A to 20C indicates a subframe boundary.

에 위치해 있다고 하면, 0번 송신 안테나와 연관된 CRS가 위치한 수신 RE가 채널 추정을 위해 이용되며, 서브프레임과 OFDM 심볼 인덱스는

이다. 도 20a의 그리드 2012에서 음영 표시된 영역의 3개의 OFDM 심볼들이 서브프레임

에 위치해 있다고 하면, 0번 송신 안테나와 연관된 CRS가 위치한 수신 RE가 채널 추정을 위해서 이용되며, 서브프레임과 OFDM 심볼 인덱스는

이다. 도 20a의 그리드 2014의 음영 표시된 영역의 4개의 OFDM 심볼들이 서브프레임

에 위치해 있다고 하면, 0번 송신 안테나와 연관된 CRS가 위치한 수신 RE가 채널 추정을 위해서 이용되며, 서브프레임과 OFDM 심볼 인덱스는

이다. 도 20a의 그리드 2016의 음영 표시된 영역의 3개의 OFDM 심볼들이 서브프레임

에 위치해 있다고 하면, 0번 송신 안테나와 연관된 CRS가 위치한 수신 RE가 채널 추정을 위해서 이용되며, 서브프레임과 OFDM 심볼 인덱스는

이다. 20A shows a CRS pattern used for each OFDM symbol area with respect to antenna 0. Referring to FIG. Looking at the CRS related to the 0 transmit antenna shown in FIG. 20A, 4 OFDM symbols in the shaded area in the grid 2010 of FIG. 20A are subframes.

Assuming that it is located in , the received RE in which the CRS associated with transmit antenna 0 is located is used for channel estimation, and the subframe and OFDM symbol index are

to be. In the grid 2012 of FIG. 20A, three OFDM symbols in the shaded area are subframes

Assuming that it is located in , the received RE in which the CRS associated with transmit antenna 0 is located is used for channel estimation, and the subframe and OFDM symbol index are

to be. 4 OFDM symbols in the shaded area of the grid 2014 of FIG. 20A are subframes

Assuming that it is located in , the received RE in which the CRS associated with transmit antenna 0 is located is used for channel estimation, and the subframe and OFDM symbol index are

to be. 3 OFDM symbols in the shaded area of the grid 2016 of FIG. 20A are subframes

Assuming that it is located in , the received RE in which the CRS associated with transmit antenna 0 is located is used for channel estimation, and the subframe and OFDM symbol index are

to be.

20B shows a CRS pattern used for each OFDM symbol area with respect to antenna No. 1; Channel estimation using the CRS associated with the No. 1 transmit antenna shown in FIG. 20B may be understood similarly to FIG. 20A.

에 위치해 있다고 하면, 2번 송신 안테나와 연관된 CRS가 위치한 수신 RE가 채널 추정을 위해서 이용되며, 서브프레임과 OFDM 심볼 인덱스는

이다. 도 20c의 그리드 2030에서 음영 및 빗금 영역의 7개의 OFDM 심볼들이 서브프레임

에 위치해 있다고 하면, 2번 송신 안테나와 연관된 CRS가 위치한 수신 RE가 채널 추정을 위해서 이용되며, 서브프레임과 OFDM 심볼 인덱스는

이다.20C shows a CRS pattern used for each OFDM symbol area with respect to antenna #2. In the grid 2030 of FIG. 20c, 7 OFDM symbols in shaded and hatched areas are subframes

If it is located in , the received RE in which the CRS associated with the second transmit antenna is located is used for channel estimation, and the subframe and the OFDM symbol index are

to be. In grid 2030 of FIG. 20c, 7 OFDM symbols in shaded and hatched areas are subframes

If it is located in , the received RE in which the CRS associated with the second transmit antenna is located is used for channel estimation, and the subframe and the OFDM symbol index are

to be.

20D shows a CRS pattern used for each OFDM symbol area with respect to antenna #3. Channel estimation using CRS associated with transmit antenna No. 3 for grids 2240 and 2242 of FIG. 20D may be understood similarly to FIG. 20C.

번째 서브프레임 이전에 수신한 CRS의 개수를 늘리는 경우까지 포함하며, 이러한 경우까지 포함하는 실시 예에도 적용될 수 있다.20A to 20D exemplify the case of observing the RE in which the CRS is located for channel estimation by allowing up to 8 OFDM symbol delays based on the first OFDM symbol in the shaded or hatched area. However, the scope of the present disclosure allows up to a greater symbol delay than the embodiment shown in FIGS. 20A to 20D or

It includes a case in which the number of CRSs received before the th subframe is increased, and may be applied to an embodiment including even such a case.

번째 수신 안테나의 수신 신호에서

번째 송신 안테나와 연관된 CRS 값을 나눈 신호를 RE 인덱스 순서대로 나열한 신호를 벡터

로 정의하면, 이하의 수학식 72와 같이 표현된다.For channel estimation in the shaded or hatched area in FIGS. 20A to 20D

In the received signal of the second receiving antenna

A vector signal in which the signal obtained by dividing the CRS value associated with the th transmit antenna is listed in the order of the RE index

When defined as , it is expressed as in Equation 72 below.

는

번째 수신 안테나의 수신 신호에서

번째 송신 안테나와 연관된 CRS 값을 나눈 신호를 RE 인덱스 순서대로 나열한 신호 벡터,

는 시스템

Is

In the received signal of the second receiving antenna

A signal vector listing the signal obtained by dividing the CRS value associated with the th transmit antenna in the order of the RE index,

is the system

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 신호가 겪는 채널 벡터,

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 잡음 신호 벡터이다.procession,

Is

received by the second receiving antenna

The channel vector experienced by the signal from the th transmit antenna,

Is

received by the second receiving antenna

The noise signal vector from the th transmit antenna.

는 수학식 73과 같다.System matrix for the shaded areas in grid 2010 and grid 2014 of FIG. 20A using the CRS associated with transmit antenna 0, and shaded areas in grids 2022 and 2026 in FIG. 20B using the CRS associated with transmit antenna 1

is equal to Equation 73.

는 시스템 행렬,

는 단위 행렬(identity matrix),

는

번째 행과

번째 열

의 엔트리를

로 갖는 행렬,

는 크로네커 곱 연산자,

는 OFDM 심볼

에서 CRS가 차지하는 전체 RE의 개수,

은 지연 확산 값이다.here,

is the system matrix,

is the identity matrix,

Is

second row and

second column

entry of

matrix with ,

is the Kronecker product operator,

is the OFDM symbol

the total number of REs occupied by the CRS,

is the delay spread value.

는 수학식 74와 같다.System matrix for the shaded areas in grids 2012 and 2016 of FIG. 20A using the CRS associated with transmit antenna 0, and the shaded areas in grids 2020 and 2024 of FIG. 20B using the CRS associated with transmit antenna 1

is equal to Equation 74.

는 시스템 행렬,

는 단위 행렬(identity matrix),

는

번째 행과

번째 열

의 엔트리를

로 갖는 행렬,

는 크로네커 곱 연산자,

는 OFDM 심볼

에서 CRS가 차지하는 전체 RE의 개수,

은 지연 확산 값이다.here,

is the system matrix,

is the identity matrix,

Is

second row and

second column

entry of

matrix with ,

is the Kronecker product operator,

is the OFDM symbol

the total number of REs occupied by the CRS,

is the delay spread value.

는 수학식 75와 같다.System matrix for shaded and hatched areas in grid 2030 of FIG. 20C using the CRS associated with transmit antenna 2, and shaded and hatched areas of 2042 in FIG. 20D using the CRS associated with transmit antenna 3

is equal to Equation 75.

는 시스템 행렬,

는 단위 행렬(identity matrix),

는

번째 행과

번째 열

의 엔트리를

로 갖는 행렬,

는 크로네커 곱 연산자,

는 OFDM 심볼

에서 CRS가 차지하는 전체 RE의 개수,

은 지연 확산 값이다.here,

is the system matrix,

is the identity matrix,

Is

second row and

second column

entry of

matrix with ,

is the Kronecker product operator,

is the OFDM symbol

the total number of REs occupied by the CRS,

is the delay spread value.

는 수학식 76과 같다.System matrix for shaded and hatched regions in grid 2032 of FIG. 20C using the CRS associated with transmit antenna 2, and shaded and hatched regions of 2040 in FIG. 20D using the CRS associated with transmit antenna 3

is equal to Equation 76.

In the case of the shaded area of the grid 2010 of FIG. 20A using the CRS associated with the No. 0 transmit antenna, it can be expressed as Equations 77 to 79.

In the case of the shaded area of the grid 2012 of FIG. 20A using the CRS associated with transmit antenna 0, it can be expressed as in Equations 80 to 82.

In the case of the shaded area of the grid 2014 of FIG. 20A using the CRS associated with the No. 0 transmit antenna, it can be expressed as Equations 83 to 85.

In the case of the shaded area of the grid 2016 of FIG. 20A using the CRS associated with the No. 0 transmit antenna, it can be expressed as in Equations 86 to 88.

는

번째 수신 안테나의 수신 신호에서

번째 송신 안테나와 연관된 CRS 값을 나눈 신호를 RE 인덱스 순서대로 나열한 신호 벡터,

는 시스템 행렬,

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 신호가 겪는 채널 벡터,

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 잡음 신호 벡터이다.

는

번째 수신 안테나의 OFDM 심볼

에서

번째 송신 안테나와 연관된 CRS가 위치한 주파수 영역 수신 신호를 오름차순으로 나열한 벡터,

는 CIR 벡터,

는 수신 안테나의 OFDM 심볼

에서

번째 송신 안테나와 연관된 CRS가 위치한 주파수 영역 잡음을 오름차순으로 나열한 벡터이다.In Equations 77 to 88,

Is

In the received signal of the second receiving antenna

A signal vector listing the signal obtained by dividing the CRS value associated with the th transmit antenna in the order of the RE index,

is the system matrix,

Is

received by the second receiving antenna

The channel vector experienced by the signal from the th transmit antenna,

Is

received by the second receiving antenna

The noise signal vector from the th transmit antenna.

Is

OFDM symbol of the second receive antenna

at

A vector listing, in ascending order, the frequency domain reception signal in which the CRS associated with the th transmit antenna is located;

is the CIR vector,

is the OFDM symbol of the receiving antenna

at

It is a vector listing the frequency domain noise in which the CRS associated with the th transmit antenna is located in ascending order.

,

, 및

을 정의할 수 있다.In the case of the CRS associated with the No. 1 transmit antenna shown in FIG. 20B, in a method similar to Equations 77 to 88

,

, and

can be defined.

The shaded and hatched regions of the grid 2030 of FIG. 20C using the CRS associated with the second transmit antenna can be expressed as in Equations 89 to 91.

The shaded and hatched regions of the grid 2032 of FIG. 20C using the CRS associated with the second transmit antenna can be expressed as Equations 92 to 94.

는

번째 수신 안테나의 수신 신호에서

번째 송신 안테나와 연관된 CRS 값을 나눈 신호를 RE 인덱스 순서대로 나열한 신호 벡터,

는 시스템 행렬,

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 신호가 겪는 채널 벡터,

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 잡음 신호 벡터이다.

는

번째 수신 안테나의 OFDM 심볼

에서

번째 송신 안테나와 연관된 CRS가 위치한 주파수 영역 수신 신호를 오름차순으로 나열한 벡터,

는 CIR 벡터,

는 수신 안테나의 OFDM 심볼

에서

번째 송신 안테나와 연관된 CRS가 위치한 주파수 영역 잡음을 오름차순으로 나열한 벡터이다.In Equations 89 to 94,

Is

In the received signal of the second receiving antenna

A signal vector listing the signal obtained by dividing the CRS value associated with the th transmit antenna in the order of the RE index,

is the system matrix,

Is

received by the second receiving antenna

The channel vector experienced by the signal from the th transmit antenna,

Is

received by the second receiving antenna

The noise signal vector from the th transmit antenna.

Is

OFDM symbol of the second receive antenna

at

A vector listing, in ascending order, the frequency domain reception signal in which the CRS associated with the th transmit antenna is located;

is the CIR vector,

is the OFDM symbol of the receiving antenna

at

It is a vector listing the frequency domain noise in which the CRS associated with the th transmit antenna is located in ascending order.

,

, 및

을 정의할 수 있다.In the case of the CRS associated with the third transmit antenna, in a method similar to Equations 89 to 94

,

, and

can be defined.

의 RE

에서의 채널 주파수 응답(channel frequency response, CFR)을 아래의 수학식 95와 같이 정의할 수 있다.OFDM symbols within the shaded or hatched areas of FIGS. 20A to 20D

RE of

A channel frequency response (CFR) in can be defined as in Equation 95 below.

는

번째 송신 안테나로부터

번째 수신 안테나 OFDM 심볼

의 RE

에서의 CFR,

은 OFDM 심볼

마다

번째 송신 안테나로부터

번째 수신 안테나 사이의

번째 채널 탭 값인

을 샘플링 한 채널 값,

은 시스템 행렬의 크기이다.

Is

from the second transmit antenna

th receive antenna OFDM symbol

RE of

CFR in

is the OFDM symbol

every

from the second transmit antenna

between the second receiving antenna

The second channel tap value

the channel value from which it was sampled,

is the size of the system matrix.

의 추정치를 LMMSE 방식을 이용하여 구하면 아래의 수학식 96과 같이 표현된다.When using the CRS associated with transmit antennas 0 and 1,

When the estimate of is obtained using the LMMSE method, it is expressed as in Equation 96 below.

는

번째 송신 안테나로부터

번째 수신 안테나 OFDM 심볼

의 RE

에서의 CFR

의 추정치,

는 평균값 이고,

는

번째 수신 안테나의 수신 신호에서

번째 송신 안테나와 연관된 CRS 값을 나눈 신호를 RE 인덱스 순서대로 나열한 신호 벡터,

는 단위 행렬,

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 신호가 겪는 채널 벡터,

는 시스템 행렬,

는 잡음 신호

의 분산,

는 복소 공액 전치 연산자,

는 역행렬 또는 의사 역행렬 연산자이다.here,

Is

from the second transmit antenna

th receive antenna OFDM symbol

RE of

CFR in

estimate of,

is the average value,

Is

In the received signal of the second receiving antenna

A signal vector listing the signal obtained by dividing the CRS value associated with the th transmit antenna in the order of the RE index,

is the identity matrix,

Is

received by the second receiving antenna

The channel vector experienced by the signal from the th transmit antenna,

is the system matrix,

is the noise signal

dispersion of,

is the complex conjugate transposition operator,

is the inverse matrix or pseudo-inverse matrix operator.

의 영이 아닌 엔트리(non-zero entry)의 위치(이하,

의 서포트(support)로 지칭한다)를 수신기가 알고 있는 것을 내포하고 있다. 실제 수신기에서는 기지국의 송신 필터가 어떻게 구성되는지 알 수 없고, 채널 탭의 서포트의 위치를 잡음이 있는 환경에서 정확히 알기 어려우며, 위 수학식 95 및 수학식 96에서

와

같은 채널의 2차 모멘트(second moment) 값을 알아야 구할 수 있으므로, LMMSE 추정기는 실제 수신기 모뎀에서 구현이 매우 어려운 추정기라고 할 수 있다.Equation 95 and Equation 96 are

The position of the non-zero entry of (hereinafter,

(referred to as support of) implies that the receiver knows. In an actual receiver, it is not known how the transmit filter of the base station is configured, and it is difficult to know exactly the position of the channel tap support in a noisy environment, and in Equations 95 and 96 above,

Wow

Since the second moment value of the same channel can be known to obtain it, the LMMSE estimator is an estimator that is very difficult to implement in an actual receiver modem.

가 희소 벡터(sparse vector)이므로 압축 센싱(CS) 기법을 활용하여 서포트 회복(support recovery)를 수행하면, 최적 추정기에 근접한 성능을 얻을 수 있다. 여기서

의 서포트의 개수를 희소성(sparsity)라고 지칭할 수 있다. CS 기법 중 대표적으로 알려진 OMP (orthorgonal matching pursuit) 알고리즘이 있고, 이 방식을 개량한 StOMP(stagewise OMP) 및 CoSaMP(compressive sampling matching pursuit) 알고리즘이 있다. CoSaMP 방식은 희소성을 알아야 하는 단점이 있으며, OMP 보다 상대적으로 강인(robust)하게 동작하는 StOMP 알고리즘을 이용할 수 있다.However, as described above, the vector

Since is a sparse vector, performance close to the optimal estimator can be obtained by performing support recovery using a compression sensing (CS) technique. here

The number of supports of may be referred to as sparsity. Among CS techniques, there is the well-known orthogonal matching pursuit (OMP) algorithm, and there are the StOMP (stagewise OMP) and CoSaMP (compressive sampling matching pursuit) algorithms that have improved this method. The CoSaMP method has a disadvantage in that it is necessary to know the scarcity, and the StOMP algorithm that operates relatively more robustly than the OMP can be used.

A system equation assuming a constant channel is expressed as Equation 97 below.

는

번째 수신 안테나의 수신 신호에서

번째 송신 안테나와 연관된 CRS 값을 나눈 신호를 RE 인덱스 순서대로 나열한 신호 벡터,

는 시스템 행렬,

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 신호가 겪는 채널 벡터,

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 잡음 신호 벡터이다.here,

Is

In the received signal of the second receiving antenna

A signal vector listing the signal obtained by dividing the CRS value associated with the th transmit antenna in the order of the RE index,

is the system matrix,

Is

received by the second receiving antenna

The channel vector experienced by the signal from the th transmit antenna,

Is

received by the second receiving antenna

The noise signal vector from the th transmit antenna.

In the case of the shaded areas of grids 2010 and 2014 of FIG. 20A using the CRS associated with transmit antenna No. 0, and the shaded areas of grids 2022 and 2026 of FIG. 20B using the CRS associated with transmit antenna No. 1, Equation 98 below and can be expressed together.

In the case of the shaded areas of grids 2012 and 2016 of FIG. 20A using the CRS associated with transmit antenna 0, and the shaded areas of grids 2020 and 2024 of FIG. 20B using the CRS associated with transmit antenna No. 1, Equation 99 below and can be expressed together.

In the case of the shaded and hatched area of grid 2030 of FIG. 20C using the CRS associated with the second transmit antenna, and the shaded area of 2042 in FIG. 20D that uses the CRS associated with the third transmit antenna, it will be expressed as Equation 100 below. can

In the case of the shaded and hatched area of grid 2032 of FIG. 20C using the CRS associated with the No. 2 transmit antenna, and the shaded area of 2040 of FIG. 20D that uses the CRS associated with the No. 3 transmit antenna, it will be expressed as in Equation 100 below. can

는 시스템 행렬,

는

번째 행과

번째 열

의 엔트리를

로 갖는 행렬,

는 OFDM 심볼

에서 CRS가 차지하는 전체 RE의 개수,

은 지연 확산 값이다.In Equations 98 to 101,

is the system matrix,

Is

second row and

second column

entry of

matrix with ,

is the OFDM symbol

the total number of REs occupied by the CRS,

is the delay spread value.

는 아래의 수학식 102와 같이 표현된다.In this case, vector

is expressed as in Equation 102 below.

는

번째 수신 안테나가 수신한

번째 송신 안테나로부터의 신호가 겪는 채널 벡터,

는 채널 값,

은 지연 확산 값이다.

Is

received by the second receiving antenna

The channel vector experienced by the signal from the th transmit antenna,

is the channel value,

is the delay spread value.

21A shows the operation of the StOMP algorithm for channel estimation in a MIMO system.

는

번째 반복 단계에서의 잔여 벡터(residual vector)이며,

에서

이고, 셋

은

이다. 이하

번째 반복 스테이지에서의 동작에 대해 설명한다. In Figure 21a,

Is

is the residual vector in the th iteration step,

at

and three

silver

to be. Below

An operation in the second iteration stage will be described.

번째 반복 스테이지에서의 잔여 벡터

에 정합 필터를 적용하여 벡터

를 출력한다.

의 정합 필터 출력

의

번째 엔트리는

와

의

번째 열 벡터(column vector)가 어느 정도 매칭되는지 알려주는 벡터이다.In step 2110, the receiving end 120

Residual vector in the second iteration stage

vector by applying a matched filter to

to output

matched filter output of

of

the second entry

Wow

of

This is a vector indicating how much the second column vector matches.

의 엔트리 중 에너지 값 또는 절대값이 기준 값

보다 큰 엔트리들의 집합

를 출력한다. In step 2120, the receiving end 120

Entry of energy value or absolute value is the reference value

a larger set of entries

to output

과 합집합을 구하고 그 결과를

로 정의하여 출력한다.

는 항상 오름차순으로 나열되어 있다고 가정한다.

번째 반복 스테이지에서는 이전 반복 단계에서 구한

가 합집합의 입력으로 들어가며, 이 과정에서 딜레이 유닛(delay unit)이 이용된다.In step 2130, the receiving end 120 stores the stored data in the previous iteration stage.

Find the union of and

is defined and output.

is always assumed to be in ascending order.

In the second iteration stage, the value obtained in the previous iteration stage is

is entered as the input of the union, and a delay unit is used in this process.

번째 단계에서의 시스템 행렬

와 수신 신호 벡터 를 입력으로 하여 제로 포싱 수신기를 통하여 추정된 CIR 벡터

가 출력 벡터로서 출력된다.In step 2140, the receiving end 120

system matrix in the first step

The CIR vector estimated through the zero-forcing receiver with the received signal vector and

is output as an output vector.

로부터 추청된 CIR 벡터

가 기여하는 값을 빼기 위해

를 간섭 벡터로 정의함으로써 간섭 벡터를 형성한다. 수신 신호

에서

을 빼준 값은

번째 반복 스테이지에서의 잔여 벡터

로 정의되며, 이러한 과정을 위해 딜레이 유닛이 이용될 수 있다.In step 2150, the receiving end 120 receives a signal

CIR vector derived from

to subtract the value contributed by

An interference vector is formed by defining as an interference vector. receive signal

at

The value minus

Residual vector in the second iteration stage

is defined, and a delay unit may be used for this process.

이 기준 값보다 작을 때 멈출 수도 있으며,

의 엔트리 중 절대값이 가장 큰 엔트리가 기준 값보다 작을 때 멈출 수 있다. StOMP 알고리즘을 이용하여 추정된 CIR 벡터

는

집합이 갖는 인덱스에서는 영이 아닌 값을 가지며,

의 여집합이 갖는 인덱스에서는 0을 갖게 된다. 추정된

로부터 아래의 수학식 103과 같이 CFR을 추정할 수 있다.The repetition operation shown in FIG. 21A determines the maximum number of repetitions and stops or stops when the number of repetitions is reached.

It can also stop when it is less than this reference value,

It can be stopped when the entry with the largest absolute value among the entries of is less than the reference value. CIR vector estimated using StOMP algorithm

Is

It has a non-zero value at the index of the set,

It has 0 at the index of the complement set of . estimated

CFR can be estimated from Equation 103 below.

는 추정된 CFR,

은 추정된 채널 값,

은 지연 확산 값이다.

is the estimated CFR,

is the estimated channel value,

is the delay spread value.

21B to 21C show BLER performance according to various channel estimation techniques in a MIMO system.

21B shows a graph showing BLER performance when MCS is 0 in various channel estimation techniques. In Fig. 21B, the horizontal axis indicates the average CNR in dB, and the vertical axis indicates BLER.

와

이며, 채널의 다중 경로들은 각각 독립하고, 영 평균 정규(zero mean normal) 분포를 따르고, 도플러 주파수는 300Hz이다. 위 분포를 따르도록 구현된 채널은 총 전력이 1이 되도록 스케일된다. 안테나 간의 상관(correlation)이 없는 MIMO 채널을 가정한다. 시스템 대역(system bandwidth)는 10Mhz이고, 변조 방식은 QPSK(quadrature phase shift keying)이며, 단말에게 스케줄된 PRB의 개수는 50개이며, 전송 레이어(transmission layer)의 개수는 2이고, 송신 안테나는 4개, 수신 안테나는 2개, 랭크 적은(rank adaptation)이 없는 개 루프 MIMO(open loop MIMO)이며, 전송 블록 크기(transport block size)가 2792 비트(bit)이며, 50 PRB에 포함될 수 있는 비트는 27200 비트이므로 유효 채널 코딩율(effective channel code rate)는 2792/27200=0.1026이며, 채널 디코더(channel decoder)는 맥스 로그-맵(max log-map) 방식으로 동작한다. 송수신의 유효 펄스 형성 필터(effective pulse shaping filter)는 9 탭을 가지는 sinc 함수를 가정한다. 터보 디코더(turbo decoder)의 반복 동작 회수는 8회이며, 2000 서브프레임 구간 동안의 채널을 대상으로 한다. 도 21b에 도시된 BLER 곡선을 살펴보면, StOMP 방식이 BLER=0.1에서 최적 채널 추정기 (LMMSE 방식)에 비해 0.93dB 성능 열화를 보이는 것을 알 수 있다. Genie는 채널을 알고 있을 때의 BLER 곡선을 의미한다.The channel is an ETU channel, and the delay and power profile are each

Wow

, the multiple paths of the channel are each independent, follow a zero mean normal distribution, and the Doppler frequency is 300 Hz. Channels implemented to follow the above distribution are scaled so that the total power is 1. A MIMO channel without correlation between antennas is assumed. The system bandwidth is 10Mhz, the modulation method is quadrature phase shift keying (QPSK), the number of PRBs scheduled to the terminal is 50, the number of transmission layers is 2, and the transmit antenna is 4 There are two receive antennas, and it is an open loop MIMO (open loop MIMO) without rank adaptation, the transport block size is 2792 bits, and the bits that can be included in 50 PRB are Since it is 27200 bits, the effective channel code rate is 2792/27200=0.1026, and the channel decoder operates in a max log-map method. An effective pulse shaping filter of transmission and reception assumes a sinc function having 9 taps. The number of repetitions of the turbo decoder is 8, and the channel during the 2000 subframe period is the target. Looking at the BLER curve shown in FIG. 21B, it can be seen that the StOMP method exhibits 0.93 dB performance degradation compared to the optimal channel estimator (LMMSE method) at BLER=0.1. Genie means the BLER curve when the channel is known.

21C is a graph showing BLER performance when MCS is 22 among various channel estimation techniques. In Fig. 21C, the horizontal axis indicates the average CNR in dB, and the vertical axis indicates BLER. The channel environment is the same as in the case of FIG. 21B except for the MCS. Since the transport block size is 46888 bits and the number of bits that can be included in 50 PRBs is 81600 bits, the effective channel coding rate is 46888/81600=0.5746. Looking at the BLER curve, it can be seen that BLER < 0.01 is possible for the optimal channel estimator (LMMSE method) and Genie method. On the other hand, the channel estimation method using the StOMP algorithm exhibits an error floor phenomenon in a high CNR (carrier to noise ratio) region. can do. Therefore, it is necessary to approximate the BLER performance to the optimal channel estimator using a more advanced channel estimation method.

According to an embodiment of the present disclosure, reference signal time interpolation (RSTI) may be used to describe a channel that changes with time. In the system equation expressed according to an embodiment of the present disclosure, CFR is estimated using a one-dimensional StOMP or a one-dimensional block StOMP method. From the estimated CFR, the CFR estimate of the OFDM symbol without CRS can be obtained.

22A to 22D show reference signal patterns used for each OFDM symbol area when the number of transmit antennas is 4 in a MIMO system. A line protruding from the top of FIGS. 22A to 22D means a subframe boundary.

에 위치한다고 하면, OFDM 심볼 0의 3번과 4번 RE 사이의 검정색 표시된 RE 상의 수신 신호는 서브 프레임

의 OFDM 심볼 11에 위치한 2번 RE 상의 수신 신호를 그 위치의 CRS 값으로 나눈 신호와 서브프레임

의 OFDM 심볼 4에 위치한 6번 RE 상의 수신 신호를 그 위치의 CRS 값으로 나눈 신호를 선형 보간함으로써 획득된다. 또한, OFDM 심볼 0의 3번 RE 위쪽의 검정색 표시된 RE 상의 수신 신호는 서브프레임

의 OFDM 심볼 11에 위치한 1번 RE 상의 수신 신호를 그 위치의 CRS 값으로 나눈 신호와 서브프레임

의 OFDM 심볼 4에 위치한 5번 RE 상의 수신 신호를 그 위치의 CRS 값으로 나눈 신호를 선형 보간함으로써 획득된다. 또한, OFDM 심볼 4의 6번 아래쪽 검정색 표시된 RE 상의 수신 신호는 서브프레임

의 OFDM 심볼 0에 위치한 4번 RE 상의 수신 신호를 그 위치의 CRS 값으로 나눈 신호와 서브프레임

의 OFDM 심볼 7에 위치한 8번 RE상의 수신 신호를 그 위치의 CRS 값으로 나눈 신호를 선형 보간함으로써 획득된다. 또한, OFDM 심볼 4의 5번과 6번 RE 사이의 검정색 표시된 RE 상의 수신 신호는 서브프레임

의 OFDM 심볼 0에 위치한 3번 RE 상의 수신 신호를 그 위치의 CRS 값으로 나눈 신호와 서브프레임

의 OFDM 심볼 7에 위치한 7번 RE 상의 수신 신호를 그 위치의 CRS 값으로 나눈 신호를 선형 보간함으로써 획득된다. 즉, 한 개의 검정색 표시된 RE 상의 수신 신호는 화살표로 연결된 두 개의 RE 상의 신호를 선형 보간함으로써 획득될 수 있다. 22A shows a CRS pattern associated with transmit antenna No. 0; 4 OFDM symbols in the shaded area of the grid 2200 of FIG. 22A are subframes

Assuming that it is located in , the received signal on the black marked RE between the 3rd and 4th REs of OFDM symbol 0 is a subframe

A signal obtained by dividing the received signal on RE 2 located in OFDM symbol 11 of

It is obtained by linear interpolation of a signal obtained by dividing the received signal on RE 6 located in OFDM symbol 4 of the position by the CRS value of that position. In addition, the received signal on the black marked RE above the 3rd RE of OFDM symbol 0 is a subframe

A signal obtained by dividing the received signal on RE 1 located in OFDM symbol 11 of

It is obtained by linear interpolation of a signal obtained by dividing the received signal on RE 5 located in OFDM symbol 4 of the position by the CRS value of that position. In addition, the received signal on the black marked RE at the 6th bottom of OFDM symbol 4 is a subframe

A signal obtained by dividing the received signal on RE 4 located at OFDM symbol 0 of

It is obtained by linear interpolation of the signal obtained by dividing the received signal on RE 8 located in OFDM symbol 7 of the position by the CRS value of that position. In addition, the received signal on the black marked RE between the 5th and 6th REs of OFDM symbol 4 is a subframe

A signal obtained by dividing the received signal on RE 3 located at OFDM symbol 0 of

It is obtained by linear interpolation of the signal obtained by dividing the received signal on RE 7 located in OFDM symbol 7 of the position by the CRS value of that position. That is, a received signal on one RE indicated in black may be obtained by linearly interpolating signals on two REs connected by arrows.

의 OFDM 심볼

의 CRS가 위치한 RE에서의 수신 신호를 CRS 값으로 나눈 신호 그리고 선형 보간에 의해 획득된 신호를 RE 인덱스 순서대로 나열한 신호를 벡터

으로 정의하면, 아래의 수학식 104와 같다.The received signals on the grids 2202, 2204, and 2206 of FIG. 22A and the REs indicated in black in the shaded or hatched areas on FIGS. 22B to 22D may also be obtained in the same manner as described above. Then, the subframe in which the CRS exists

OFDM symbol of

The signal obtained by dividing the received signal at the RE where the CRS of the CRS is located by the CRS value and the signal obtained by linear interpolation are listed in the order of the RE index.

It is defined as Equation 104 below.

의 엔트리는 송신 안테나

, 수신 안테나

, 서브프레임

, OFDM 심볼

에서, 선형 보간되지 않은 수신 신호에 대하여는 그 RE 상의 잡음 신호이며, 선형 보간하여 획득된 수신 신호에 대하여는 보간할 2개의 RE 상의 잡음을 선형 보간한 잡음과 채널이 선형으로 변화하지 않을 때 발생할 수 있는 간섭의 합이다.

는 OFDM 심볼

에서

번째 송신 안테나와 연괸된 CRS와 보간에 의해 생성된 CRS가 위치한 RE들의 인덱스에 해당하는

의 행 벡터들과 열 벡터 0부터 열 벡터

까지 열 벡터들로 구성된 부분 행렬이다. 그러므로

의 크기는

이며, 도 22a 내지 도 22d에서와 같이 보간에 의해 송신 안테나 그리고 OFDM 심볼 인덱스와 무관하게 동일한 RE의 인덱스를 가진다.vector here

The entry in the transmit antenna

, receiving antenna

, subframe

, OFDM symbol

For a received signal that is not linearly interpolated, it is a noise signal on the RE, and for a received signal obtained by linear interpolation, noise obtained by linearly interpolating noise on two REs to be interpolated and a channel that may occur when the channel does not change linearly is the sum of the interference.

is the OFDM symbol

at

The CRS associated with the th transmit antenna and the index of the REs in which the CRS generated by interpolation is located

row vectors and column vectors of 0 to column vectors

It is a submatrix of column vectors up to . therefore

the size of

and has the same RE index irrespective of the transmit antenna and OFDM symbol index by interpolation as in FIGS. 22A to 22D.

23 illustrates the operation of the StOMP algorithm according to an embodiment of the present disclosure in a MIMO system.

번째 반복 스테이지에서의 잔여 벡터

에 정합 필터를 적용하여 벡터

를 출력한다.

의 정합 필터 출력

의

번째 엔트리는

와 시스템 행렬

의

번째 열 벡터(column vector)가 어느 정도 매칭되는지 알려주는 벡터이다.In step 2310, the receiving end 120

Residual vector in the second iteration stage

vector by applying a matched filter to

to output

matched filter output of

of

the second entry

and system matrix

of

This is a vector indicating how much the second column vector matches.

의 엔트리 중 에너지 값 또는 절대 값이 기준 값

보다 큰 엔트리들의 집합

를 출력한다. In step 2320, the receiving end 120

Entry of energy value or absolute value is the reference value

a larger set of entries

to output

과 합집합을 구하고 그 결과를

로 정의하여 출력한다.

는 항상 오름차순으로 나열되어 있다고 가정한다.

번째 반복 스테이지에서는 이전 반복 단계에서 구한

가 합집합의 입력으로 들어가며, 이 과정에서 딜레이 유닛(delay unit)이 이용된다.In step 2330, the receiving end 120 receives the stored data from the previous iteration stage.

Find the union of and

is defined and output.

is always assumed to be in ascending order.

In the second iteration stage, the value obtained in the previous iteration stage is

is entered as the input of the union, and a delay unit is used in this process.

번째 단계에서의 시스템 행렬

와 수신 신호 벡터 를 입력으로 하여 제로 포싱 수신기를 통하여 추정된 CIR 벡터

가 출력 벡터로서 출력된다.In step 2340, the receiving end 120

system matrix in the first step

The CIR vector estimated through the zero-forcing receiver with the received signal vector and

is output as an output vector.

로부터 추청된 CIR 벡터

가 기여하는 값을 빼기 위해

를 간섭 벡터로 정의함으로써 간섭 벡터를 형성한다. 수신 신호

에서

을 빼준 값은

번째 반복 스테이지에서의 잔여 벡터

로 정의되며, 이러한 과정을 위해 딜레이 유닛이 이용될 수 있다.In step 2350, the receiving end 120 receives the signal

CIR vector derived from

to subtract the value contributed by

An interference vector is formed by defining as an interference vector. receive signal

at

The value minus

Residual vector in the second iteration stage

is defined, and a delay unit may be used for this process.

은 아래의 수학식 105와 같이 계산된다.of the StOMP algorithm shown in FIG.

is calculated as in Equation 105 below.

는 미리 정의된 계수,

는

번째 반복 단계에서의 잔여 벡터,

는

의 2-놈(norm),

는 OFDM 심볼

에서 CRS가 차지하는 전체 RE의 개수이다.here,

is the predefined coefficient,

Is

the residual vector in the second iteration,

Is

The 2-norm of

is the OFDM symbol

is the total number of REs occupied by the CRS.

는

집합이 갖는 인덱스에서는 0이 아닌 값을 가지며,

의 여집합이 갖는 인덱스에서는 0의 값을 갖는다. 추정된

로부터 아래의 수학식 106와 같이 CFR을 추정할 수 있다.CIR vector estimated using StOMP algorithm

Is

It has a non-zero value at the index of the set,

The index of the complement set has a value of 0. estimated

CFR can be estimated from Equation 106 below.

는 추정된 CFR,

은 추정된 채널 값,

은 지연 확산 값이다.here,

is the estimated CFR,

is the estimated channel value,

is the delay spread value.

대신 CRS의 패턴에 따라

또는

을 이용할 수 있으며,

는

번째 행과

번째 열

의 엔트리를

로 갖는 행렬이다.When RSTI is not applied, the system matrix is

Instead, according to the pattern of CRS

or

is available,

Is

second row and

second column

entry of

is a matrix with

When there are four transmit antennas, a case in which the RSTI technique is not applied to a CRS associated with any transmit antenna is defined as rsti=[0 0 0 0]. When the RSTI technique is applied to the CRS associated with the No. 0 transmit antenna and the No. 1 transmit antenna and the RSTI technique is not applied to the CRS associated with the No. 0 transmit antenna and the No. 1 transmit antenna, rsti = [1 1 0 0] is defined. do. When the RSTI technique is applied to all CRSs associated with the transmit antenna, rsti = [1 1 1 1] is defined.

와

을 선형 보간하여 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정한다. 도 22a 내지 도 22d에서 0번 송신 안테나와 관련된 CRS를 이용하는 그리드 2202의 음영 표시된 영역의 경우,

와

을 선형 보간함으로써 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정한다. 0번 송신 안테나와 연관된 CRS를 이용하는 그리드 2204의 음영 표신된 영역의 경우,

와

을 선형 보간함으로써 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정한다. 0번 송신 안테나와 연관된 CRS를 아용하는 청색 영역의 경우,

와

을 선형 보간하여 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정한다. 1번 송신 안테나와 연관된 도 22b의 그리드 2210, 2212, 2214, 2216에 도시된 CRS의 경우에도 도 22a에서의 0번 송신 안테나에 대한 경우와 유사한 방법으로 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정할 수 있다.In the case of the shaded area of the grid 2200 using the CRS associated with the No. 0 transmit antenna in FIGS. 22A to 22D ,

Wow

is linearly interpolated to estimate the CFR of an OFDM symbol in which CRS does not exist. In the case of the shaded area of the grid 2202 using the CRS associated with the No. 0 transmit antenna in FIGS. 22A to 22D,

Wow

Estimate the CFR of an OFDM symbol in which CRS does not exist by linear interpolation. For the shaded area of grid 2204 using the CRS associated with transmit antenna 0,

Wow

Estimate the CFR of an OFDM symbol in which CRS does not exist by linear interpolation. In the case of the blue region using the CRS associated with transmit antenna 0,

Wow

is linearly interpolated to estimate the CFR of an OFDM symbol in which CRS does not exist. In the case of the CRS shown in grids 2210, 2212, 2214, and 2216 of FIG. 22B associated with the first transmit antenna, the CFR of the OFDM symbol in which the CRS does not exist is estimated in a similar manner to the case of the 0 transmit antenna in FIG. 22A can do.

와

을 선형 보간하여 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정한다. 2번 송신 안테나와 연관된 CRS를 이용하는 도 22c의 그리드2222의 음영 표시 및 빗금 표시된 영역의 경우,

와

를 선형 보간하여 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정한다. 3번 송신 안테나와 연관된 도 22d의 그리드 2230 및 2232에 도시된 CRS의 경우에도 도 22c에서의 2번 송신 안테나에 대한 경우와 유사한 방법으로 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정할 수 있다.In the case of the shaded and hatched areas of the grid 2220 of FIG. 22c using the CRS associated with the No. 2 transmit antenna,

Wow

is linearly interpolated to estimate the CFR of an OFDM symbol in which CRS does not exist. In the case of the shaded and hatched areas of the grid 2222 of FIG. 22c using the CRS associated with the No. 2 transmit antenna,

Wow

is linearly interpolated to estimate the CFR of an OFDM symbol in which CRS does not exist. In the case of the CRS shown in grids 2230 and 2232 of FIG. 22D associated with the third transmit antenna, the CFR of the OFDM symbol in which the CRS does not exist may be estimated in a similar manner to the case of the second transmit antenna in FIG. 22C.

24 is a flowchart for channel estimation using the RSTI technique according to an embodiment of the present disclosure in a MIMO system.

In step 2405, the receiving end 120 linearly interpolates between the two descrambled signals on REs where the reference signal or CRS is located. The receiving terminal 120 may convert the received signal into a frequency domain by performing FFT on the received signal, and then check the RE in which the CRS is located. Thus, the receiving terminal 120 may obtain a signal for the RE between the REs in which the CRS is located by descrambles a signal on the REs in which the CRS is located and linearly interpolates between the descrambled signals.

In step 2410, the receiver 120 repeatedly estimates the one-dimensional CIR in the time domain through StOMP or block StOMP. The receiver 120 may perform channel estimation in the time domain using a signal in which the CRS is located and a signal obtained through linear interpolation in step 2405 . Specifically, the receiver 120 may estimate the CIR value using the StOMP or block StOMP algorithm.

In step 2415, the receiver 120 estimates CFR by converting the estimated CIR values into a frequency domain. The receiver 120 may obtain CFR, which is a channel estimation value in the frequency domain, by performing FFT on the CIR values estimated in step 2410 .

In step 2420, the receiving end 120 linearly interpolates between the CFR estimates. That is, the receiving terminal 120 may obtain channel values for the remaining REs by performing linear interpolation on REs located between the CFRs obtained in step 2415 on the frequency axis.

에서

의

개

의 모든 엔트리가 모두 0이 아니거나 모두 0이어야 한다는 성질을 이용하는 알고리즘이다. 시스템 수학식은 아래의 수학식 107과 같이 나타낼 수 있다.As another embodiment of the present disclosure, a block StOMP algorithm may be used. The block StOMP algorithm has a given lag.

at

of

dog

It is an algorithm that utilizes the property that all entries of are not all zeros or must be all zeros. The system equation can be expressed as Equation 107 below.

는 수신 신호,

는 시스템 행렬,

는 채널 벡터,

는 잡음 벡터이다.

is the received signal,

is the system matrix,

is the channel vector,

is the noise vector.

In the case of rsti=[0 0 0 0], which does not apply the RSTI technique to any antenna, for OFDM symbols 0 of antennas 0 and 1, and OFDM symbols 1 of antennas 2 and 3, the system matrix is It can be expressed as Equations 108 to 111 as follows.

For OFDM symbols 0 and 4 of antennas 0 and 1, and OFDM symbols 11 of antennas 0 and 1, the system matrix may be expressed as in Equations 112 to 115 below.

For OFDM symbols No. 7 of antennas 0 and 1, and OFDM symbol No. 8 of antennas 2 and 3, the system matrix may be expressed as in Equations 116 to 119 below.

는 수신 신호,

는 시스템 행렬,

는 채널 벡터,

는 잡음 벡터이고,

는

번째 행과

번째 열

의 엔트리를

로 갖는 행렬이다. In Equations 108 to 119,

is the received signal,

is the system matrix,

is the channel vector,

is the noise vector,

Is

second row and

second column

entry of

is a matrix with

는

번째 반복 스테이지에서의 잔여 벡터이며, 최초 반복 (

)에서

이고, 집합

는

이다. 25 illustrates an operation of a block StOMP algorithm according to an embodiment of the present disclosure. Referring to Figure 25,

Is

Residual vector in the second iteration stage, the first iteration (

)at

and set

Is

to be.

번째 반복 스테이지에서의 잔여 벡터

에 정합 필터를 적용하여 벡터

를 출력한다.

의 정합 필터 출력

의

번째 엔트리는

와 시스템 행렬

의

번째 열 벡터가 어느 정도 매칭되는지 알려주는 벡터이다.

의 엔트리의 개수는

이고,

개의 엔트리들을 하나의 블록으로 정의할 수 있다.In step 2510, the receiving end 120

Residual vector in the second iteration stage

vector by applying a matched filter to

to output

matched filter output of

of

the second entry

and system matrix

of

This is a vector indicating how much the second column vector matches.

The number of entries in

ego,

Entries can be defined as one block.

의

(

)번째 블록 내의 엔트리들의 파워의 합이 기준 값

보다 큰 블록들의 집합을

라고 할 수 있다. 다시 말해, 수신단 120은

의 엔트리 중 에너지 값 또는 절대 값이 기준 값

보다 큰 엔트리들의 집합

를 출력한다. 실시 예에 따라, 기준 값과 비교되는 블록 별 파워 값은 블록 내부의 엔트리들의 1-놈(norm), 2-놈, 3-놈 등의 값이 될 수 있다. 또 다른 실시 예에서, 기준값

는 매 반복마다, 일정 주기, 또는 일정 기준에 따라 변경될 수 있다. 일 실시 예에서, 블록 StOMP 알고리즘의 기준 값

는 아래의 수학식 119와 같이 계산될 수 있다.At step 2520,

of

(

) the sum of the powers of the entries in the block is the reference value

a larger set of blocks

it can be said In other words, the receiving end 120

Entry of energy value or absolute value is the reference value

a larger set of entries

to output According to an embodiment, the power value for each block compared with the reference value may be a 1-norm, 2-norm, 3-norm, etc. of entries in the block. In another embodiment, the reference value

may be changed at every iteration, at a predetermined period, or according to a predetermined criterion. In one embodiment, the reference value of the block StOMP algorithm

can be calculated as in Equation 119 below.

는 블록 StOMP 알고리즘의 기준 값,

는 미리 정의된 계수,

는

번째 반복에서의 잔여 벡터,

는 이용되는 CRS의 개수이다.here,

is the reference value of the block StOMP algorithm,

is the predefined coefficient,

Is

the residual vector at the th iteration,

is the number of CRSs used.

과 합집합을 구하고 그 결과를

로 정의하여 출력한다.

는 항상 오름차순으로 나열되어 있다고 가정한다.

번째 반복 스테이지에서는 이전 반복 단계에서 구한

가 합집합의 입력으로 들어가며, 이 과정에서 딜레이 유닛이 이용된다. 예를 들면,

,

이고, RSTI에 의하여 선형 보간이 수행되면

이고, 선형 보간이 수행되지 않으면

이다. In step 2530, the receiver 120 receives the stored data from the previous iteration stage.

Find the union of and

is defined and output.

is always assumed to be in ascending order.

In the second iteration stage, the value obtained in the previous iteration stage is

is input to the union, and a delay unit is used in this process. For example,

,

, and when linear interpolation is performed by RSTI,

, and if linear interpolation is not performed

to be.

번째 단계에서의 시스템 행렬

와 수신 신호 벡터

를 입력으로 하여 제로 포싱 수신기를 통하여 추정된 CIR 벡터

가 출력 벡터로서 출력된다.In step 2540, the receiving end 120

system matrix in the first step

and receive signal vector

CIR vector estimated through a zero-forcing receiver with

is output as an output vector.

로부터 추청된 CIR 벡터

가 기여하는 값을 빼기 위해

를 간섭 벡터로 정의함으로써 간섭 벡터를 형성한다. 수신 신호

에서

을 빼준 값은

번째 반복 스테이지에서의 잔여 벡터

로 정의되며, 이러한 과정을 위해 딜레이 유닛이 이용될 수 있다.In step 2350, the receiving end 120 receives a signal

CIR vector derived from

to subtract the value contributed by

An interference vector is formed by defining as an interference vector. receive signal

at

The value minus

Residual vector in the second iteration stage

is defined, and a delay unit may be used for this process.

이 기준 값보다 작아질 때 멈출 수도 있으며,

의 엔트리 중 절대값이 가장 큰 엔트리가 기준 값보다 작아질 때 멈출 수 있다. 블록 StOMP 알고리즘을 이용하여 추정된 CIR 벡터

는 집합이 갖는 인덱스에서는 0이 아닌 값을 가지며,

의 여집합이 가지는 인덱스에서는 0인 값을 갖게 된다. 이후, 수학식 103을 이용하여 CFR을 추정할 수 있으며, RSTI가 적용된 StOMP 방식과 유사하게 선형을 수행하여 CRS가 존재하지 않는 OFDM 심볼의 CFR을 추정할 수 있다.The above-described operation is repeated a certain number of times, and according to an embodiment, the repeated operation stops when reaching a predefined maximum number of times, or

It may stop when it becomes less than this reference value,

It can be stopped when the entry with the largest absolute value among the entries in is smaller than the reference value. CIR vector estimated using block StOMP algorithm

has a non-zero value at the index of the set,

The index of the complement set has a value of 0. Thereafter, the CFR may be estimated using Equation 103, and the CFR of the OFDM symbol in which the CRS does not exist may be estimated by performing linearization similar to the StOMP method to which the RSTI is applied.

26A to 26B show graphs of BLER performance according to various channel estimation techniques in a MIMO system.

26A shows BLER performance when MCS is 0 in various channel estimation schemes. 26A, the horizontal axis represents the average CNR in dB, and the vertical axis represents BLER. The experimental environment is the same as in the case of FIG. 21B. Looking at the BLER curve, it can be seen that the block StOMP method to which rsti=[1 1 1 1] is applied shows about 0.4dB performance degradation compared to the LMMSE method, which is close to the optimal channel estimator at BLER=0.1. From these results, it can be seen that there is a 0.5dB gain compared to the StOMP method assuming a static channel. In addition, it can be seen that the StOMP method to which rsti = [1 1 1 1] is applied shows performance degradation of about 0.7 dB compared to the optimal channel estimator at BLER = 0.1. Among rsti=[0 0 0 0], rsti=[1 1 0 0], rsti=[1 1 1 1], the best performance is obtained when rsti = [1 1 1 1] in both StOMP and block StOMP methods. Able to know.

26B shows BLER performance when MCS is 22 in various channel estimation schemes. 26B, the horizontal axis represents the average CNR in dB, and the vertical axis represents BLER. The experimental environment is the same as in the case of FIG. 21C. The channel estimation method using StOMP assuming a constant channel exhibits an error floor phenomenon, whereas the StOMP and block StOMP methods applying rsti=[0 0 0 0] and rsti=[1 1 0 0] show BLER < 0.01 is satisfied. That is, in a high Doppler frequency and high CNR region, linear interpolation of a received signal in an RE in which CRS is present may not be suitable.

27 is a flowchart illustrating an adaptive selection of rsti selection of a channel estimation method using StOMP and block StOMP using linear interpolation of CFR according to an embodiment of the present disclosure. The flowchart shown in FIG. 27 may be generally performed using a Doppler detector and an SNR measurer that are well known to those skilled in the art, and the Doppler detector and the SNR measurer may be included in the receiving end. rsti indicates whether the RSTI technique is applied to each transmit antenna, for example, rsti = [0 0 0 0], when the RSTI technique is not applied to any antenna, rsti = [1 1 0 0 ] indicates that the RSTI technique is not applied to the CRSs associated with the No. 0 transmit antenna and the No. 1 transmit antenna, and rsti = [1 1 1 1] indicates a case in which the RSTI method is applied to both the CRSs associated with the transmit antenna.

In step 2705, the receiving end 120 determines whether the Doppler frequency is a low frequency or whether the receiving end 120 is moving at a low speed. If it is determined that the low-speed movement is in progress, the receiver 120 applies rsti=[1 1 1 1] in step 2740.

If it is not determined that the terminal is moving at a low speed, in step 2710, the receiving terminal 120 determines whether the Doppler frequency is an intermediate frequency, that is, whether the receiving terminal 120 is moving at an intermediate speed. When it is determined that the receiving end 120 moves at the intermediate movement speed, the receiving end 120 proceeds to step 2715 and determines whether the instantaneous SNR measured by the slot or subframe rate is greater than a specific reference value T _m,1 . If the instantaneous SNR is greater than T _m,1 , proceed to step 2745 and apply rsti=[1 1 1 1]. If the instantaneous SNR is less than or equal to T _m,1 , go to step 2750 and rsti=[1 1 0 0] is applied.

When the receiving end 120 determines that the Doppler frequency is a high frequency in step 2710, that is, when it is determined that the receiving end 120 is moving at high speed, the receiving end 120 proceeds to step 2720 to determine the instantaneous SNR measured by slot or subframe rate according to a specific standard It is determined whether the value T _h,1 is greater than or not. If the instantaneous SNR is greater than T _h,1 , the receiving end 120 proceeds to step 2755 and applies rsti=[1 1 1 1]. If the instantaneous SNR is less than or equal to T _h,1 , the receiving terminal 120 proceeds to step 2730 to determine whether the instantaneous SNR is greater than a specific reference value T _h,2 . If the instantaneous SNR is greater than T _h,2 , the receiving end 120 proceeds to step 2760 and applies rsti=[1 1 0 0], and if the instantaneous SNR is less than or equal to T _h,2 , the receiving end 120 proceeds to step 2765 Go ahead and apply rsti=[0 0 0 0]. According to an embodiment, an average value of SNRs of OFDM symbols occupied by CRS for which channel estimation is used in one subframe may be defined as an instantaneous SNR, and SNRs of OFDM symbols received over several subframes and in which CRS exists It can also be defined using a finite impulse response (FIR) filter or an infinite impulse response (IIR) filter output.

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

Such software may be stored in a computer-readable storage medium. The computer-readable storage medium includes at least one program (software module), at least one program including instructions that, when executed by at least one processor in an electronic device, cause the electronic device to implement the method of the present disclosure save the

Such software may be in the form of a volatile or non-volatile storage device such as ROM (Read Only Memory), or RAM (random access memory), memory chips, device or in the form of memory, such as integrated circuits, or compact disc ROMs (CD-ROMs), Digital Versatile Discs (DVDs), magnetic disks or magnetic tapes ( optical or magnetically readable media, such as magnetic tape).

Storage devices and storage media are embodiments of machine-readable storage means suitable for storing a program or programs comprising instructions that, when executed, implement the embodiments. Embodiments provide a program comprising code for implementing an apparatus or method as claimed in any one of the claims of this specification, and a machine-readable storage medium storing such a program. Furthermore, such programs may be transmitted electronically by any medium, such as a communication signal transmitted over a wired or wireless connection, and embodiments suitably include equivalents.

In the above-described specific embodiments, elements included in the invention are expressed in the singular or plural according to the specific embodiments presented. However, the singular or plural expression is appropriately selected for the situation presented for convenience of description, and the above-described embodiments are not limited to the singular or plural component, and even if the component is expressed in plural, it is composed of a singular or , even a component expressed in a singular may be composed of a plural.

On the other hand, although specific embodiments have been described in the description of the invention, various modifications are possible without departing from the scope of the technical idea contained in the various embodiments. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined by the claims described below as well as the claims and equivalents.

Claims (20)

In the method of operation of the receiving end,
The process of obtaining a basis expansion model (BEM) basis vector (vector) based on the reference signals received from the transmission device (transmission device),
A process of obtaining at least one BEM coefficient corresponding to at least one entry of a channel impulse response (CIR) based on the BEM basis vector, wherein the at least one entry has a non-zero value ,
obtaining the CIR based on the at least one BEM coefficient and the BEM basis vector;
The process of obtaining a channel frequency response (CFR) from the CIR, and
and processing received signals using the CFR.
The method according to claim 1,
The method further comprising: obtaining a second set of channel values of the CIR by performing interpolation on the first set of channel values of the CIR based on the at least one BEM coefficient and the BEM basis vector .
3. The method according to claim 2,
The process of determining at least one entry (entry) having a non-zero value of the CIR,
The method further comprises estimating the at least one BEM coefficient corresponding to the at least one entry having a non-zero value,
The CIR includes the first set of channel values and the second set of channel values.
4. The method according to claim 3,
The process of determining at least one entry having a non-zero value includes:
The process of generating a correlation vector by applying a matched filter to the residual signal, and
determining at least one entry of the correlation vector in which an energy value of the at least one entry of the correlation vector is greater than a reference value;
The residual signal includes at least a portion of the reference signals.
4. The method according to claim 3,
The process of determining at least one entry having a non-zero value includes:
and determining at least one block from among the blocks in which all values of entries in the CIR are non-zero,
The process of estimating the at least one BEM coefficient,
estimating the at least one BEM coefficient corresponding to entries included in the at least one block;
wherein each of the blocks includes one or more entries that are all zero or non-zero.
6. The method of claim 5,
The process of determining the at least one block includes:
A process of generating a correlation vector by applying a matched filter to the residual signal;
determining at least one block in which the energy value of at least one block of the correlation vector is greater than a reference value;
The residual signal includes at least a portion of the reference signals.
The method according to claim 1,
The BEM basis vector, the method comprising a Legendre polynomial (Legendre polynomial). 3. The method according to claim 2,
The method further comprising determining the number of the basis for the BEM basis vector based on at least one of a Doppler frequency of a channel associated with the CIR or a channel quality for the channel.
The method according to claim 1,
The reference signals are transmitted through a plurality of antennas of the transmitting device,
determining a set of antennas to perform linear interpolation among the plurality of antennas based on at least one of a Doppler frequency of a channel associated with the CIR or a channel quality for the channel; and
The method further comprises the step of determining a set of the second channel value of the CIR by using the reference signals for each of a plurality of antennas including antennas included in the set of antennas,
and the set of second channel values for the antenna included in the antenna set is determined from the reference signals and an interpolated signal obtained by linear interpolation between reference signals.
10. The method of claim 9,
determining at least one entry having a non-zero value of the CIR comprising a first set of channel values;
Further comprising the step of determining at least one value corresponding to the at least one entry having a non-zero value,
The CIR includes the first set of channel values and the second set of channel values.
In the device of the receiving end,
a transceiver for receiving reference signals from a transmitting device;
At least one processor operatively connected to the transceiver, the processor comprising:
obtaining a basis expansion model (BEM) basis vector based on the reference signals,
at least one BEM coefficient corresponding to at least one entry of a channel impulse response (CIR) is obtained based on the BEM basis vector, wherein the at least one entry has a non-zero value;
obtaining the CIR based on the at least one BEM coefficient and the BEM basis vector;
Obtaining a channel frequency response (CFR) from the CIR,
An apparatus for processing received signals using the CFR.
The method of claim 11 , wherein the at least one processor comprises:
The apparatus further configured to obtain a second set of channel values of the CIR by performing interpolation on the first set of channel values of the CIR based on the at least one BEM coefficient and the BEM basis vector.
13. The method of claim 12,
The at least one processor,
determining at least one entry having a non-zero value of the CIR;
and estimating the at least one BEM coefficient corresponding to the at least one entry having a non-zero value;
The CIR includes the first set of channel values and the second set of channel values.
14. The method of claim 13,
The at least one processor,
A correlation vector is generated by applying a matched filter to the residual signal,
and determine at least one entry in the correlation vector in which an energy value of the at least one entry in the correlation vector is greater than a reference value;
The residual signal is at least a part of the reference signals.
14. The method of claim 13,
The at least one processor,
Determining at least one block from among the blocks in which the values of the entries in the CIR are not all 0,
and estimating the at least one BEM coefficient corresponding to entries included in the at least one block,
Each of the blocks includes one or more entries that are all zero or non-zero.
16. The method of claim 15,
The at least one processor,
A correlation vector is generated by applying a matched filter to the residual signal,
further configured to determine at least one block in which an energy value of at least one block of the correlation vector is greater than a reference value,
The residual signal includes at least a portion of the reference signals.
12. The method of claim 11,
The BEM basis vector comprises a Legendre polynomial (Legendre polynomial).
13. The method of claim 12,
The at least one processor,
and determine the number of basis for the BEM basis vector based on at least one of a Doppler frequency of a channel associated with the CIR or a channel quality for the channel.
12. The method of claim 11,
The reference signals are transmitted through a plurality of antennas of the transmitting device,
The at least one processor,
determining a set of antennas to perform linear interpolation among the plurality of antennas based on at least one of a Doppler frequency of a channel associated with the CIR or a channel quality for the channel;
The method further comprises the step of determining a set of the second channel value of the CIR by using the reference signals for each of a plurality of antennas including antennas included in the antenna set;
The set of second channel values for the antenna included in the antenna set is determined from an interpolation signal obtained by linear interpolation between reference signals and reference signals.
20. The method of claim 19,
The at least one processor,
determine at least one entry having a non-zero value of the CIR comprising a first set of channel values;
and determine at least one value for the at least one entry having a non-zero value;
The CIR includes the first set of channel values and the second set of channel values.

Images