KR102500154B1 - Low-complexity coherent PCID and RFN detection method - Google Patents

Abstract

The present invention relates to a coherent synchronization signal detection using NSSS in an NB-IoT system. The present invention, by using a symmetric complex conjugate property of a synchronization signal, takes an absolute value of an imaginary part of an objective function, and reduces a complexity by first finding another value and simply finding a remaining value at the end. In addition, because a coherent method is used, the present invention can be used in a condition requiring the channel information. According to comparison results of the complexity through the present invention, by comparing to a conventional coherent method, the present invention allows an advantage of showing half the complexity to be checked. Therefore, the present invention is capable of enabling a burden of a UE to be relieved.

Description

Low-complexity coherent PCID and RFN detection method {Low-complexity coherent PCID and RFN detection method}

The present invention provides a method for estimating physical cell ID (PCID) and radio frame number (RFN) by reducing complexity by using the characteristics of a narrowband secondary synchronization signal (NSSS) signal in a narrowband Internet of Things (NB-IoT) system. it's about

The NB-IoT system estimates the symbol timing offset (STO) and carrier frequency offset (CFO) on the time axis through the narrowband primary synchronization signal (NPSS), passes through Fourier transform (FFT), has several NSSS signals, and PCID Estimate values and RFN information. In order to obtain the PCID value and the RFN value, 4032 possible comparison values must be obtained, which may impose a burden on the user equipment (UE) because it is computationally complex.

The NSSS signal is composed of Zadoff-Chu (ZC) and Hadamard-Walsh sequences, and can be divided into coherent or non-coherent detection depending on whether channel information is used or not. Among them, when coherent detection is used, the complexity does not increase much in the case of a frequency non-selective channel, but the complexity tends to increase in the case of a frequency selective channel.

The present invention relates to a method for obtaining channel characteristics and estimating a PCID and an RFN using characteristics of an NSSS signal when channel information is required in a NB-IoT system.

Using the fact that the signal of NSSS has the characteristic of being a symmetric conjugate complex number, the absolute value of the imaginary part is obtained and the correlation value is obtained, thereby providing a method of reducing complexity when obtaining specific information.

According to the present invention, in the NB-IoT system, a user equipment (UE) can know channel information, a PCID value, and RFN information using an NSSS signal. Therefore, it can be used when channel information is required. In addition, according to the present invention, the load on the UE can be reduced by using the characteristics of the NSSS signal to reduce the overall complexity while making the probability of an error similar to that of other methods.

, (b)

을 보여주는 도면이다.
[도 4] pedestrian 채널에서 기존의 방식들과 발명의 방법에 성능 비교 (a) 완벽한 채널 추정, (b)추정한 채널을 보여주는 도면이다.
[도 5] vehicle 채널에서 기존의 방식들과 발명의 방법에 성능 비교
(a) 완벽한 채널 추정, (b)추정한 채널을 보여주는 도면이다.
[도 7]

에 따른 기존의 방식들과 발명의 방법에 성능비교 (a) SNR=-5 dB, (b)SNR=0 dB를 보여주는 도면이다.[Fig. 1] is a diagram showing a frame structure for a synchronization signal of NB-IoT.
[Fig. 2] is a diagram showing the overall flow of the present invention.
[Figure 3] AWGN, probability of false detection as NSSS signal in flat-fading channel (a)

, (b)

is a drawing showing
[Figure 4] Comparison of performance between existing methods and the method of the present invention in a pedestrian channel is a diagram showing (a) perfect channel estimation and (b) estimated channel.
[Figure 5] Performance comparison between existing methods and the method of the invention in the vehicle channel
(a) perfect channel estimation, (b) a diagram showing the estimated channel.
[Figure 7]

It is a diagram showing performance comparison between the existing methods and the method according to the invention (a) SNR = -5 dB, (b) SNR = 0 dB.

With reference to the accompanying drawings, an embodiment of the present invention will be described in detail so that those skilled in the art can easily practice it.

The present invention relates to a method for estimating a channel information value, a PCID value, and an RFN value required for communication using NSSS, which is a synchronization signal, in a NB-IoT system. To understand this, it is necessary to know the structure of NSSS signals.

(1) Narrowband Secondary Synchronization Signal (NSSS) Acquisition Step

Referring to [Fig. 1], it can be seen the configuration of the overall synchronization signal. The synchronization signal is largely divided into an even radio frame and an odd radio frame, and it can be seen that there is an NSSS signal in the 9th subframe. The mathematical expression of this NSSS signal is shown in Equation 1 below.

는 NSSS의 전체 길이로 132이고,

이며

으로 PCID 값이다.

,

,

이며 PCID를 결정할 수 있는 값이다. 또한

으로, RFN의

값에 의하여 정해진다.

는 128개의 Hadamard-Walsh 시퀀스를 나타낸다.

is 132 as the total length of NSSS,

is

is the PCID value.

,

,

It is a value that can determine the PCID. also

As, RFN's

determined by the value

represents 128 Hadamard-Walsh sequences.

는 11개의 연속적인 심볼 블록을 그룹화 한 값으로 정의한다. 이것을 바탕으로 l번째 신호가 k번째 subcarrier에 있는 것은

로 표현할 수 있다 (

). 그리고

,

,

을 각각

,

,

으로 표현하면 다음과 같은 수식이 된다. 이는 NSSS 신호이다.

is defined as a grouping value of 11 consecutive symbol blocks. Based on this, the lth signal is in the kth subcarrier

can be expressed as (

). and

,

,

respectively

,

,

Expressed as: This is the NSSS signal.

,

,

이다. here

,

,

am.

(2) Channel information estimation step

The channel can be estimated by receiving the signal represented by the above equation at the receiving end and using the received signal in Equation 3.

이고,

는 NSSS 전체의 서브캐리어 집합이며,

는 RFN을 결정 지을 수 있는 값이다.

는 수학식 2에서의 n값이 될 수 있는 값이고,

은 u가 될 수 있는 후보군이다. 또한

는 w가 될 수 있는 값이다. 마지막으로

는 radio frame 안에서 OFDM block number를 가리킨다. Y는 수신신호를 의미하며, 이들로부터 추정한 수힉식 3의 H는 추정한 채널신호이다. l은 신호의 순서, k는 subcarrier의 순서를 나타내는 것으로서,

는 수신된 신호이고,l번째 신호가 k번째 subcarrier에 있는 것은

로 표현할 수 있다. here

ego,

is a set of subcarriers of the entire NSSS,

is a value that can determine the RFN.

Is a value that can be an n value in Equation 2,

is a candidate group that can be u . also

is a value that can be w . finally

indicates the OFDM block number in the radio frame. Y means a received signal, and H in Equation 3 estimated from them is an estimated channel signal. l represents the order of signals, k represents the order of subcarriers,

is the received signal, and the l -th signal is in the k -th subcarrier

can be expressed as

In addition, based on Equation 3, an objective function for obtaining the maximum value among the cross-correlation values of the estimated candidate group and the received signal can be obtained as follows.

이다. 수학식 4로부터 각각의 후보군 (m,n,u,w)과 수신 된 신호의 상관값을 구하게 된다. 앞서 수학식 3에서 설명한 것과 같이,

는 RFN을 결정 지을 수 있는 값이다.

는 수학식 2에서의 n값이 될 수 있는 값이고,

은 u가 될 수 있는 후보군이다. 또한

는 w가 될 수 있는 값이다.

는 NSSS 전체의 서브캐리어 집합이며,

는 radio frame 안에서 OFDM block number이다. 이 때, m은 NSSS가 전송되는 RFN 값을 결정하는 인덱스값, n은 셀 특정 파라미터, u는 순환 쉬프트값, w는 루트 시퀀스 값이다.
l은 신호의 순서, k는 subcarrier의 순서를 나타내는 것으로서,

는 수신된 신호이고,l번째 신호가 k번째 subcarrier에 있는 것은

로 표현할 수 있다.
수학식 4에서, 만약 (a,b,c,d)=(m,n,u,w)라면, 각 채널은 수학식 5와 같이 표현 할 수 있다. here

am. From Equation 4, a correlation value between each candidate group (m, n, u, w) and the received signal is obtained. As described in Equation 3 above,

is a value that can determine the RFN.

Is a value that can be an n value in Equation 2,

is a candidate group that can be u . also

is a value that can be w .

is a set of subcarriers of the entire NSSS,

is the number of OFDM blocks in a radio frame. At this time, m is an index value for determining the RFN value through which the NSSS is transmitted, n is a cell specific parameter, u is a cyclic shift value, and w is a root sequence value.
l represents the order of signals, k represents the order of subcarriers,

is the received signal, and the l -th signal is in the k -th subcarrier

can be expressed as
In Equation 4, if ( a,b,c,d )=( m,n,u,w ), each channel can be expressed as Equation 5.

이고,

는 잔여 carrier frequency offset(CFO)값을 나타내고,

는 잔여 Symbol timing offset(STO) 값을 나타낸다. here

ego,

Represents the residual carrier frequency offset (CFO) value,

represents the remaining Symbol timing offset (STO) value.

delete

(3) Channel information simplification step

가 되도록 ZC 시퀀스가 켤레 쌍인 복소수가 형성되므로, 수학식 4는 다음과 같이 허수부와 실수부를 분리하여 정리할 수 있다.

Since the ZC sequence is a complex number that is a conjugate pair, Equation 4 can be arranged by separating the imaginary part and the real part as follows.

는 다음과 같이 정의할 수 있다.Here, the subscripts I and Q represent the real and imaginary parts of the variable, respectively.

can be defined as:

이다. 이는 오른쪽 항의 부호만 다르다는 것 알 수 있고, 이러한 특성을 이용하여 다음과 같이 목적 함수를 표현하면 검색 공간을 줄일 수 있다. 즉, 아래 수학식 8은, 수학식 6의 허수 부분에 절대값을 취하여 계산의 복잡도를 감소시킨 것이다. and

am. It can be seen that only the sign of the right term is different, and the search space can be reduced by using this characteristic to express the objective function as follows. That is, in Equation 8 below, the complexity of calculation is reduced by taking the absolute value of the imaginary part of Equation 6.

는

의 가설 값이고,

는 절대값에 의하여 제거된다. 수학식 3, 4에서

는 w 의 후보군이었던 것을, 여기에서는

이므로,

로 표현한 것이다. here

Is

is the hypothesis value of

is eliminated by the absolute value. In Equations 3 and 4

was a candidate group for w , where

Because of,

is expressed as

을 구할 수 있다.Based on the above formula, the RFN and PCID detection method of the present invention is divided into two steps, and in the first step, the estimated values of m, n, u, w' in Equation 9 as follows

can be obtained.

(m은 NSSS가 전송되는 RFN 값을 결정하는 인덱스값, n은 셀 특정 파라미터, u는 순환 쉬프트값, 루트 시퀀스 값 w가 될 수 있는 값이

일때,

는 그 후보값으로

를 가지는 루트 시퀀스 값,

는 각각 m,n,u,w' 의 추정 값이다)

(m is an index value that determines the RFN value through which NSSS is transmitted, n is a cell-specific parameter, u is a cyclic shift value, and a value that can be a root sequence value w is

when,

is the candidate value

A root sequence value having

are estimated values of m, n, u, w', respectively)

이기 때문에, 이는 계산을 줄일 수 있다. 이 단계에서 RFN은

과

그리고 ZC root index인

또는

에 의해서 구한다. 이는 수학식 1에서,

으로, RFN의

값에 의하여 정해지며, 수학식 1의 인덱스 추정을 통하여 구해질 수 있다. here

Because of this, it can reduce the calculation. At this stage the RFN is

class

And the ZC root index

or

saved by In Equation 1,

As, RFN's

It is determined by the value and can be obtained through index estimation in Equation 1.

Next, the following formula can be used to obtain the substantially transmitted NSSS.

는 가능한 후보군이다. 여기서는 단지 2가지의 가설 값이 있으므로 다음의 수식으로 표현이 가능하다.here,

is a possible candidate group. Here, since there are only two hypothesis values, it can be expressed in the following formula.

Also, the above expression can be expressed as:

값이 될 것이고, 최종적으로 v는

으로 추정할 수 있다. 이렇게 추정한 v 값은 PCID의 추정값이 된다. 이 때,

는 가능한 후보군이다.

는 각각 n,u, w'의 추정 값이다.

는 radio frame 안에서 OFDM block number를 가리킨다.
<실시 예>according to the above formula

will be the value, and finally v is

can be estimated as The value of v thus estimated becomes the estimated value of the PCID. At this time,

is a possible candidate group.

are estimated values of n, u, and w', respectively.

indicates the OFDM block number in the radio frame.
<Example>

delete

3, 4, 5, and 6 are simulation results using the present invention. The methods compared here are DFT-based PCID and RFN detection (DPRD), low-complexity PCID and RFN detection (LPRD), and sequential PCID and RFN detection (SPRD). 3 is a graph showing the probability of erroneous estimation, and it can be seen that using several NSSSs as expected with AWGN and Flat-fading channels shows a noise attenuation effect, and the performance is similar to that of DPRD, and the performance is much superior to that of LPRD. can confirm that Figure 4 shows the performance in channels of PedA and PedB. If the channel information is known, the coherent method outperforms the noncoherent method regardless of the channel. In particular, in the case of PedB, when non-coherent estimation is performed, it can be seen that an error floor occurs as the SNR increases. 5 is a performance comparison between VehA and VehB, and it can be seen that the performance is much better than that of LPRD and almost similar to that of DPRD. This shows that the present invention is more suitable in various channels. 6 shows that the noncoherent method shows better performance in the PeDA method with the probability of incorrect estimation when the SNR is -5 and the SNR is 0, but the average effect is insignificant regardless of the detection method in the PedA channel due to the flat fading condition. It can be seen that the average effect becomes larger in frequency selective channels such as the VehA channel. And in the case of this VehA, the proposed invention reduces complexity by 57.7% compared to DPRD, but similar performance can be seen.

Claims (5)

In the method of estimating the PCID and RFN of a channel using NSSS,
NSSS acquisition step of acquiring a narrowband secondary synchronization signal from the synchronization signal; [Equation 1, 2]
Channel information estimation step of estimating a channel from the obtained NSSS signal; [Equation 3, 5]
obtaining a cross-correlation of estimated values according to an objective function represented by Equation 1 below, using the estimated channel information;
Method for estimating PCID and RFN of a channel including a.

(Equation 1)

(here,

It is a relationship. And Y is a received signal, E is a candidate signal. also,

is a value that can determine the RFN and is a candidate group of m values.

is a value that can be an n value,

is a candidate group that can be u . also

is a value that can be w . At this time, m is an index value for determining the RFN value through which the NSSS is transmitted, n is a cell specific parameter, u is a cyclic shift value, and w is a root sequence value.

is a set of subcarriers of the entire NSSS,

is the number of OFDM blocks in a radio frame. l represents the order of signals, k represents the order of subcarriers,

is the received signal, and the l -th signal is in the k -th subcarrier

can be expressed as) According to claim 1,
Dividing the estimated channel information into an imaginary part and a real part; and
Separating the imaginary part and the real part from the cross-correlation according to the objective function using the characteristic of the symmetric complex conjugate of the NSSS signal as shown in Equation 2 below;
Method for estimating PCID and RFN of a channel including a.
(Formula 2)

(here,

is,

, where subscript I means the imaginary part of the variable, subscript Q means the real part of the variable, and the remaining variables are the same as defined in Equation 1 of claim 1) According to claim 2,
a complexity reduction step of reducing complexity as in Equation 3 by taking an absolute value of the imaginary part of Equation 2;
PCID and RFN estimation method of a channel further comprising.
(Formula 3)

(here,

Is

is the hypothesis value of

is eliminated by the absolute value. w' is a root sequence value,

is a value that can be w')) According to claim 3,
Estimating ( m,n,u,w' ) through Equation 4 below; and
estimating an RFN using the estimated ( m,n,u,w' );
PCID and RFN estimation method of a channel further comprising.
(Formula 4)

(m is an index value that determines the RFN value through which NSSS is transmitted, n is a cell-specific parameter, u is a cyclic shift value, w, w' are root sequence values,

When is a value that can be w,

is a value that can be w',

are estimated values of m, n, u, and w', respectively) According to claim 4,
From Equation 5 below

By estimating the value and calculating it with the estimated m, n, and u values,

estimating a PCID value from;
PCID and RFN estimation method of a channel further comprising.

(Formula 5)

(

is a possible candidate group.

are estimated values of n, u, and w', respectively.

indicates the OFDM block number in the radio frame)

Images