CN101355543A - Channel Estimation Method for MIMO-SCFDE System Based on Orthogonal Training Sequence - Google Patents

Abstract

The present invention provides a MIMO-SCFDE system channel estimation method based on an orthogonal training sequence, which includes the following steps: (1) at the sending end, firstly design an orthogonal training sequence according to the number of transmitting antennas, and then use the designed training The sequence is sent out through each transmitting antenna in the corresponding order; (2) Temporarily store the received signal at each receiving antenna, and after receiving all the training sequence signals, first transform to the frequency domain, and then receive the corresponding The linear equations are solved according to each frequency-domain sub-channel between the antennas, and the state information of each frequency-domain sub-channel between the corresponding transmitting and receiving antennas is obtained; (3) the obtained channel state information between the corresponding transmitting and receiving antennas is transformed into In the time domain, window the time-domain channel to remove the influence of noise on the channel estimate, and then transform to the frequency domain to obtain the channel estimate. The invention is suitable for the characteristics of the MIMO-SCFDE system, has low complexity and can improve channel estimation precision.

Description

Channel Estimation Method for MIMO-SCFDE System Based on Orthogonal Training Sequence

technical field

The invention relates to a channel estimation method for a single carrier frequency domain equalization (MIMO-SCFDE) system with multiple transmitting antennas and multiple receiving antennas, and belongs to the technical field of broadband wireless communication.

Background technique

Orthogonal Frequency Division Multiplexing (hereinafter referred to as OFDM: Orthogonal Frequency Division Multiplexing) is a multi-carrier transmission technology that uses N subcarriers to divide the entire broadband channel into N parallel narrowband subchannels that are orthogonal to each other.

OFDM systems have many compelling advantages:

1. OFDM has very high spectral efficiency. There is no guard interval between sub-channels of OFDM, and the spectrum main lobes of signals in adjacent sub-channels overlap, which greatly improves the spectral efficiency of OFDM systems.

2. Strong anti-multipath interference ability and anti-fading ability. Since the general OFDM system adopts the cyclic prefix (hereinafter referred to as CP: Cyclic Prefix) method, it can completely eliminate the multipath interference caused by multipath propagation under certain conditions, and completely eliminate the destruction of carrier orthogonality caused by multipath propagation. Therefore, the OFDM system has a good ability to resist multipath interference; OFDM subcarriers divide the entire wideband channel into several narrow subchannels. Although the entire wideband channel may be a very uneven fading channel, the fading on each subchannel is Approximately flat, which makes the equalization of OFDM signals particularly simple, often requiring only a one-tap equalizer.

3. The implementation is relatively simple. When the quadrature amplitude modulation (QAM: Quadrature Amplitude Modulation) or multi-ary phase shift keying (MPSK: M-ary Phase Shift Keying) modulation method is used on the sub-channel, the modulation process can use the inverse discrete Fourier transform (hereinafter Referred to as IDFT: Inverse Discrete Fourier Transform), the demodulation process can be completed by discrete Fourier transform (hereinafter referred to as DFT: Discrete Fourier Transform), and they can use a fast algorithm: Fast Fourier Transform (hereinafter referred to as FFT: Fast Fourier Transform) and Inverse Fast Fourier Transform (hereinafter referred to as IFFT: Inverse Fast Fourier Transform) implementation. Neither multiple banks of oscillators are used to generate the carrier signal nor bandpass filter banks are used to separate the signals.

It is these advantages that make OFDM become a research hotspot in the past ten years, and it is considered as the supporting technology for future communication, especially broadband wireless communication. However, many shortcomings of the OFDM system itself, especially its peak-to-average power ratio (PAPR: Peak to Average Power Ratio) is too large, which limits its practicality; and frequency-domain balanced single carrier (hereinafter referred to as SC-FDE: Single Carrier with Frequency Domain Equalization) has the above-mentioned advantages of OFDM, and there is no PAPR problem of OFDM, and its performance and efficiency are basically equivalent to OFDM. The PAPR problem is a problem that the OFDM system itself is difficult to solve with a low cost (spectrum efficiency and power efficiency), so SC-FDE technology is currently receiving more and more attention.

Both OFDM and SC-FDE are block transmission technologies based on CP, which show outstanding advantages in resisting multipath fading environment of wireless channels, but their spectrum efficiency is still not very high. In recent years, multiple-input multiple-output (MIMO: Multiple-input Multiple-output) communication systems based on multi-antenna technology have shown great advantages in terms of spectral efficiency, and are generally considered to be one of the main supporting technologies for future wireless communications.

The MIMO system adopts the method of using multiple antennas at the transmitting end and the receiving end, and utilizes the diversity effect between antennas and the independence of channel characteristics between different antennas caused by multi-path transmission in the wireless environment to improve the channel capacity, thereby obtaining a higher channel capacity than a single antenna. The greater spectrum utilization of the system has received more and more attention in the increasingly tight wireless spectrum resources and broadband communications. MIMO systems can obtain diversity gain against signal fading or channel capacity gain through different methods. Generally speaking, MIMO technologies can be divided into three categories:

(1) Power efficiency can be improved by maximizing space diversity. The main methods include: delay diversity, space-time block codes (STBC: Space-Time Block Codes) and space-time trellis codes (STTC: Space-Time Trellis Codes) ;

(2) Use the MIMO channel to perform spatial multiplexing through a layered method to improve the effectiveness of the system, among which the most commonly used is the V-BLAST algorithm proposed by G.J.Foschini et al.;

(3) Using channel state information at the sending end, by applying SVD (Singular Value Decomposition) decomposition to the channel matrix, using the decomposed unitary matrix as the forward filter and backward filter at the sending end and receiving end can reach the limit capacity .

With the development of MIMO technology, many researchers have formed a broadband MIMO-OFDM system according to the characteristics of OFDM technology and MIMO technology, which greatly improves the bandwidth effectiveness and further improves the spectrum utilization rate. Simply put, the MIMO-OFDM system converts the broadband wireless channel into several narrowband wireless channels transmitted in parallel by OFDM, which makes it easier to deal with the problems caused by the delay extension of the broadband channel; at the same time, each narrowband channel uses MIMO technology for transmission, resulting in high spectral efficiency. In order to obtain good performance in the MIMO-OFDM system, the receiver often uses the V-BLAST detection algorithm or a similar detection algorithm. The process uses multiple matrix pseudo-inversions, and the complexity is relatively high. Due to the similarity between SC-FDE and OFDM, SC-FDE can also be combined with MIMO to form a system similar to MIMO-OFDM: MIMO-SCFDE.

In the wireless communication environment, under normal circumstances, the transmitted signal propagates through multipath, which often causes the channel to have obvious frequency selective fading characteristics. OFDM and SC-FDE systems need to estimate the channel state when performing frequency domain equalization on the signal at the receiving end. Information (CSI: Channel State Information), the accuracy of channel estimation directly affects the performance of the system; similarly, for MIMO-OFDM and MIMO-SCFDE systems, most signal detection algorithms at the receiving end also require CSI, therefore, channel estimation in MIMO -OFDM and MIMO-SCFDE systems are often required.

In a single-antenna system, channel state information is generally estimated by sending training sequences or by using some blind estimation algorithms. There are many relatively mature methods for channel estimation of MIMO-OFDM systems, which can be roughly divided into least square (LS: Least Square) algorithm, linear minimum mean square error algorithm, maximum likelihood algorithm and some blind estimation algorithms. Although MIMO-OFDM and MIMO-SCFDE systems have great similarities, they are different in channel estimation methods: (1) MIMO-OFDM systems generally do not have PAPR requirements for training sequence time domain signals, while for MIMO-OFDM systems The PAPR problem of the time-domain signal of the training sequence of the SCFDE system is often unacceptable to the system; (2) MIMO-OFDM signals generally do not have the PAPR problem in the frequency domain, while the MIMO-SCFDE signal needs to consider the influence of its frequency-domain PAPR problem on channel estimation . Therefore, in general, channel estimation methods for MIMO-OFDM systems cannot be directly used for channel estimation in MIMO-SCFDE systems.

For the MIMO-SCFDE system, there are still relatively few corresponding channel estimation algorithms. After searching the literature of the prior art, it is found that there is only one recursive reconstruction (RR: Recursive Reconstruction) algorithm, which uses a specific training sequence The structure obtains the initial value of part of the channel state information of channel estimation, and then gradually obtains all the channel state information through recursive reconstruction. Although the algorithm only uses one frame of training sequence signal, which improves the utilization rate of bandwidth, the accuracy of channel estimation is not as good as High, especially as the number of transmitting and receiving antennas increases, the accuracy of channel estimation will decrease.

Contents of the invention

The invention aims at the problems existing in the channel estimation algorithm technology of the existing MIMO-SCFDE system, and provides a channel estimation method of the MIMO-SCFDE system based on an orthogonal training sequence with high channel estimation accuracy.

The MIMO-SCFDE system channel estimation method based on the orthogonal training sequence of the present invention comprises the following implementation steps:

(1) At the sending end, first design an orthogonal training sequence according to the number of transmitting antennas, and then send the designed training sequence through each transmitting antenna in a corresponding order;

(2) Temporarily store the received signals at each receiving antenna, wait until all the training sequence signals are received, first transform to the frequency domain, and then solve the linearity according to each frequency domain sub-channel between the antennas corresponding to each transmitting and receiving Equations to obtain the state information of each frequency-domain sub-channel between the corresponding transmitting and receiving antennas;

(3) Windowing in the time domain to remove the influence of noise, transform the obtained channel state information between the corresponding transmitting and receiving antennas into the time domain, add windowing to the channel in the time domain, remove the influence of noise on the channel estimate, and then transform into the frequency domain to obtain channel estimates.

The detailed implementation method of the above steps is as follows:

Step (1), at the sending end, first design an orthogonal training sequence according to the number of transmitting antennas, and then send the designed training sequence through each transmitting antenna in a corresponding order;

The specific method of designing the training sequence according to the number of transmitting antennas is as follows:

Assuming that the number of transmitting antennas is Nt and the number of receiving antennas is Nr, in order to obtain all the frequency-domain sub-channel information between the corresponding transmitting and receiving antennas at the receiving end by solving the equation, it is necessary to Nt training sequences are sent on each antenna, and when designing the training sequence, it is necessary to ensure that the coefficient matrix is full rank when solving the linear equation system. On each frequency domain sub-channel, the training sequence data sent before and after each transmitting antenna are orthogonal to each other. If the above conditions are met, the PAPR problem of the training sequence signal in the time domain and frequency domain should also be considered when designing the training sequence, and the PAPR problem of the training sequence signal in the time domain and frequency domain should be avoided as much as possible.

Let's take Nt=2, Nr=2 as an example to illustrate:

Assuming the p-th receiving antenna, the discrete time-domain channel impulse response between the q-th transmitting antenna is h _p,q ={h _p,q;l ,l=0,1,...,L-1}, where L is the channel length in the discrete time domain; the frequency domain form obtained after the FFT transformation of the channel impulse response in the discrete time domain is Hp, q={H _{p, q; n} , n=0, 1,..., N-1}, where N is The length of one frame of SC-FDE signal;

For a system with Nt=2, in order to obtain the channel state of each sub-channel by solving the linear equations at the receiving end, it is necessary to send two frames of training sequences on each transmit antenna, and in each frequency domain sub-channel The training sequence data sent before and after different antennas are orthogonal, assuming s _{i, q} = {s _{i, q; n} , n=0, 1,..., N-1}, i=1, 2, q=1, 2, Represents the time-domain form of the i-th frame of training sequence data sent on the q-th antenna, S _{i, q} = {S _{i, q; n} , n=0, 1,..., N-1}, i=1, 2 , q=1, 2, represents the frequency domain form of the i-th frame of training sequence data sent on the qth antenna, and the training sequence data forms a training sequence matrix:

Time domain format: $\left[\begin{array}{cc}{\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout></figure-callout>}}_{\mathrm{1,1}}& {\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{1,2}}\\ {\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{2,1}}& {\mathrm{the\; s}}_{\mathrm{2,2}}\end{array}\right]$

Frequency domain form: $\left[\begin{array}{cc}{\mathrm{<figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout>}}_{\mathrm{1,1}}& {\mathrm{<figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout>}}_{\mathrm{1,2}}\\ S_{\mathrm{2,1}}& S_{\mathrm{2,2}}\end{array}\right]=\mathrm{FFT}\left\{\left[\begin{array}{cc}{\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout></figure-callout>}}_{\mathrm{1,1}}& {\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{1,2}}\\ {\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{2,1}}& {\mathrm{the\; s}}_{\mathrm{2,2}}\end{array}\right]\right\}$

In order to satisfy the orthogonality mentioned above, for example, one can make $\left[\begin{array}{cc}{\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout></figure-callout>}}_{\mathrm{1,1}}& {\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{1,2}}\\ {\mathrm{the\; <figure-callout\; id="1"\; label="s"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{2,1}}& {\mathrm{the\; s}}_{\mathrm{2,2}}\end{array}\right]=c\left[\begin{array}{cc}+\mathrm{the\; s}& +\mathrm{the\; s}\\ +\mathrm{the\; s}& -\mathrm{the\; s}\end{array}\right],$ in $c=1/\sqrt{2},$ Form an orthogonal matrix, then the frequency domain form satisfies $\left[\begin{array}{cc}{\mathrm{<figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout>}}_{\mathrm{1,1}}& {\mathrm{<figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout>}}_{\mathrm{1,2}}\\ S_{\mathrm{2,1}}& S_{\mathrm{2,2}}\end{array}\right]=c\left[\begin{array}{cc}+S& +S\\ +S& -S\end{array}\right],$ Also an orthogonal matrix;

Where s is a time-domain training sequence signal of one frame, and S is the corresponding frequency-domain form, which is actually not limited to this form, as long as the above orthogonal conditions are met.

When designing the training sequence, the PAPR problem generated by the training sequence signal in the time domain and frequency domain should also be considered, and the PAPR of the training sequence signal in the time domain and frequency domain should be made as small as possible. For example, the use of Chu sequence or Newmann sequence and in the The sequence obtained by performing various transformations on the basis, for example, can use cyclic shift to obtain a sequence with the same properties; the Chu sequence has a constant amplitude in the time domain and the frequency domain, and the Chu sequence with a length of N can be generated by the following formula :

where Z and N are mutually prime;

Newmann sequence is another sequence with constant amplitude in time domain and frequency domain, and its construction is relatively simple, which is more suitable for channel estimation of SC-FDE system. A Newmann sequence with a length of N can be generated by the following formula:

$\mathrm{the\; s}\left(\mathrm{no}\right)=e^{\mathrm{j\&pi;}{\mathrm{no}}^2/N},$ n=0,1,...,N-1

In step (2), the received signals are temporarily stored at each receiving antenna, and after all the training sequence signals are received, they are converted to the frequency domain first, and then each frequency domain is used between the corresponding antennas for each transmission and reception. The channel solves the linear equations to obtain the state information of each frequency domain sub-channel between the corresponding transmitting and receiving antennas;

The method of solving the linear equations according to each frequency domain sub-channel between the antennas corresponding to each transmission and reception is:

Assume that the signal sent by the qth transmit antenna in the i-th frame is received by the p-th receive antenna as:

r _{i, p, q} = {r _{i, p, q; n} , n=0, 1, ..., N-1}

r _{i, p} is the signal received on the pth receiving antenna in the i-th frame. Due to the effect of CP, in the discrete time domain, the linear convolution of the signal and the channel impulse response is converted into circular convolution, then:

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; \&CircleTimes;</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \}</span>$

Wherein, v _{i, p} = {v _{i, p; n} , n=0, 1, ..., N-1} is the additive white Gaussian noise contained in the signal received by the pth receiving antenna in the i frame;

Transform to the frequency domain:

R _i,p =FFT{r _i,p }={R _i,p;n ,n=0,1,...,N-1}

Circular convolution in the time domain behaves as a multiplication in the frequency domain:

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \}</span>$

Among them, V _{i, p} is the frequency domain form of additive Gaussian white noise;

Through the above relationship, the channel state information can be obtained by solving a system of equations.

Also take Nt=2, Nr=2 as an example to illustrate the method of solving the state information of the nth frequency domain sub-channel information between the two transmitting antennas and the first receiving antenna:

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="R"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="R"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">R</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">h</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">h</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="V"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="V"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">V</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="R"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">R</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">h</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 2,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">h</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="V"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">V</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right.$

Solving the system of equations directly gives:

$\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">h</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="h"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">h</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>{\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}& {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}& {\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 2,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="R"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="R"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">R</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="R"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">R</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]<span\; class="notranslate">\; -</span>{\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}& {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}& {\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 2,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="V"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="V"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">V</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="V"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">V</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]$

In fact, through the previous orthogonal design of the training sequence, the matrix orthogonality can be used to simplify the solution of the above formula. According to the matrix orthogonality theory, the inverse of the orthogonal matrix can be obtained by its conjugate rank matrix:

${\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}& {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}& {\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 2,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>{\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}"><figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}& {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="S"\; filenames="A20081013842200122.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2,1</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}& {\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 2,2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]}^{\mathrm{<span\; class="notranslate">\; h</span>}}$

Among them, A ^H represents the conjugate rank matrix of matrix A; therefore, it is not necessary to invert the matrix when solving the equation system, which greatly reduces the complexity of the solution, and at the same time, this solution does not amplify the noise; in addition, for a fixed The value of the training sequence is specific, and the calculation data to be used can be temporarily stored in advance, without repeated calculations;

The rest of the channel state information solution method is the same as the above method, and all the corresponding frequency domain sub-channel state information between the transmitting and receiving antennas will be obtained after all the solutions are completed;

Step (3), windowing in the time domain to remove the influence of noise, transform the obtained channel state information between the corresponding transmitting and receiving antennas into the time domain, and windowing the channel in the time domain to remove the influence of noise on the channel estimate , and then transformed to the frequency domain to obtain the channel estimate;

Transform the obtained channel state information between the corresponding transmitting and receiving antennas into the time domain, and the specific method of windowing the time domain channel is as follows:

Suppose h′ _{p, q} is the transformed time-domain channel impulse response, namely:

h′ _{p, q} = IFFT{H _{p, q} }

h″ _{p, q} represent the time-domain channel impulse response after windowing, which can be obtained by the following formula:

h″ _{p, q} = h′ _{p, q} W

Among them, W is a windowing matrix, which is a diagonal matrix, and the diagonal elements are 1 within the channel delay extension range, and the rest are all 0;

Then transform to the frequency domain through FFT:

H′ _{p, q} = FFT{h″ _{p, q} }

H' _{p, q} is the estimated value of the channel to be obtained.

The channel delay spread can be obtained by an estimation method, which is a well-known technology in the prior art, and related documents can be referred to, so it will not be discussed here.

The MIMO-SCFDE system channel estimation method proposed by the invention is based on an orthogonal training sequence, is suitable for the characteristics of the MIMO-SCFDE system, has low complexity, and can improve channel estimation accuracy.

Description of drawings

Figure 1 is a block diagram of a system using the method of the present invention.

Fig. 2 is a block diagram of the channel estimation method of the present invention.

FIG. 3 is a comparison diagram of mean square error curves between the channel estimation method of the present invention and the RR channel estimation method.

Fig. 4 is a comparison diagram of the channel estimation method of the present invention and the system bit error rate performance curve under ideal conditions.

Detailed ways

Example

The embodiment gives the baseband simulation result, and does not consider the influence of the synchronization error, that is, the synchronization is an ideal synchronization; the training sequence in the embodiment adopts a Newmann sequence.

Fig. 1 has provided the MIMO-SCFDE system block diagram that uses the method proposed in the present invention, send data first through serial/parallel conversion process, after adding CP, D/A conversion, become radio frequency signal and send out through each transmitting antenna; Each antenna at the receiving end performs down-conversion, A/D conversion, and de-CP processing on the received signal. If it is a training sequence signal, it is sent to the channel estimation module to estimate the channel state information. If it is a data frame, it is transformed into the frequency domain by FFT. , and then equalize the channel state information obtained by channel estimation, IFFT transformation to the time domain to complete signal detection, and obtain the original transmission data after parallel/serial transformation.

Fig. 2 has provided the block diagram of the channel estimation method that the present invention proposes, on each antenna of the receiving end, the received training sequence signal is sent to solve the initial channel state information module after going through CP, FFT transformation, and solves the channel estimation initial value ; Then IFFT transforms to the time domain, removes the influence of noise through time domain windowing, and then obtains the final channel estimation value through FFT transformation.

The simulation parameters of this embodiment:

Simulation environment: MATLAB R2007a

Total number of sub-channels: N=256

CP length: 32

Number of transmitting antennas: Nt=2

Number of receiving antennas: Nr=3

Modulation method: QPSK

The range of the average received signal-to-noise ratio selected by simulation: SNR＝4～16(dB)

Simulation channel environment: between each transmitting antenna and each receiving antenna is a broadband multipath channel, the number of discretized channel paths is CP length, and each path is a complex Gaussian variable that obeys independent and identical distribution, with a mean of 0 and a variance of 1 .

Fig. 3 shows the baseband simulation results of the mean square error (MSE: Mean Square Error) using the channel estimation method of the present invention and the RR channel estimation method, from which it can be seen that: under the same signal-to-noise ratio, the channel estimation method of the present invention The accuracy is higher than the RR method.

Fig. 4 has provided and used the system bit error rate (BER:Bit Error Rate) performance comparison of the present invention, did not use coding in the emulation, can find out from this, use the system BER performance of the channel estimation method of the present invention to compare ideally The difference is small, for example, when BER=10^(-4), the SNR difference is less than 0.5dB, therefore, the channel estimation of the method of the present invention has less SNR loss to the system.

To avoid confusion, some terms mentioned in this manual are explained as follows:

1. Symbol: refers to the data after information bits are modulated and mapped (also called symbol mapping). Usually a complex number whose real and imaginary parts are integers.

2. One frame signal: For a single-antenna SC-FDE system, one frame signal refers to N information symbols between two adjacent CPs at the sending end, and N information symbols that are transformed by FFT after removing the CP at the receiving end. symbols; for the MIMO-SCFDE system, a frame signal at the transmitting end refers to N information symbols between two CPs on each transmitting antenna, and at the receiving end refers to the FFT transformation after removing the CP on each receiving antenna N symbols. At the sending end, when sending each frame signal on each transmitting antenna of the MIMO-SCFDE system, CP must be added according to the signal processing method of the SC-FDE system; at the receiving end, each frame signal on each receiving antenna must be processed according to SC-FDE The system signal processing method goes to CP.

3. Sub-channel: For SC-FDE discrete baseband signals, a sub-channel refers to a frequency point after FFT at the receiving end. For a radio frequency channel, a subchannel refers to a section of frequency spectrum of the radio frequency channel.

Claims (4)

1. a MIMO-SCFDE system channel estimation method based on orthogonal training sequences, is characterized in that: comprise following realization steps: (1) At the sending end, first design an orthogonal training sequence according to the number of transmitting antennas, and then send the designed training sequence through each transmitting antenna in a corresponding order; (2) Temporarily store the received signals at each receiving antenna, wait until all the training sequence signals are received, first transform to the frequency domain, and then solve the linearity according to each frequency domain sub-channel between the antennas corresponding to each transmitting and receiving Equations to obtain the state information of each frequency-domain sub-channel between the corresponding transmitting and receiving antennas; (3) Windowing in the time domain to remove the influence of noise, transform the obtained channel state information between the corresponding transmitting and receiving antennas into the time domain, add windowing to the time domain channel, remove the influence of noise on the channel estimate, and then transform into the frequency domain to obtain channel estimates. 2. according to the described MIMO-SCFDE system channel estimation method based on orthogonal training sequence of claim 1, it is characterized in that: in the described step (1), the concrete method according to the number design training sequence of transmitting antenna is: Assuming that the number of transmitting antennas is Nt and the number of receiving antennas is Nr, in order to obtain all the frequency-domain sub-channel information between the corresponding transmitting and receiving antennas at the receiving end by solving the equation, it is necessary to Nt training sequences are sent on each antenna, and when designing the training sequence, it is necessary to ensure that the coefficient matrix is full rank when solving the linear equation system. On each frequency domain sub-channel, the training sequence data sent before and after each transmitting antenna are orthogonal to each other. If the above conditions are met, the PAPR problem in the time domain and frequency domain of the training sequence signal should be avoided when designing the training sequence. 3. according to the described MIMO-SCFDE system channel estimation method based on orthogonal training sequence of claim 1, it is characterized in that: in the described step (2), according to each frequency domain sub-channel solution between the corresponding antennas of each transmitting and receiving The method for a system of linear equations is: Assume that the signal sent by the qth transmit antenna in the i-th frame is received by the p-th receive antenna as: r _{i, p, q} = {r _{i, p, q; n} , n=0, 1, ..., N-1} r _{i, p} is the signal received on the pth receiving antenna in the i-th frame. Due to the effect of CP, in the discrete time domain, the linear convolution of the signal and the channel impulse response is converted into circular convolution, then: ${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; \&CircleTimes;</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \}</span>$ Wherein, v _{i, p} = {v _{i, p; n} , n=0, 1, ..., N-1} is the additive white Gaussian noise contained in the signal received by the pth receiving antenna in the i frame; Transform to the frequency domain: R _i,p =FFT{r _i,p }={R _i,p;n ,n=0,1,...,N-1} Circular convolution in the time domain behaves as a multiplication in the frequency domain: ${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}{<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \}</span>$ Among them, V _{i, p} is the frequency domain form of additive Gaussian white noise; Through the above relationship, the channel state information can be obtained by solving a system of equations. 4. according to the described MIMO-SCFDE system channel estimation method based on orthogonal training sequence of claim 1, it is characterized in that: in the described step (3), the channel state information between the corresponding transmitting and receiving antennas obtained is converted to time domain, the specific method of windowing the time domain channel is as follows: Suppose h′ _{p, q} is the transformed time-domain channel impulse response, namely: h′ _{p, q} = IFFT{H _{p, q} } h″ _{p, q} represent the time-domain channel impulse response after windowing, which can be obtained by the following formula: h″ _{p, q} = h′ _{p, q} W Among them, W is a windowing matrix, which is a diagonal matrix, and the diagonal elements are 1 within the channel delay extension range, and the rest are all 0; Then transform to the frequency domain through FFT: H′ _{p, q} = FFT{h″ _{p, q} } H' _{p, q} is the estimated value of the channel to be obtained.

Images