CN106101042A - A kind of CFO method of estimation based on many noises - Google Patents

Abstract

A kind of CFO methods of estimation based on many noises disclosed by the invention, the method includes that many noises trigger, and captures composite signal；Estimation deflection factor and estimation CFO；The CFO methods of estimation based on many noises of gained of the present invention, utilize multiple noise, produce different reception signals.The character docking collection of letters number utilizing unitary matrice transposition reversible synthesizes, and the dependency in conjunction with channel draws the frequency deflection launched with receive signal, on the premise of certain signal to noise ratio, possesses less least mean-square error.

Description

A kind of CFO method of estimation based on many noises

Technical field

The CFO (Carrier Frequency Offset, carrier frequency offset) that the present invention relates to a kind of LTE system estimates Method, a kind of CFO methods of estimation based on many noises.

Background technology

LTE is the communication system that 3GPP is the most developing, and it is by taking OFDM (Orthogonal Frequency Division Multiplexing, OFDM) multichannel mode, each subcarrier is overlapped, greatly Width promote system spectrum efficiency, be very suitable for being wirelessly transferred of high data rate, this be also LTE compare 2/3G technology can One of reason that cell throughout is greatly improved.Problematically, competence exertion when OFDM needs each subcarrier to keep its orthogonality Performance, the most sensitive to synchronous error, and ISI (Intern present in the multidiameter fading channel of practical radio communication environment Symbol Interference, intersymbol interference) carrier frequency offset CFO and symbol time offset STO (Symbol can be caused Timing Offset), cause the bit error rate to improve, LTE system hydraulic performance decline.It addition, transmitter carrier frequency is local with receiver The tiny frequency deviation existed between agitator, and the Doppler frequency shift that the receiving-transmitting sides relative motion owing to communicating produces is also The generation of CFO can be caused.To this end, the present invention designs a kind of CFO method of estimation MNE (Multiple Noises based on many noises Based Evaluation), utilize multiple noise to disturb launching signal, produce different reception signals, and to reception Signal carries out capture synthesis, and then estimates frequency deviation, it is possible to the problem solving tradition blind estimate algorithm acquisition accuracy difference.

Summary of the invention

The invention aims to solve CFO deviation in LTE network and cause the decline problem of OFDM performance, thus evade Better profit from the spectrum efficiency of whole communication system in a multi-path environment, it is provided that higher cell traffic throughput, promote visitor Family perception, it is proposed that a kind of CFO methods of estimation based on many noises.

In order to achieve the above object, the CFO methods of estimation based on many noises designed by the present invention, touch including many noises Send out, capture composite signal, estimate the steps such as deflection factor, estimation CFO: add multiple interchannel noise to launching signal, to difference Reception signal synthesize, and then frequency deviation is estimated.

Preferably scheme comprises the following steps:

Step 1: many noises trigger；

Step 1-1: set constant N_g, for a length of N=N_g ²Transmitting signal x (n) (n=0..N-1), it receives letter Number y (n)=x (n) * δ (ε)+w (n), wherein δ is the signal deflection factor caused due to CFO, and it is the function of ε, and ε is to be evaluated CFO, w (n) be interchannel noise；

Step 1-2: set amount of noise k, produces unitary matrice training sequence

U_k={ u₁(n), u₂(n) ... u_k(n) }, stochastic generation interchannel noise W_k={ w₁(n), w₂(n) ... w_k(n) }, utilize The channel model that step 1-1 is set up produces corresponding reception signal Y respectively_k={ y₁(n), y₂(n) ... y_k(n)}；

Step 2: capture composite signal；

Step 2-1: for W_kAnd Y_k, calculate composite signal Z_k=Y_k-W_k={ z₁(n), z₂(n) ... z_k(n) }, wherein z_i (n)=y_i(n)-w_i(n), i=0 ... k；

Step 2-2: utilize unitary matrice U_k, capture causes the deflection variable of CFO

Wherein U^*Represent the conjugate transpose computing of U matrix, r_fi (n)=u_i ^*(y_i(n)-w_i(n)), i=0 ... k；

Step 3: estimation deflection factor

Step 3-1: calculate deflection inequality

Step 3-2: calculate deflection factor

Step 4: estimation CFO；

CFO is estimated by the deflection factor utilizing step 3-2 to produce, Wherein, Re () represents the real part functions seeking imaginary number, and j represents imaginary unit.

The CFO methods of estimation based on many noises of gained of the present invention, utilize multiple noise, produce different reception signals. The character docking collection of letters number utilizing unitary matrice transposition reversible synthesizes, thus draws the frequency deflection launched with receive signal.

The CFO evaluation methods based on many noises of gained of the present invention, it is possible to achieve same transmitting signal is made an uproar in different background Training under sound, utilizes the dependency of channel to realize the estimation to frequency deflection.

Accompanying drawing explanation

Fig. 1 CFO based on many noises method of estimation flow process；

Fig. 2 is MNE algorithm and tradition method of estimation comparison diagram based on CP.

Detailed description of the invention

The invention will be further described to combine accompanying drawing below by embodiment, and idiographic flow can refer to Fig. 1.

Embodiment 1:

The CFO methods of estimation based on many noises that this example describes, including including that many noises trigger, capture composite signal, Estimation deflection factor, estimates the steps such as CFO.

Preferably scheme comprises the following steps:

Step 1: many noises trigger；

Step 1-1: set constant N_g, for a length of N=N_g ²Transmitting signal

X (n) (n=0..N-1), its reception signal y (n)=x (n) * δ (ε)+w (n), wherein δ is the letter caused due to CFO Number deflection factor, it is the function of ε, and ε is CFO to be evaluated, and w (n) is interchannel noise；

Step 1-2: set amount of noise k, produces unitary matrice training sequence

U_k={ u₁(n), u₂(n) ... u_k(n) }, stochastic generation interchannel noise W_k={ w₁(n), w₂(n) ... w_k(n) }, utilize The channel model that step 1-1 is set up produces corresponding reception signal Y respectively_k={ y₁(n), y₂(n) ... y_k(n)}；

Step 2: capture composite signal；

Step 2-1: for W_kAnd Y_k, calculate composite signal Z_k=Y_k-W_k={ z₁(n), z₂(n) ... z_k(n) }, wherein z_i (n)=y_i(n)-w_i(n), i=0 ... k；

Step 2-2: utilize unitary matrice U_k, capture causes the deflection variable of CFO

Wherein U^*Represent the conjugate transpose computing of U matrix, r_fi (n)=u_i ^*(y_i(n)-w_i(n)), i=0 ... k；

Step 3: estimation deflection factor

Step 3-1: calculate deflection inequality

Step 3-2: calculate deflection factor

Step 4: estimation CFO；

CFO is estimated by the deflection factor utilizing step 3-2 to produce,

Wherein, Re () represents the real part functions seeking imaginary number, and j represents imaginary number Unit.

Below with N=4, N_gAs a example by=2 being specifically described this method, representative basis data are as shown in table 1.

Step 1: many noises trigger；

Step 1-1: utilize additive white Gaussian noise to generate multiple channel noise signals

w₁(n)=awgn (rand (2,2)+i*rand (2,2), 15)=[9.0949,8.5906；3.6409,2.6249],

w₂(n)=awgn (rand (2,2)+i*rand (2,2), 15)=[4.4900,8.2094；5.1739,1.7341],

W₂={ w₁(n), w₂(n)}；

Step 1-2: generate unitary matrice training sequence u₁(n)=[-0.5257 ,-0.8507；-0.8507,0.5257], U₂ (n)=[-0.7678,0.6407；-0.6407 ,-0.7678], U₂={ u₁(n), u₂(n) }, utilize channel model to produce respectively and connect The collection of letters number

y₁(n)=[7.7186-0.0025i, 7.2142-0.0108i；3.3159+0.0016i, 2.3000-

0.0004i],

y₂(n)=[4.3630+0.0019i, 8.0824+0.0012i；3.7654-0.0023i, 0.3257-

0.0107i],

Y₂={ y₁(n), y₂(n)}；

Step 2: capture composite signal；

Step 2-1: calculate composite signal

Z₂=Y₂-W₂=

{ [-1.3764-0.0025i ,-1.3763-0.0108i；-0.3249+0.0016i ,-0.3249-

0.0004i]；[-0.1270+0.0019i ,-0.1270+0.0012i；-1.4085-

0.0023i ,-1.4085-0.0107i] }；

Step 2-2: capture causes the deflection variable of CFO

S_Rf=∑_kR_fk=∑_kU_k ^**Z_k=[2.0000-0.0000i, 2.0000+0.0119i；2.0000+

0.0060i, 1.9999+0.0179i]；

Step 3: estimation deflection factor

Calculate deflection inequalityDeflection factor

Step 4: estimation CFO；

CFO is estimated by the deflection factor utilizing step 3-2 to produce,

$\mathrm{\ϵ}=\mathrm{Re}\left(log\right(\mathrm{\δ}\left(\mathrm{\ϵ}\right)/N)*\frac{N}{\mathrm{\π}*(N-1)*j})=\mathrm{0.0019.}$

We, to the MNE algorithm designed by the present invention and tradition method of estimation based on CP, estimate based on leading Moose Three kinds of method performances such as method and Classen estimation based on pilot tone are contrasted, and see shown in accompanying drawing 2.Tied by contrast Fruit can be seen that, Moose and Classen method of estimation is under noise jamming, and performance shake is little, and based on CP and MNE estimation side Method shake is relatively large, main reason is that MNE considers the reason of many noises.But under identical signal to noise ratio snr premise, MNE possesses less least mean-square error MSE.

Wherein, the table mentioned in embodiment is as follows:

Table 1

Sequence number	Project	Data
			1	Signal to noise ratio (SNR)	15
2	Noise signal	AWGN
			3	Amount of noise (k)	2

。

Claims (1)

1. a CFO method of estimation based on many noises, it is characterised in that:

Step 1: many noises trigger；

Step 1-1: set constant N_g, for a length of N=N_g ²Transmitting signal x (n) (n=0..N-1), it receives signal y N ()=x (n) * δ (ε)+w (n), wherein δ is the signal deflection factor caused due to CFO, and it is the function of ε, and ε is to be evaluated CFO, w (n) are interchannel noise；

Step 1-2: set amount of noise k, produces unitary matrice training sequence U_k={ u₁(n), u₂(n) ... u_k(n) }, stochastic generation Interchannel noise W_k={ w₁(n), w₂(n) ... w_k(n) }, the channel model utilizing step 1-1 to set up produces corresponding reception respectively Signal Y_k={ y₁(n), y₂(n) ... y_k(n)}；

Step 2: capture composite signal；

Step 2-1: for W_kAnd Y_k, calculate composite signal Z_k=Y_k-W_k={ z₁(n), z₂(n) ... z_k(n) }, wherein z_i(n)=y_i (n)-w_i(n), i=0 ... k；

Step 2-2: utilize unitary matrice U_k, capture causes the deflection variable of CFO Wherein U^*Represent the conjugate transpose computing of U matrix, r_fi(n)=u_i ^*(y_i(n)-w_i(n)), i=0 ... k；

Step 3: estimation deflection factor

Step 3-1: calculate deflection inequality

Step 3-2: calculate deflection factor

Step 4: estimation CFO；

CFO is estimated by the deflection factor utilizing step 3-2 to produce,Wherein, Re () represents the real part functions seeking imaginary number, and j represents imaginary unit.