KR20120033761A - Method for estimating snr(signal to noise ratio) in wireless communication system - Google Patents

Abstract

PURPOSE: A method for estimating a SNR(Signal To Noise Ratio) in a wireless communication system is provided to accurately estimate a SNR without a channel estimation process by using a preamble which a receiver and a transmitter of an IEEE 802.11n system know in common. CONSTITUTION: Transmission signal power estimation is performed with a predetermined algorithm for estimating a SNR(Signal To Noise Ratio). Noise power is obtained by using an average absolute value of the differences between two received preambles. The SNR is estimated in fixed transmission power through the relative change of noise power. An algorithm for estimating the SNR is expressed in a predetermined expression relation. Each SNR estimation algorithm uses a preamble made into two consecutive OFDM symbol sizes consisting of BPSK(Binary Phase Shift Keying) or QPSK(Quadrature Phase Shift Keying).

Description

Method for estimating SNR (signal to noise ratio) in wireless communication system}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for estimating signal to noise ratio (SNR) in a wireless communication system. In particular, a wireless communication capable of accurately estimating an SNR without using a channel estimation process by using preambles known to each other at a transmitting and receiving end of an IEEE 802.11n system. SNR estimation method in the system.

Techniques such as link adaptation, subcarrier allocation, power allocation, and the like are used to increase communication reliability and transmission rate in a wireless communication environment. These techniques require feedback information about the channel state, and in order to operate properly to improve performance, the receiver needs to accurately estimate the channel state information and design a low complexity SNR estimator.

Conventional preamble-based SNR estimation algorithms include a Boumard SNR estimation algorithm, a Ren SNR estimation algorithm, a Milan SNR estimation algorithm, and the like.

The Boumard SNR estimation algorithm performs SNR estimation under the assumption that there is little channel change between adjacent subcarriers in a 2 × 2 MIMO (Orthogonal Frequency Division Multiplexing) system, and two identical and consecutive preamble symbols Use In this case, the transmission preamble may be expressed as the following equation.

Here, n = 1, ..., N number of carriers, 0 represents the first transmission preamble, 1 represents the second transmission preamble.

는 추정된 전체 전송 프레임의 평균 파워를 나타내고,

는 평균 노이즈 파워를 나타낸다.

는 두 개의 연속된 수신 프리앰블 Y(0,n)과 Y(1,n)및 원래 전송 프리앰블 C(n)을 이용하여 수식 3과 같이

가 구해지며, 최종적으로 수식 4와 같이

의 절대값의 평균을 구함으로써

가 구해진다.SNR estimation is performed by estimating signal and noise power as shown in Equation 2 below. In Equation 2

Denotes the average power of the estimated total transmitted frames,

Represents the average noise power.

Is shown in Equation 3 using two consecutive receive preambles Y (0, n) and Y (1, n) and the original transmit preamble C (n).

Is finally obtained, as shown in Equation 4.

By taking the average of the absolute values of

Is obtained.

The Boumard SNR estimation algorithm can perform SNR estimation without channel estimation required by a conventional maximum likelihood (ML) or minimum mean squared error (MMSE) SNR estimator. However, since the noise power is estimated as shown in Equation 5 under the assumption that there is little channel change between adjacent subcarriers, the algorithm is not suitable when the channel change is large.

Next, the Ren SNR estimation algorithm is based on a preamble having two identical OFDM symbol sizes as in the Boumard SNR estimation algorithm, as shown in Equation 6 below, and unlike the Boumard SNR estimation algorithm, as in Equation 7 Noise power is estimated on the same subcarrier. Therefore, the Ren SNR estimation algorithm is an improvement over the Boumard SNR estimation algorithm, which is sensitive to the selective characteristics of the channel frequency. The signal power estimate is obtained by removing the estimated noise power from the total received signal power as in Equation 8.

는 상기 수식 3과 동일하게 구해진다.Determined by channel estimation

Is obtained in the same manner as in Equation 3.

Next, the Milan SNR estimation algorithm is an algorithm using a plurality of identical preambles in the time domain. If the preamble consisting of all N subcarriers is formed in the form of Q identical periodic preamble structures in the time domain as shown in FIG. 1 (a), in the frequency domain, the subcarriers of Q intervals as shown in FIG. Each time a signal appears periodically and a Null (zero) signal is generated between intervals. The Milan SNR estimation algorithm is a method of estimating SNR using this characteristic. That is, the estimation of S after FFT (fast Fourier transform) is performed at every Q interval in which a signal appears as shown in Equation 9 below, and the estimation of W is performed in a null carrier part as shown in Equation 10.

Y _p (m) is a received signal that appears every Q interval on the frequency axis and is expressed as in Equation 11 below.In Equation 12, Y _z (mQ + q) is m = 0,1, ..., 63, q = 0 , Every 1 indicates a signal received at a null subcarrier position. eta (k, n) represents additive white Gaussian noise (AWGN) of magnitude 1 and average 0 in the kth sample of the nth preamble.

는 위의 수식 11 및 12를 이용하여 다음의 수식 13과 같이 구해진다.Finally, the SNR estimate obtained by the Milan SNR estimation algorithm

Is obtained as in Equation 13 below using Equations 11 and 12 above.

However, the conventional SNR estimation algorithms as described above have been pointed out as a problem that the calculation process for SNR estimation is complicated and the feedback delay is large.

The present invention has been made in view of the above-described matters, and provides an SNR estimation method in a wireless communication system capable of accurately estimating SNR without using a channel estimation process by using preambles known to each other at a transmitting and receiving end of an IEEE 802.11n system. Has its purpose.

In order to achieve the above object, SNR estimation method in a wireless communication system according to the present invention,

By using a predetermined algorithm ρ _{av, NEW} for estimating SNR, the transmission signal power is estimated by using a size of a transmission preamble of 1, and the noise power is obtained by using an average of absolute values of the difference between two reception preambles. It is characterized by estimating SNR through change in relative noise power at fixed transmission power.

Here, the algorithms ρ _{av, NEW} for estimating the SNR may be expressed by the following mathematical relationship.

In the above formula, Y (0, n) and Y (1, n) are the size of a continuous transmission preamble having a size of 1 and having the same pattern of binary phase shift keying (BPSK) or quadrature phase shift keying (QPSK), respectively. The received signal after the FFT is shown.

According to the present invention, there is an advantage in that the SNR can be accurately estimated without a channel estimation process by using the preambles known to each other at the transmitting and receiving end of the IEEE 802.11n system.

1 is a diagram showing a preamble structure used in a conventional Milan SNR estimation algorithm.
2 is a diagram illustrating a transmission preamble frame structure employed in an SNR estimation method in a wireless communication system according to the present invention.
3 shows a comparison between the SNR value estimated by the present invention and conventional SNR estimation algorithms and the actual SNR value in a Rayleigh flat fading channel.
FIG. 4 shows a comparison of normalized mean square error (NMSE) performance of the present invention and conventional SNR estimation algorithms in a Rayleigh flat fading channel. FIG.
FIG. 5 shows a comparison between the actual SNR value and the SNR value estimated by the present invention and conventional SNR estimation algorithms in Rayleigh selective fading channel A. FIG.
FIG. 6 shows a comparison of the NMSE performance of the present invention and conventional SNR estimation algorithms in Rayleigh selective fading channel A. FIG.
FIG. 7 shows a comparison between the SNR value estimated by the present invention and conventional SNR estimation algorithms and the actual SNR value in Rayleigh selective fading channel B. FIG.
FIG. 8 shows a comparison of the NMSE performance of the present invention and conventional SNR estimation algorithms in Rayleigh selective fading channel B. FIG.
9 shows packet error rates of respective multipoint communication services (MCS) levels in a Rayleigh flat fading channel.
10 shows the rate of each MCS level in a Rayleigh flat fading channel.
FIG. 11 illustrates the transmission rates of the present invention and conventional SNR estimation algorithms when adaptive modulation and coding (AMC) is applied in a Rayleigh flat fading channel. FIG.
12 shows packet error rates of respective MCS levels in Rayleigh selective fading channel A. FIG.
FIG. 13 shows transmission rates of respective MCS levels in Rayleigh selective fading channel A. FIG.
FIG. 14 shows the transmission rates of the present invention and conventional SNR estimation algorithms when AMC is applied in Rayleigh selective fading channel A. FIG.
FIG. 15 shows packet error rates of respective MCS levels in Rayleigh selective fading channel B. FIG.
FIG. 16 shows transmission rates of respective MCS levels in Rayleigh selective fading channel B. FIG.
FIG. 17 illustrates the rate of the present invention and conventional SNR estimation algorithms when AMC is applied in Rayleigh selective fading channel B. FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

2 is a diagram illustrating a transmission preamble frame structure employed in an SNR estimation method in a wireless communication system according to the present invention.

In the SNR estimation method in the wireless communication system according to the present invention, a transmission signal power is obtained by using a transmission preamble having a size of 1 as shown in FIG. 2 by a predetermined algorithm ρ _{av, NEW} for estimating SNR. The noise power is estimated using the average of the absolute values of the difference between the two reception preambles, and the SNR is estimated through the change of the relative noise power at the fixed transmission power.

Here, the algorithms ρ _{av, NEW} for estimating the SNR may be expressed by the following mathematical relationship.

In the above formula, Y (0, n) and Y (1, n) are the size of a continuous transmission preamble having a size of 1 and having the same pattern of binary phase shift keying (BPSK) or quadrature phase shift keying (QPSK), respectively. The received signal after the FFT is shown.

Then, the performance of the SNR estimation algorithm (ρ _{av, NEW} ) employed in the SNR estimation method of the present invention and the conventional SNR estimation algorithm will be compared and analyzed based on data obtained through experiments.

Table 1 shows simulation parameters and Table 2 shows channel parameters, respectively.

Simulation parameters             parameter                     value        System Bandwidth (BW)                    20 MHz        1 OFDM symbol time      4㎲ (3.2㎲: FFT section + 0.8㎲: CP length) Data per spatial stream (SS)
Symbol count                    468       Data subcarriers                    64         Subcarrier spacing                  312.5 KHz             MIMO                Layered 2 × 2             Noise                    AWGN           FFT size                   64 point          GI (CP) size                   16 point       1 OFDM symbol sample                     80         SNR Estimation Technique          Boumard, Milan, Ren, New       Preamble Consecutive same in QPSK or BPSK
2 OFDM symbol           Modulation technique             BPSK, QPSK, 16QAM      Convolutional coding                  1/2, 3/4           Transport packet                    25000

Channel parameters          channel     Delay Path (samples)           Rayleigh power  Rayleigh Selective Channel A      3Path: [0 12 15]       [-1.92 -5.92 -9.92]  Rayleigh Selective Channel B     4Path: [0 12 15 18]    [-1.92 -5.92 -9.92 -12.92]  Rayleigh flat channel          No delay                 -

The system parameters are based on IEEE 802.11n MIMO-OFDM and performance analysis is performed on two channels with two transmit and receive antennas.

Each SNR estimation algorithm used a preamble consisting of two consecutive OFDM symbol sizes (64 points) consisting of BPSK or QPSK. The channel environment was experimented with a Rayleigh flat fading channel with little change in channel state, Rayleigh selective fading A with a maximum delay sample less than the cyclic prefix length, and Rayleigh selective channel B with a maximum delay sample larger than the CP length. It was.

As a performance evaluation method, a method of roughly comparing the estimated SNR estimated by each algorithm with the actual SNR value was adopted. For a detailed performance comparison, a normalized mean square error (NMSE) was calculated as shown in Equation 15 below. The analysis was based on this.

는 i번째 패킷의 수신 프리앰블에서 추정된 SNR 값,

는 실제 SNR 값을 나타낸다.Where N _t is 25,000 packets.

Is the SNR value estimated from the received preamble of the i th packet,

Represents the actual SNR value.

On the other hand, when comparing the SNR value estimated by each technique and the actual SNR value in the Rayleigh flat channel in FIG. 3, the Ren algorithm has a relatively higher SNR estimation error than other algorithms in the low SNR range. It can be seen that the case is almost identical to the actual SNR.

In addition, when confirming the more accurate performance difference between each technique through the NMSE performance results of FIG. In the order of Milan, Ren, the performance is good. As described above, the Boumard algorithm is a suitable algorithm when there is little change in the channel through the performance results, and the new SNR estimation algorithm of the present invention also shows the performance of the Boumard degree.

5 and 6 show performance curves and NMSE performance results comparing the actual SNR and the SNR estimated by each algorithm in Rayleigh selective channel A, respectively.

Referring to FIG. 5, the performance of Boumard can be seen that the SNR estimated up to -4dB is estimated close to the actual SNR, and then the estimation error is increased by the frequency selective characteristic of the channel. On the other hand, as shown in FIG. 6, in the case of Milan or Ren, the estimation error decreases gradually, and as the SNR increases, the NMSE decreases gradually and remains constant at about 0.3. In the case of the algorithm of the present invention, it can be seen that the NMSE can be estimated close to the actual SNR without affecting the frequency selective of the channel by about 10 ⁻³ .

7 and 8 show performance in a channel worse than Channel A with performance results in Rayleigh selective channel B having four multipaths and a channel maximum delay greater than CP.

Thus, the performance of each SNR estimation algorithm exhibits a high estimation error when compared to channel A. In FIG. 7, it can be seen that Boumard has not estimated the actual SNR value since -2 dB. In the case of the algorithm of the present invention and the Milan and Ren algorithm, it can be seen that the estimation can be made close to the actual SNR value. Among them, the algorithm of the present invention has a lower estimation error in terms of NMSE. Although the performance of the algorithm of the present invention is shown from 28 dB in FIG. 8 due to the influence of the frequency selective characteristic, it can be seen that the performance of all the channels is superior to other techniques.

In order to estimate the SNR at the receiving end and give this information as a feedback, to evaluate the transmission rate performance according to the estimated SNR performance, the adaptive modulation and coding (AMC) technique is applied based on each SNR estimation algorithm and analyzed.

Table 3 below shows MCS levels selected by appropriately considering a multipoint communication services (MCS) table of IEEE 802.11n to apply AMC.

MCS level selected for AMC application
MCS

modulation

Code
Rate

MCP

Data Rate [Mbps]
(Data
Subscriber = 64,
CP = 16)
Nss


Nsts

Stream1
Stream2
   One   BPSK   BPSK   1/2    One        16.0    2    2    2   QPSK   BPSK   1/2   1.5        24.0    2    2    3   QPSK   QPSK   1/2    2        32.0    2    2    4   QPSK   QPSK   1/2    3        48.0    2    2    5   16QAM   16QAM   1/2    4        64.0    2    2

                  * Nss: Number of spatial streams, * Nsts: Number of space-time streams

                  * MCP: Modulation and Coding Product

FIG. 9 shows packet error rates of respective MCS levels corresponding to Table 3 in a Rayleigh flat fading channel, and FIG. 10 shows a rate performance.

9 and 10, it can be seen that the packet error rate performance is not as good as the MCP increases (that is, the larger the MCS level), while the maximum transmission rate is larger. In FIG. 10, SNR estimation must be accurately performed to properly select an MCS level that optimally operates in an SNR within a reference range based on an SNR corresponding to an intersection point between transmission rates of each level. Therefore, the more accurate the SNR estimation is, the better the transmission rate is when AMC is applied. As shown in FIG. You can see that.

In conclusion, we can see that the proposed algorithm and Boumard have better transmission rate than the Ren or Milan algorithm in the flat fading channel.

12 and 13 illustrate packet error rate and transmission rate performance of each MCS level corresponding to Table 3 in Rayleigh selective fading channel A. FIG.

The performance aspect between the levels is similar to that in the flat fading channel, but the overall packet error rate performance is poor due to multipath fading. In particular, it can be seen that the performance of the MCS 4 and MCS 5 shows a relatively large performance difference when compared to the performance in the flat fading channel. Similarly, in the case of the data rate, the maximum data rate is satisfied at a higher SNR point than in the case of the flat fading channel.

FIG. 14 is a diagram illustrating each SNR estimation algorithm rate performance when AMC is applied in Rayleigh selective fading channel A. FIG.

As shown in FIG. 14, the Boumard algorithm in channel A can confirm that SNR estimation is not properly performed from about 10 dB and only the maximum data rate corresponding to MCS level 1 is satisfied. On the other hand, the algorithm of the present invention, and the Ren and Milan algorithms generally have good SNR estimation, and satisfy the maximum data rate corresponding to MCS 5. In addition, as shown in the NMSE performance result of FIG. 6, the algorithm of the present invention in channel A has a lower SNR estimation error rate than that of the Ren or Milan algorithm, so it can be seen that the average transmission rate is the best.

15 and 16 show packet error rate and rate performance of each MCS level in Rayleigh selective channel B. FIG.

15 and 16, since the maximum channel delay is larger than the protection interval (CP) when compared with the channel A, an InterSymbol Interference (ISI) occurs, and thus, even in the case of a packet error rate, the performance deterioration occurs. It can be seen that the data rate also does not satisfy the original maximum data rate.

FIG. 17 is a diagram illustrating a rate performance result of applying AMC as feedback information to an SNR estimate obtained by each algorithm in channel B. FIG.

Referring to FIG. 17, when looking at Boumard's AMC data rate in channel B, as in channel A, it shows a bad data rate due to a high SNR estimation error. In the other three algorithms, the SNR estimation is performed to some extent while satisfying the maximum data rate, and among them, it is confirmed that the data rate of the algorithm of the present invention is good on average.

As described above, the SNR estimation method in a wireless communication system has an advantage of accurately estimating SNR without using a channel estimation process by using preambles known to each other at the transmitting and receiving end of the IEEE 802.11n system.

As mentioned above, although the present invention has been described in detail with reference to the preferred embodiment, the present invention is not limited thereto, and it will be apparent to those skilled in the art that various changes and applications can be made without departing from the technical spirit of the present invention. Accordingly, the true scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of the same should be construed as being included in the scope of the present invention.

Claims (2)

By using a predetermined algorithm ρ _{av, NEW} for estimating SNR, the transmission signal power is estimated by using a size of a transmission preamble of 1, and the noise power is obtained by using an average of absolute values of the difference between two reception preambles. SNR estimation method in a wireless communication system for estimating SNR by changing relative noise power at fixed transmission power. The method of claim 1,
The algorithm (ρ _{av, NEW} ) for estimating the SNR is expressed by the following mathematical relationship.

Here, Y (0, n) and Y (1, n) are each composed of binary phase shift keying (BPSK) or quadrature phase shift keying (QPSK), and have a size equal to 1 and follow the FFT of consecutive transmission preambles having the same pattern. Indicates a received signal.

Images