CN103475620B - A kind of channel estimation methods based on OFDM transmission technology in aviation communication system - Google Patents

Abstract

The invention belongs to the communications field, be specifically related to the channel estimation methods of OFDM technology transmission plan in aviation communication system.Only have the characteristic of two footpaths and linear change according to aviation channel, first find out the position in these two footpaths, and then other footpath zero setting beyond these two footpaths.This not only lowers the impact of noise on channel estimating, improve the accuracy of estimation, but also decrease the pilot number into estimating large volumes of channels parameter.

Description

A kind of channel estimation methods based on OFDM transmission technology in aviation communication system

Technical field

The invention belongs to the communications field, be specifically related to the channel estimation methods of OFDM technology transmission plan in aviation communication system.

Background technology

Along with further developing of global IT application, the demand of aeronautical telecommunication services also becomes day by day urgent.The transmission technology that current aviation wireless communication system uses, from twentieth century five, the sixties use till today always.What its modulation technique adopted is double-side band-amplitude modulation, and channel spacing bandwidth is 8.33kHz.The minimum transfer bandwidth of double-side band-amplitude modulation system is 8kHz, can not improve the capacity of communication system by the method reducing channel spacing.So far, double-side band-amplitude modulation communication technology has developed into the end.In order to meet the demand of very high frequency aviation wireless communication system of Large Copacity, reliable communication, the new communication technology needs to be applied in aviation wireless communication system.

Due to the particularity of aviation channel, large multidiameter delay and Doppler frequency shift are that channel estimating proposes very large challenge, can not directly apply conventional channel method of estimation, must study new channel estimation methods.In existing public technology, OFDM is little as the key technology research of air communications, also not for the channel estimation methods of aviation channel characteristic.

According to the feature of channel linearity change in prior art, propose a kind of Quick-Change channel method of estimation, the method considers the impact of subcarrier interference, and amount of calculation is little, does not need the advantage knowing channel statistic.But when channel maximum multipath time delay is very large, and when channel parameter to be estimated is more than the number of training symbol, described Quick-Change channel method of estimation just no longer can be used.In air communications, main scene is scene of cruising, and now the communication channel on aircraft and ground is the model in two footpaths, and namely time delay is the direct projection footpath of zero, and another is the reflection footpath of maximum delay.According to these characteristics of aviation channel, the present invention proposes a kind of channel estimation methods based on OFDM transmission technology in aviation communication system.

Summary of the invention

The object of the present invention is to provide a kind of channel estimation methods based on OFDM transmission technology in aviation communication system, the method can solve the problem of long time delay in aviation channel estimation.

The object of the invention is to be achieved through the following technical solutions:

Based on a channel estimation methods for OFDM transmission technology in aviation communication system, comprise the following steps:

S1, choose N _pindividual training symbol described training symbol lays respectively at subcarrier p (1), p (2) ..., p (N _p) place;

S2, take out subcarrier p (1), p (2) described in S1 at receiving terminal ..., p (N _p) the frequency-domain received signal Y of correspondence position _p, to described frequency-domain received signal Y _pcombine, form equation group, that is,

$Y_p=\left[\begin{array}{c}{Y}_{p\left(1\right)}\\ \·\\ \·\\ \·\\ Y_{p\left(N_p\right)}\end{array}\right]=A{h}_{\mathrm{ave}}+\mathrm{B\α}+e=\mathrm{Qh}+e,$ Wherein, h _ave=[h _ave(0), h _ave(1) ... h _ave(L-1)] ^t, α=[α ₀, α ₁... α _l-1] ^t, h _averepresent the average in the every bar footpath of time domain fading channel, α represents the rate of change in the every bar footpath of time domain fading channel, and L represents the maximum multipath number after channel dispersion, L > L "=τ _max/ T _s, wherein L " is channel maximum multipath time delay τ _maxdivided by channel sample time T _s, Ah _avefor desired signal, B α is the distracter that training symbol produces, and e is error term, and the inter-sub-carrier interference that described error term e is produced by interchannel noise w and non-training symbol is formed;

S3, definition interference factor C _k-n=-(1-e ^{-j2 π (k-n)/N}) ^-1, b _n=[1, e ^{-j2 π n/N}..., e ^{-j2 π n (L-1)/N}] ^t, wherein, 0≤n≤N-1,0≤k≤N-1, N represents subcarrier number, and j is imaginary unit,

Definition

$A={\left[\begin{array}{ccc}{s}_{p\left(1\right)}{b}_{p\left(1\right)}^T& \·\·\·& s_{p\left(N_p\right)}{b}_{p\left(N_p\right)}^T\end{array}\right]}^T$

$B={\left[\begin{array}{ccc}\underset{\underset{\underset{n\∈\mathrm{pilot}}{n\≠p\left(1\right)}}{n=0}}{\overset{N-1}{\mathrm{\Σ}}}{C}_{p\left(1\right)-n}{S}_n{b}_n^T& \·\·\·& \underset{\underset{\underset{n\∈\mathrm{pilot}}{n\≠p\left(N_p\right)}}{n=0}}{\overset{N-1}{\mathrm{\Σ}}}{C}_{p\left(N_p\right)-n}{s}_n{b}_n^T\end{array}\right]}^T$

Q=[AB]

h=[h _aveα] ^T

$e=\left[\begin{array}{c}{e}_{p\left(1\right)}\\ \·\\ \·\\ \·\\ e_{p\left(N_p\right)}\end{array}\right]=\left[\begin{array}{c}\underset{\underset{n\∉\mathrm{pilot}}{n=0}}{\overset{N-1}{\mathrm{\Σ}}}{C}_{p\left(1\right)-n}{S}_n{b}_n^T\\ \·\\ \·\\ \·\\ \underset{\underset{n\∉\mathrm{pilot}}{n=0}}{\overset{N-1}{\mathrm{\Σ}}}{C}_{p\left(N_p\right)-n}{S}_n{b}_n^T\end{array}\right]\mathrm{\α}+w,$

Wherein, the n ∈ pilot in B expression formula represents that the value of n is taken from set pilot, e expression formula represent that the value of n takes from the value beyond set pilot, interchannel noise be length be N _pcolumn vector, described set pilot is N _pindividual training symbol be positioned at subcarrier p (1), p (2) ..., p (N _p) value;

S4, characteristic according to aviation channel two footpath, and the maximum multipath time delay of channel, obtain the Position Approximate in two footpaths after channel dispersion, the value near reservation, by other footpath zero setting, the time delay in the main footpath of Article 1 is zero, and after discretization, position is just at front M ₁in bar footpath, Article 2 reflection footpath is at maximum multipath time delay place, and the position after its discretization, near L ', is spaced apart leg, wherein L '=L "+J ₁, use l ₁, l ₂represent M near the main footpath of Article 1 after channel dispersion respectively ₁neighbouring (2*leg+1) bar path position of bar path position and Article 2 reflection footpath, that is,

l ₁=[0,1,…M ₁-1]，l ₂＝[L′-leg，L′-leg+1，…，L′,L′+1,…,L′+leg]，

Order

A′=A(:,[l ₁l ₂])

B′=B(:,[l ₁l ₂])

Q′=[A′B′]，

Wherein, A ', B ' represent the l taking out matrix A, B respectively ₁and l ₂the row of vector representative;

S5, according to S2 frequency-domain received signal Y _pcan obtain: h '=Q ^'+y _p+ Q ^'+e, ignores the impact of error term e, can obtain channel parameter estimation: h '=Q ^'+y _p, wherein, Q ^'+=(Q ^{' H}q ') ^-1q ^{' H}represent the pseudo inverse matrix of Q ', Q ^'+e represents evaluated error, h '=[h ' _aveα '] ^texpression length is (M ₁+ 2*leg+1) one dimension column matrix, h ' _ave=[h ' _ave(0) ..., h ' _ave(M ₁+ 2*leg)] expression length is (M ₁+ 2*leg+1) one dimension column matrix, expression length is (M ₁+ 2*leg+1) one dimension column matrix;

S6, compare h ' _avein amplitude, find out h ' respectively _ave(0) ~ h ' _ave(M ₁-1) and h ' _ave(M ₁) ~ h ' _ave(M ₁+ 2*leg) in maximum two h ' _ave(t ₁), h ' _ave(t ₂), h ' _ave(t1), h ' _ave(t ₂) position be t respectively ₁, t ₂, make l ₁=t ₁, l ₂=L "+J ₁-leg+t ₂-M ₁, by formula Q '=[A ' B '], h '=Q ^'+y _pto new channel parameter estimation value;

S7, by the h of each element of time domain channel matrix in channel new described in S6 _ave(n), α _nvalue zero setting, wherein, t ₁, (L '+J ₁-leg+t ₂) these two positions time domain channel value by

$h(k,n)=h_{\mathrm{ave}}\left(n\right)+(k-\frac{N-1}{2}){\mathrm{\α}}_{n}0\≤k\≤N-1,n=t_1,t_2$ Draw.

Further, the maximum multipath number L=|L after channel dispersion described in S2 " |+J ₁+ J ₂, wherein, | L " | for L " rounds, J ₁be the discrete footpath number before zero moment, J ₂represent the discrete footpath number after maximum delay.

Further, 5≤J ₁≤ 30,5≤J ₂≤ 30.

Further, Q ' described in S5 is sequency spectrum matrix when choosing pilot tone.

The invention has the beneficial effects as follows: the characteristic only having two footpaths and linear change according to aviation channel, first find out the position in these two footpaths, and then other footpath zero setting beyond these two footpaths.This not only lowers the impact of noise on channel estimating, improve the accuracy of estimation, but also decrease the pilot number into estimating large volumes of channels parameter.As traditionally method of estimation, even if OFDM symbol is all used for transmission pilot tone and can not estimates whole channel parameter, only need now a no more than N/4 pilot tone just can obtain whole channel parameter, improve the validity of transmission.

Accompanying drawing explanation

Fig. 1 is channel estimating flow chart.

Fig. 2 is pilot configuration schematic diagram.

Fig. 3 is simulation result.

Embodiment

Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described:

Present embodiment adopts Matlab2013a emulation platform to carry out running experiment.Ofdm system parameter: subcarrier N=1024, Cyclic Prefix 660, frequency pilot sign number N _p=256, data symbol number 768, symbol period T _s=0.05us.P (1), p (2) ..., p (N _p) value by the determining positions of pilot tone place subcarrier, every 4 one group is evenly distributed in and sends in signal, and pilot configuration is as shown in Figure 2.

Wireless channel environment is Doppler frequency shift is f _d=4533.3Hz and maximum multipath time delay τ _maxthe aviation two footpath channel model of=33us, Rice factor (direct projection footpath and the ratio reflecting footpath) is 15dB, channel width W=20MHz.J ₁value be 11.

Carrier frequency size is 2G, and modulation system is QPSK, 1/2 convolution code, interleaving mode 48*32.Now normalization maximum doppler frequency (W/ (N*f _d)) be 0.23, discretized channel maximum multipath number L "=W* τ _maxbe 660, L '=L "+J ₁=671.

Selecting All Parameters M ₁=20, leg=20, obtains l ₁=[1,2 ..., 19,20], l ₂=[651,650 ..., 680,681], obtain Q ' by Q '=[A ' B '].By h '=Q ^'+y _pobtain h ', that wherein frequency pilot sign is chosen is the N of size between 0 ~ 1 _pindividual known pseudo random sequence, this is to ensure that Q ' is for sequency spectrum matrix.Relatively h ' _avein amplitude, find out h ' respectively _ave(0) ~ h ' _aveand h ' (19) _ave(20) ~ h ' _ave(60) two maximum in is h ' _ave(11), h ' _ave(40), l is made ₁=11, l ₂=L "+J ₁-leg+t ₂-M ₁=671, then by Q '=[A ' B '] and h '=Q ^'+y _pobtain new channel parameter estimation value.The h except 11,671 these two positions _ave(n), α _nvalue zero setting, the time domain channel value of these two positions is obtained by following formula:

$h(k,n)=h_{\mathrm{ave}}\left(n\right)+(k-\frac{N-1}{2}){\mathrm{\α}}_n,0\≤k\≤N-1,n=\mathrm{11,671}.$

Finally obtain all time-domain matrix element values:

$h(k,n)=\left\{\begin{array}{cc}{h}_{\mathrm{ave}}\left(n\right)+(k-\frac{N-1}{2}){\mathrm{\α}}_n& 0\≤k\≤N-1,n=\mathrm{11,671}\\ 0& 0\≤k\≤N-1,n\≠\mathrm{11,671}\end{array}\right..$

Adopt method described in the embodiment of the present invention to carry out emulation testing, simulation result as shown in Figure 3.After estimating channel, according to document " ALow-ComplexityICICancellationSchemeinFrequencyDomainfor OFDMinTime-varyMultipathChannels (byHongmeiWang; XiangChen, ShidongZhou, YanYao.2005IEEE16 ^thinternationalSymposiumonPersonal, IndoorandMobileRadioCommunications) " method that proposes carries out equilibrium, offsets fading channel to the impact sending signal.Can find out aviation channel method of estimation better performances of the present invention by simulation result, after adding coding, the bit error rate when signal to noise ratio is 8dB reaches 1e-5 soon, only has the performance loss of 1dB with perfect channel estimation result.

Claims (4)

1. based on a channel estimation methods for OFDM transmission technology in aviation communication system, it is characterized in that: its step is as described below:

S1, choose N _pindividual training symbol described training symbol lays respectively at subcarrier p (1), p (2) ..., p (N _p) place;

S2, take out subcarrier p (1), p (2) described in S1 at receiving terminal ..., p (N _p) the frequency-domain received signal Y of correspondence position _p, to described frequency-domain received signal Y _pcombine, form equation group, that is,

$Y_p=\left[\begin{array}{c}{Y}_{p\left(1\right)}\\ \·\\ \·\\ \·\\ Y_{p\left(N_p\right)}\end{array}\right]={\mathrm{Ah}}_{ave}+B\mathrm{\α}+e=Qh+e,$ Wherein, h _ave=[h _ave(0), h _ave(1) ... h _ave(L-1)] ^t,

α=[α ₀, α ₁... α _l-1] ^t, h _averepresent the average in the every bar footpath of time domain fading channel, α represents the rate of change in the every bar footpath of time domain fading channel, and L represents the maximum multipath number after channel dispersion, L > L "=τ _max/ T _s, wherein L " is channel maximum multipath time delay τ _maxdivided by channel sample time T _s, Ah _avefor desired signal, B α is the distracter that training symbol produces, and e is error term, and the inter-sub-carrier interference that described error term e is produced by interchannel noise w and non-training symbol is formed;

S3, definition interference factor C _k-n=-(1-e ^{-j2 π (k-n)/N}) ^-1, b _n=[1, e ^{-j2 π n/N}..., e ^{-j2 π n (L-1)/N}] ^t, wherein, 0≤n≤N-1,0≤k≤N-1, N represents subcarrier number, and j is imaginary unit,

Definition

$A={\left[\begin{array}{ccc}{s}_{p\left(1\right)}{b}_{p\left(1\right)}^T& ...& s_{p\left(N_p\right)}{b}_{p\left(N_p\right)}^T\end{array}\right]}^T$

$B={\left[\begin{array}{ccc}\underset{n\∈pilot}{\overset{N-1}{\underset{n\≠p\left(1\right)}{\underset{n=0}{\mathrm{\Σ}}}}}{C}_{p\left(1\right)-n}{s}_n{b}_n^T& ...& \underset{n\∈pilot}{\overset{N-1}{\underset{n\≠p\left(N_p\right)}{\underset{n=0}{\mathrm{\Σ}}}}}{C}_{p\left(N_p\right)-n}{s}_n{b}_n^T\end{array}\right]}^T$

Q＝[AB]

h＝[h _aveα] ^T

$e=\left[\begin{array}{c}{e}_{p\left(1\right)}\\ \·\\ \·\\ \·\\ e_{p\left(N_p\right)}\end{array}\right]=\left[\begin{array}{c}\underset{n\∉pilot}{\overset{N-1}{\underset{n=0}{\mathrm{\Σ}}}}{C}_{p\left(1\right)-n}{s}_n{b}_n^T\\ \·\\ \·\\ \·\\ \underset{n\∉pilot}{\overset{N-1}{\underset{n=0}{\mathrm{\Σ}}}}{C}_{p\left(N_p\right)-n}{s}_n{b}_n^T\end{array}\right]\mathrm{\α}+w,$

Wherein, the n ∈ pilot in B expression formula represents that the value of n is taken from set pilot, e expression formula represent that the value of n takes from the value beyond set pilot, interchannel noise be length be N _pcolumn vector, described set pilot is N _pindividual training symbol be positioned at subcarrier p (1), p (2) ..., p (N _p) value;

S4, characteristic according to aviation channel two footpath, and the maximum multipath time delay of channel, obtain the Position Approximate in two footpaths after channel dispersion, the value near reservation, by other footpath zero setting, the time delay in the main footpath of Article 1 is zero, and after discretization, position is just at front M ₁in bar footpath, Article 2 reflection footpath is at maximum multipath time delay place, and the position after its discretization, near L ', is spaced apart leg, wherein L '=L "+J ₁, J ₁be the discrete footpath number before zero moment, use l ₁, l ₂represent M near the main footpath of Article 1 after channel dispersion respectively ₁neighbouring (2*leg+1) bar path position of bar path position and Article 2 reflection footpath, that is,

l ₁＝[0,1,…M ₁-1]，l ₂＝[L′-leg，L′-leg+1，…，L′,L′+1,…,L′+leg]，

Order

A′＝A(:,[l ₁l ₂])

B′＝B(:,[l ₁l ₂])

Q′＝[A′B′]，

Wherein, A ', B ' represent the l taking out matrix A, B respectively ₁and l ₂the row of vector representative;

S5, according to S2 frequency-domain received signal Y _pcan obtain: h '=Q ' ⁺y _p+ Q ' ⁺e, ignores the impact of error term e, can obtain channel parameter estimation: h '=Q ' ⁺y _p, wherein, Q ' ⁺=(Q ' ^hq ') ^-1, Q ' ^hrepresent the pseudo inverse matrix of Q ', Q ' ⁺e represents evaluated error, h '=[h ' _aveα '] ^texpression length is (M ₁+ 2*leg+1) one dimension column matrix, h ' _ave=[h ' _ave(0) ..., h ' _ave(M ₁+ 2*leg)] expression length is (M ₁+ 2*leg+1) one dimension column matrix, expression length is (M ₁+ 2*leg+1) one dimension column matrix;

S6, compare h ' _avein amplitude, find out h ' respectively _ave(0) ~ h ' _ave(M ₁-1) and h ' _ave(M ₁) ~ h ' _ave(M ₁+ 2*leg) in maximum two h ' _ave(t ₁), h ' _ave(t ₂), h ' _ave(t ₁), h ' _ave(t ₂) position be t respectively ₁, t ₂, make l ₁=t ₁, l ₂=L "+J ₁-leg+t ₂-M ₁, by formula Q '=[A ' B '], h '=Q ' ⁺y _pto new channel parameter estimation value;

S7, by the h of each element of time domain channel matrix in channel new described in S6 _ave(n), α _nvalue zero setting, wherein, t ₁, (L '+J ₁-leg+t ₂) these two positions time domain channel value by

$h(k,n)=h_{ave}\left(n\right)+(k-\frac{N-1}{2}){\mathrm{\α}}_n$ 0≤k≤N-1, n=t ₁, t ₂draw.

2. a kind of channel estimation methods based on OFDM transmission technology in aviation communication system according to claim 1, is characterized in that: maximum multipath number L=|L after channel dispersion described in S2 " |+J ₁+ J ₂, wherein, | L " | for L " rounds, J ₁be the discrete footpath number before zero moment, J ₂represent the discrete footpath number after maximum delay.

3. a kind of channel estimation methods based on OFDM transmission technology in aviation communication system according to claim 2, is characterized in that: 5≤J ₁≤ 30,5≤J ₂≤ 30.

4. a kind of channel estimation methods based on OFDM transmission technology in aviation communication system according to claim 1, is characterized in that: Q ' described in S5 is sequency spectrum matrix when choosing pilot tone.