CN101136884B - Channel Estimation Method for TDS-OFDM System - Google Patents

Abstract

The invention discloses a channel estimation method used in a TDS-OFDM system, belonging to the technical field of digital information transmission. The method includes: using the channel estimation result obtained in the last iteration of the previous frame or the last iteration of the current frame as the initial channel impulse response of the iteration of the current frame, and removing the interference of the pseudo-random sequence on the data block according to the initial channel impulse response , and perform cyclic reconstruction of the data block; perform equalization on the cyclic reconstruction result, and perform partial judgment on the equalized result; remove the interference of the data block on the pseudo-random sequence according to the partial judgment result and the initial channel impulse The cyclic reconstruction of the sequence; re-estimate the channel impulse response according to the cyclic reconstruction result, and use the channel impulse response as the channel estimate of the current frame iteration. The invention assists the channel estimation by making partial judgment on the data blocks of the TDS-OFDM system, improves the precision of the channel estimation and reduces the complexity of the operation.

Description

Channel Estimation Method for TDS-OFDM System

technical field

The invention relates to the technical field of digital information transmission, in particular to a channel estimation method for a TDS-OFDM system.

Background technique

The main problem to be solved by wireless communication technology at present is how to increase the transmission rate reliably within a limited bandwidth. OFDM (Orthogonal Frequency Division Multiplexing, Orthogonal Frequency Division Multiplexing) can reliably achieve high-speed transmission in frequency-selective fading channels, and has been widely used in wireless communications such as wireless local area networks, fixed wireless access, digital audio and video broadcasting system. There are currently two OFDM systems, one is based on CP (Cyclic Prefix, cyclic prefix) CP-OFDM system, which uses CP to fill the guard interval, and the guard interval is located between two OFDM symbols to resist the delay spread in the channel . During the transmission process of the CP-based OFDM system, the linear convolution relationship between the OFDM symbol and the channel impulse response is converted into a circular convolution relationship, so that the received signal can be equalized by using a simple frequency domain equalization technique in the receiver; One is a ZP-OFDM system based on ZP (Zero Padding, zero padding). The zero symbol replaces the CP in the CP-OFDM system, so that the ZP-OFDM system can still be guaranteed when the channel response spectrum is zero due to deep channel fading. OFDM system realizes symbol recovery, see (Muquet B, Wang Z, Giannakis G.B, Courville M.de, and Duhamel P, Cyclic Prefixing or Zero Padding for Wireless Multicarrier Transmissions, IEEETrans.on Communications, 2002, 50(12): 2136- 2148). In coherently demodulated CP-OFDM and ZP-OFDM systems, the channel impulse response is generally estimated by means of pilot symbols (corresponding to continuous stream transmission mode) or preamble sequences (corresponding to burst transmission mode).

In the above-mentioned CP-OFDM and ZP-OFDM systems, the CP and ZP symbols, together with the pilot frequency and preamble sequence, cause the reduction of spectrum efficiency. In order to make better use of the guard interval, a Chinese invention patent ZL01124144.6 "Guard Interval Filling Method for Orthogonal Frequency Division Multiplexing System" proposes an OFDM system based on TDS (Time-Domain Synchronous, time-domain synchronization). In the TDS-OFDM system, the known pseudo-random sequence is used to fill the guard interval between OFDM symbols, and the filled pseudo-random sequence can be used for receiver synchronization and channel estimation in addition to being used as a guard interval. Referring to Fig. 1, it is the frame structure figure of TDS-OFDM system, and TDS-OFDM signal frame comprises the pseudo-random sequence of OFDM data block and padding, the time-domain OFDM data block {x _n ^{( i)} } _n=0 ^N-1 is followed by a pseudo-random sequence padding section {c _n ⁽ⁱ⁾ } _n=0 ^v-1 of length v, thus forming the i-th TDS- with length N ₂ =N+v OFDM transmit signal frame {s _n ⁽ⁱ⁾ } _n=0 ^N-1 , it is assumed that the filled pseudo-random sequence is modulated by BPSK (Binary Phase Shift Keying, Binary Phase Shift Keying) and has the following autocorrelation characteristics:

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; m</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; \&ap;</span>\mathrm{<span\; class="notranslate">\; v\&delta;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span>$

Among them, δ(n) is a discrete Dirac function. It is assumed that the adopted channel model is a quasi-static L-order FIR (Finite Impulse Response, finite impulse response) filter, and is interfered by additive Gaussian white noise. Then the TDS-OFDM received signal frame {r _n ⁽ⁱ⁾ } _n=0 ^N-1 and the TDS-OFDM transmitted signal frame {s _n ⁽ⁱ⁾ } _n=0 ^N-1 satisfy the linear convolution relationship:

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; l</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="N"\; filenames="H2007101753018E00032.png,H2007101753018E00033.png"\; state="\{\{state\}\}">N</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}$

Among them, w _n ⁽ⁱ⁾ is additive Gaussian white noise. If the linear convolution result p _n of the filled pseudo-random sequence {c _n ⁽ⁱ⁾ } _n=0 ^v-1 and the channel impulse response {h _l ⁽ⁱ⁾ } _l=0 ^L-1 can be obtained in the receiver ⁽ⁱ⁾ , namely

${\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; l</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>$

Then the channel impulse response can be estimated by direct linear correlation method in time domain. Cross-correlation between the received signal p _n ⁽ⁱ⁾ and the local known sequence {c _n ⁽ⁱ⁾ } _n=0 ^v-1 can be obtained

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; pc</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; m</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ]</span>}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; v</span>}\underset{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; m</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ]</span>}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; .</span>$

Thus the channel impulse response can be estimated by the following formula

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}_{\mathrm{<span\; class="notranslate">\; l</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; v</span>}}{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; pc</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$

Or, if the circular convolution result of the filled pseudo-random sequence {c _n ⁽ⁱ⁾ } _n=0 ^v-1 and the channel impulse response {h _l ⁽ⁱ⁾ } _l=0 ^L-1 can be obtained in the receiver q _n ⁽ⁱ⁾ , namely

$q_{\mathrm{no}}^{\left(i\right)}=\underset{l=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{h}_l^{\left(i\right)}{c}_{{(\mathrm{no}-l)}_{N_2}}^{\left(i\right)}+w_{\mathrm{no}}^{\left(i\right)},$ 0≤n<v,

Then the channel impulse response can be estimated by computation of circular convolution. There are two calculation methods for circular convolution, one is the time-domain direct circular correlation method, and the other is the frequency-domain FFT-assisted method. The calculation of the circular convolution assisted by FFT (Fast Fourier Transform) is as follows:

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; IFFT</span>}}^{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; [</span>\frac{{\mathrm{<span\; class="notranslate">\; FFT</span>}}^{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ]</span>}{{\mathrm{<span\; class="notranslate">\; FFT</span>}}^{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ]</span>}<span\; class="notranslate">\; ]</span>$

Wherein, FFT ^M [·] and IFFT ^M [·] represent FFT and IFFT (Inverse Fast Fourier Transform, Inverse Fast Fourier Transform) of length M respectively, where M=v.

In the TDS-OFDM system, the guard interval filled with pseudo-random sequence replaces the traditional guard interval filled with CP or ZP. When the TDS-OFDM signal passes through the frequency selective fading channel, in the TDS-OFDM frame {s _n ⁽ⁱ⁾ } _n=0 ^N-1 , the OFDM data block {x _n ⁽ⁱ⁾ } _n=0 ^N-1 will Interference filled pseudo-random sequence {c _n ⁽ⁱ⁾ } _n=0 ^N-1 , and vice versa, and interference between the stuffing sequence and OFDM data blocks also exists in the TDS-OFDM frame. The ideal linear convolution result of the pseudo-random sequence is defined as the linear convolution result that completely removes the influence of the OFDM data block. The ideal circular convolution result of the pseudo-random sequence is defined as the circular convolution result that completely removes the influence of the OFDM data block. Since the filled pseudo-random sequence is known to the receiver, it is easy to eliminate the interference of the pseudo-random sequence to the OFDM data block when the channel impulse response is known; on the contrary, since the transmitted OFDM data block is random , and may be interfered by noise, even in the case of known channel impulse response, it is difficult to eliminate the interference of OFDM data blocks to the filled pseudo-random sequence, so that the transmitted pseudo-random sequence and The ideal linear convolution result or the ideal circular convolution result of the channel impulse response, that is, a more accurate channel estimation cannot be obtained.

Chinese invention patent 200510012127.6 "An Iterative Elimination Method for Pseudo-random Sequence Filling in OFDM Modulation System" proposes a method of removing the filled pseudo-random sequence by using iterative elimination method. The method first calculates the relationship between the pseudo-random sequence and the channel impulse response Linear convolution, and then remove the influence of the pseudo-random sequence on the data block, then perform frequency domain equalization on the data block, and finally remove the influence of the data block on the pseudo-random sequence, so as to obtain more accurate channel estimation. However, when using this method for channel estimation, it is necessary to know the channel estimation of the previous two consecutive OFDM symbols, and multiple iterations are required for each OFDM symbol to obtain a more accurate channel estimation, and the calculation is complicated.

Contents of the invention

In order to reduce the computational complexity of channel estimation in the TDS-OFDM system, the present invention provides a channel estimation method for the TDS-OFDM system. Described technical scheme is as follows:

The present invention provides a channel estimation method for a TDS-OFDM system, the method comprising:

Step A: The channel estimate obtained in the last iteration of the previous frame or the last iteration of the current frame is used as the initial channel impulse response of the current iteration of the current frame, and the pseudo-random sequence pair in the current frame is removed according to the initial channel impulse response Interference of data blocks, and cyclic reconstruction of said data blocks;

Step B: Equalize the cyclic reconstruction result, and make a partial judgment on the equalized result;

Step C: removing the interference of the data block on the pseudo-random sequence according to the result of the partial decision and the initial channel impulse response, and performing cyclic reconstruction of the pseudo-random sequence;

Step D: re-estimate the channel impulse response according to the cyclic reconstruction result of the pseudo-random sequence without data block interference, and use the channel impulse response as the channel estimate of the current frame iteration.

In the step A, according to the initial channel impulse response, the interference of the pseudo-random sequence to the data block is removed, and the specific steps of performing the cyclic reconstruction of the data block are as follows:

calculating the circular convolution of the pseudo-random sequence and the initial channel impulse response, the result of the circular convolution is the interference of the pseudo-random sequence to the data block;

Constructing the circular convolution of the data block and the channel impulse response of the current iteration according to the current frame and the next frame of the current frame;

The interference of the pseudo-random sequence on the data block is removed from the circular convolution result of the data block and the initial channel impulse response to obtain a circular reconstruction result of the data block.

The step of equalizing the cyclic reconstruction result in the step B specifically includes:

performing FFT transformation on the cyclic reconstruction result;

performing FFT transformation on the initial channel impulse response, and then performing a conjugate operation element by element;

Dot producting the result of the conjugate operation and the result of performing FFT transformation on the cyclic reconstruction result to obtain an equalized result.

The step of partially judging the equalized result in the step B specifically includes:

performing FFT transformation on the initial channel impulse response, and calculating the square of the absolute value of the transformation result;

Part of the in-phase component and the quadrature component are respectively judged on the equalized result according to the square result, the constellation symbol and the optimal decision interval parameter in the constellation diagram space.

Described step C specifically comprises:

calculating the circular convolution of the data block and the initial channel impulse response, the result of the circular convolution being the interference of the data block to the pseudo-random sequence;

Constructing the circular convolution of the pseudo-random sequence and the channel impulse response of this iteration according to the current frame and the next frame of the current frame;

The interference of the data block on the pseudo-random sequence is removed from the circular convolution result of the pseudo-random sequence and the channel impulse response of the current iteration, to obtain a circular reconstruction result of the pseudo-random sequence.

Described step D specifically comprises:

performing FFT transformation on the cyclic reconstruction result of the pseudo-random sequence from which data block interference has been removed;

performing FFT transformation on the pseudo-random sequence;

dividing the FFT transformation result of the cyclic reconstruction result by the FFT transformation result of the pseudo-random sequence, performing IFFT transformation on the division result to obtain a new channel impulse response, and converting the new channel impulse response as the channel estimate of the current iteration of the current frame.

The beneficial effects of the technical solution provided by the invention are:

The present invention reduces the operation length of FFT/IFFT by constructing circular convolution for interference removal and channel estimation; through partial decision-assisted channel estimation, the convergence speed of iteration is accelerated; and by using partial decision in channel estimation, the reduction of The complexity of the algorithm makes the channel estimation result more accurate.

Description of drawings

FIG. 1 is a frame structure diagram of a TDS-OFDM system in the prior art;

Fig. 2 is the baseband model figure of the TDS-OFDM system that the embodiment of the present invention provides;

FIG. 3 is a schematic diagram of an iterative flow chart of a channel estimation method for a TDS-OFDM system provided by an embodiment of the present invention;

FIG. 4 is a flowchart of a channel estimation method for a TDS-OFDM system provided by an embodiment of the present invention;

FIG. 5 is a schematic diagram of a partial decision interval of a QPSK signal provided by an embodiment of the present invention;

FIG. 6 is a schematic diagram of a partial decision interval of a 16-QAM signal provided by an embodiment of the present invention;

Fig. 7 is a schematic diagram of the variation of the channel estimation error simulated under channel 1 with the number of iterations provided by the embodiment of the present invention;

Fig. 8 is a schematic diagram of the variation of channel estimation error with the number of iterations simulated under channel II provided by an embodiment of the present invention;

FIG. 9 is a BER simulation curve diagram of a TDS-OFDM system using a QPSK constellation diagram provided by an embodiment of the present invention under a multipath channel;

FIG. 10 is a BER simulation graph of a TDS-OFDM system using a 16-QAM constellation diagram provided by an embodiment of the present invention under a multipath channel.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below in conjunction with the accompanying drawings.

When channel estimation is performed in the embodiment of the present invention, circular convolution is constructed, and partial judgment is performed on data blocks of TDS-OFDM frames to assist channel estimation, thereby improving convergence speed of iterations and reducing computational complexity.

In the embodiment of the present invention, the TDS-OFDM frame generally refers to the TDS-OFDM received signal frame, and the pseudo-random sequence filled in each transmitted signal frame in the TDS-OFDM signal is the same. Referring to Figure 2, it is a baseband model diagram of a TDS-OFDM system, in which S/P and P/S represent serial/parallel conversion and parallel/serial conversion respectively, and AWGN (Additive White Gaussian Noise) represents additive white Gaussian noise. In the figure, the i-th OFDM frequency-domain transmission data block {X _k ⁽ⁱ⁾ } _k=0 ^N-1 is first subjected to IFFT transformation to obtain the i-th OFDM data block {x _n ⁽ⁱ⁾ } _n=0 ^N-1 , As shown in formula (1), where N represents the length of IFFT.

$x_{\mathrm{no}}^{\left(i\right)}=\frac{1}{\sqrt{N}}\underset{k=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{x}_k^{\left(i\right)}{e}^{\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="H2007101753018E00032.png,H2007101753018E00033.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;nk}/N},$ 0≤n<N; (1)

如式(2)所示，其中N₂＝N+v。Secondly, a fixed pseudo-random sequence {c _n } _n=0 ^v-1 is inserted after each OFDM data block to be transmitted {x _n ⁽ⁱ⁾ } _n=0 ^N-1 to form the i-th TDS-OFDM transmit signal frame

As shown in formula (2), wherein N ₂ =N+v.

${\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; N</span>}\\ {\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; N</span>}}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

表示l对N₂取模。Assume that the adopted channel model is a quasi-static L-order FIR filter and is interfered by AWGN noise w _n ⁽ⁱ⁾ . Assuming that the length of the filled pseudo-random sequence is not less than the channel length, that is, v≥L-1, then corresponding to the i-th TDS-OFDM transmission signal frame The TDS-OFDM received signal frame is r _n ⁽ⁱ⁾ , as shown in formula (3), where h _l ⁽ⁱ⁾ is the channel impulse response,

Indicates that l is modulo _N2 .

$r_{\mathrm{no}}^{\left(i\right)}=\underset{l=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{h}_l^{\left(i\right)}{\mathrm{the\; s}}_{{(\mathrm{no}-l)}_{N_2}}^{\left(i\right)}+w_{\mathrm{no}}^{\left(i\right)},$ 0≤n<N ₂ (3)

Since each TDS-OFDM transmit signal frame is filled with a fixed pseudo-random sequence, in a frequency selective fading channel, the pseudo-random sequence in the i-1th TDS-OFDM frame is In other words, it is equivalent to the CP of the CP-OFDM system, so that the linear convolution relationship between the transmitted signal and the channel impulse response can be converted into the circular convolution relationship described by formula (3).

Referring to FIG. 3 , the channel estimation method for the TDS-OFDM system provided by the embodiment of the present invention is performed in an iterative form, and uses partial judgment on data blocks to assist channel estimation. In the TDS-OFDM system described above On the basis, referring to FIG. 4, an embodiment of the present invention provides a channel estimation method for a TDS-OFDM system, which specifically includes the following steps:

Step 101: Initialize the channel impulse response, set the iteration label I and the number of iterations J.

Among them, the specific process of initializing the channel impulse response is as follows:

时，若信道估计器刚开始工作，即当i＝1时，信道冲激响应初始化为全零；否则，将上一帧最后一次或当前帧上一次迭代得到的信道估计结果作为本次迭代的初始信道冲激响应，即

或

The channel impulse response of the i-th TDS-OFDM frame at initialization

, if the channel estimator just starts to work, that is, when i=1, the channel impulse response is initialized to all zeros; otherwise, the channel estimation result obtained in the last iteration of the previous frame or the last iteration of the current frame is used as the initial channel impulse response of this iteration, that is,

or

The iteration label I is set to 0 during initialization, that is, I=0; the number of iterations J≥1.

Step 102: Remove the interference of the pseudo-random sequence on the data block in the current frame according to the initial channel impulse response, and perform cyclic reconstruction of the data block.

The specific process is as follows:

的循环卷积

即为伪随机序列c_n对数据块的干扰。循环卷积

的频域计算方法如式(4)所示，其中⊙表示点乘(逐元素相乘)。1) Calculate the pseudo-random sequence c _n and the initial channel impulse response

circular convolution of

That is, the interference of the pseudo-random sequence c _n on the data block. circular convolution

The frequency domain calculation method of is shown in formula (4), where ⊙ represents dot multiplication (element-by-element multiplication).

^{2 ) Construct} _{the circular} ^{convolution result} _{y n (} ⁱ⁾ , as shown in formula (5), the channel impulse response h _n ⁽ⁱ⁾ represents the channel impulse response of this iteration, and the loop result includes the interference of the pseudo-random sequence c _n on the data block x _n ⁽ⁱ⁾ .

${\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; N</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; L</span>}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; N</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

或 3) Remove the interference of the pseudo-random sequence c _n on the data block x _n ⁽ i) in the above circular convolution result y _n ⁽ⁱ⁾ As shown in formula (6), the cyclic reconstruction result of data block x _n ⁽ⁱ⁾ is obtained

or

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; q</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; L</span>}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; N</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

Compared with linear convolution, the FFT/IFFT operation length of circular convolution is small, so in the above steps, by constructing the circular convolution of the pseudo-random sequence and the initial channel impulse response and reconstructing the loop of the data block and the channel impulse response Convolution reduces the operation length of FFT/IFFT, improves the convergence speed of iterations, and reduces the operation complexity.

进行均衡，然后对均衡的结果进行部分判决。Step 103: cyclic reconstruction result of data block x _n ⁽ⁱ⁾

Perform equalization, and then make partial judgments on the result of the equalization.

进行相位均衡。对

进行相位均衡的过程具体如下：Equalization includes phase equalization and amplitude equalization, and the equalization in this embodiment refers to

Perform phase equalization. right

The process of phase equalization is as follows:

和初始信道冲激响应进行FFT变换，得到

和

即， ${\hat{Y}}_k^{\left(i\right)}={\mathrm{FFT}}^N\left[{\hat{y}}_n^{\left(i\right)}\right],$ ${\hat{H}}_k^{\left(i\right)}={\mathrm{FFT}}^N\left[{\hat{h}}_n^{\left(i\right)}\right].$ 1) Respectively for the loop reconstruction results

and the initial channel impulse response Perform FFT transformation to get

and

Right now, ${\hat{Y}}_k^{\left(i\right)}={\mathrm{FFT}}^N\left[{\widehat{\mathrm{the\; y}}}_{\mathrm{no}}^{\left(i\right)}\right],$ ${\hat{h}}_k^{\left(i\right)}={\mathrm{FFT}}^N\left[{\hat{h}}_{\mathrm{no}}^{\left(i\right)}\right].$

按元素进行共轭运算，将

按元素进行共轭运算的结果和

进行点乘，得到

均衡的结果

如式(7)所示，其中(·)^*表示按元素进行共轭运算。2 pairs

Conjugate element-wise, the

The result of the element-wise conjugate operation and

Do point multiplication to get

balanced result

As shown in Equation (7), where (·) ^* represents the element-wise conjugation operation.

进行部分判决的具体过程如下：The result of the above equilibrium

The specific process of making partial judgments is as follows:

进行部分判决，其中

表示

部分判决的结果，

X表示星座符号，Θ₁和Θ₂表示星座图空间的划分；可以看出，在区间Θ₁对

进行硬判决，在区间Θ₂对

进行零判决。According to formula (8) for the result of equilibrium

part of the judgment, of which

express

the results of part of the judgment,

X represents the constellation symbol, and Θ ₁ and Θ ₂ represent the division of the constellation diagram space; it can be seen that in the interval Θ ₁ pair

Make a hard decision, in the interval Θ ₂ pairs

Make zero verdicts.

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; min</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; |</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Z</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&eta;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; ,</span>& {\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Z</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; \&Theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ \mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>& {\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Z</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; \&Theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

进行部分判决是通过对的同相分量和正交分量分别进行判决得到的，具体过程如下：假设星座符号X的同相分量与正交分量均取值于符号集合[-1，+1]，对

的同相分量按照式(9)进行部分判决。Further, for the QPSK (Quadrature Phase Shift Keying, Quadrature Phase Shift Keying) constellation diagram, for

Part of the judgment is made by the The in-phase component and the quadrature component of the constellation symbol X are determined separately, and the specific process is as follows: Assume that the in-phase component and the quadrature component of the constellation symbol X both take values from the symbol set [-1, +1], for

in-phase component of Part of the judgment is made according to formula (9).

部分判决的结果，且

是

的同相分量。参见图5，为QPSK信号的部分判决区间。对

的正交分量进行部分判决的方法与同相分量相同。Among them, α represents the best decision interval parameter in the QPSK constellation space, 0<α<1, express

the result of a partial judgment, and

yes

in-phase component of . Referring to FIG. 5 , it is a partial decision interval of a QPSK signal. right

The quadrature component of the partial judgment method is the same as that of the in-phase component.

进行部分判决也是通过对

的同相分量和正交分量分别进行判决得到的，具体过程如下：假设星座符号X的同相分量与正交分量均取值于符号集合[-3，-1，+1，+3]，对的同相分量

按照式(10)进行部分判决。Further, for 16-QAM (QuadratureAmplitude Modulation, quadrature amplitude modulation) constellation diagram, for

Part of the judgment is also passed on

The in-phase component and the quadrature component of the constellation symbol X are determined separately, and the specific process is as follows: Assume that the in-phase component and the quadrature component of the constellation symbol X both take values from the symbol set [-3, -1, +1, +3], for in-phase component of

Part of the judgment is made according to formula (10).

表示

部分判决的结果，且

是

的同相分量。参见图6，为16-QAM信号的部分判决区间。对

的正交分量进行部分判决的方法与同相分量相同。Among them, β represents the best decision interval parameter in the 16-QAM constellation space, 0<β<1,

express

the result of a partial judgment, and

yes

in-phase component of . Referring to FIG. 6 , it is a partial decision interval of a 16-QAM signal. right

The quadrature component of the partial judgment method is the same as that of the in-phase component.

Step 104: Remove the interference of the data block x _n ⁽ⁱ⁾ on the pseudo-random sequence c _n , and perform cyclic reconstruction of the pseudo-random sequence c _n .

The specific process is as follows:

计算数据块x_n ⁽ⁱ⁾和初始信道冲激响应

的循环卷积

即数据块x_n ⁽ⁱ⁾对伪随机序列c_n的干扰。循环卷积

的频域计算方法如式(11)所示：1) According to the partial decision result of the data block x _n ⁽ⁱ⁾ obtained in step 103

Compute data block x _n ⁽ⁱ⁾ and initial channel impulse response

circular convolution of

That is, the interference of the data block x _n ⁽ⁱ⁾ to the pseudo-random sequence c _n . circular convolution

The frequency domain calculation method of is shown in formula (11):

2) Construct the circular convolution result y _n ⁽ⁱ⁾ of the pseudo-random sequence c _n and the channel impulse response h _n ^{( i) according to the current frame r n (i) and the next frame r n+N (i )} _{of the} current frame , as shown in formula (12), the result includes the interference of the data block x _n ⁽ⁱ⁾ on the pseudo-random sequence c _n .

${\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; N</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 12</span>}<span\; class="notranslate">\; )</span>$

如式(13)所示，得到伪随机序列的循环重建结果

3) Remove the interference of the data block x _n ⁽ⁱ⁾ on the pseudo-random sequence c _n from the circular convolution result y _n ⁽ⁱ⁾

As shown in formula (13), the cyclic reconstruction result of the pseudo-random sequence is obtained

${\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; q</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; L</span>}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; v</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 13</span>}<span\; class="notranslate">\; )</span>$

Step 105: Update channel estimation.

重新估计信道冲激响应，并将重新估计后得到的信道冲激响应

作为当前帧本次迭代的信道估计结果。信道冲激响应

迫零算法的频域计算方法如式(14)所示。The result of the cyclic reconstruction from the pseudorandom sequence c _n

Re-estimate the channel impulse response, and the re-estimated channel impulse response

As the channel estimation result of this iteration of the current frame. channel impulse response

The frequency domain calculation method of the zero-forcing algorithm is shown in formula (14).

${\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; IFFT</span>}}^{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; FFT</span>}}^{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; [</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; q</span>}}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; /</span>{\mathrm{<span\; class="notranslate">\; FFT</span>}}^{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 14</span>}<span\; class="notranslate">\; )</span>$

Step 106: Let I=I+1, judge whether the iteration number I is equal to the preset iteration number J, if yes, end the channel estimation; otherwise, return to step 101.

Based on the above description, a computer simulation is performed on the channel estimation method for the TDS-OFDM system provided by the embodiment of the present invention. See Table 1 for the system parameters used in the simulation. Two multipath channel models are used in the simulation: channel I and channel II, see Table 2 for the multipath channel parameters used in the simulation. Assuming that the two multipath channels are subject to independent Rayleigh fading, and the channel impulse response is normalized so that the power of the signal after passing through the multipath channel remains unchanged, the decision interval parameter of the QPSK constellation is α=0.1875 , the decision interval parameter of the 16-QAM constellation diagram is β=0.0938.

Table 1

name value The number of subcarriers (FFT length) N 3780 Subcarrier Constellation Diagram 4-QAM, 16-QAM Baseband symbol rate 7.56MHz Pseudo-random sequence length v 420

Table 2

Referring to FIG. 7 , it is a graph showing the variation of the channel estimation error with the number of iterations simulated under channel I; see FIG. 8 , which is a graph showing the variation of the channel estimation error with the number of iterations simulated under channel II. Both channel I and channel II are assumed to be static. The channel estimation error ξ is defined as:

$\mathrm{<span\; class="notranslate">\; \&xi;</span>}<span\; class="notranslate">\; =</span>\sqrt{\underset{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; `</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; |</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}}_{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

和h_l分别表示估计的信道冲激响应和真实的信道冲激响应。从图7和图8可以看出，经过3至4次的迭代运算，信道估计误差几乎就不减小了。在准静态或者是缓慢变化的信道中，如果将前一个符号的信道冲激响应作为下一个符号的初始信道冲激响应，可以在无需迭代的情况下获得较准确的信道冲激响应估计性能。in,

and h _l respectively represent the estimated channel impulse response and the real channel impulse response. It can be seen from FIG. 7 and FIG. 8 that after 3 to 4 iterations, the channel estimation error hardly decreases. In a quasi-static or slowly changing channel, if the channel impulse response of the previous symbol is used as the initial channel impulse response of the next symbol, more accurate channel impulse response estimation performance can be obtained without iteration.

FIG. 9 is a BER (Bit Error Rate, bit error rate) simulation curve of a TDS-OFDM system using a QPSK constellation diagram under a multipath channel. The horizontal axis in the figure is the signal-to-noise ratio E _b /N ₀ of each bit, The maximum Doppler frequency shift is set to 18Hz, so that the channel is approximately constant within a TDS-OFDM symbol, but the channel changes from one TDS-OFDM symbol to the next; 900 consecutive TDS-OFDM symbols The bit error rate is calculated by simulation, but the bit error rate of the first two TDS-OFDM symbols is not counted; the minimum lower bound in the figure represents the bit error rate under the ideal channel estimation situation. As can be seen from the figure, under channel I and channel II, using the channel estimation method provided in this embodiment, the bit error rate of the TDS-OFDM system obtained can approach the minimum without iteration (J=0) Nether.

Fig. 10 is a BER simulation graph of a TDS-OFDM system using a 16-QAM constellation diagram under a multipath channel, from which a conclusion similar to that of a QPSK system can be obtained.

In the embodiment of the present invention, the channel estimation is performed by constructing the circular convolution of the data block and the channel impulse response, so that the length of the FFT/IFFT operation is minimized; the channel estimation is assisted by partial judgment with a small delay, which is convenient to speed up the convergence speed of the iteration; and in Partial decisions are used in channel estimation to make the estimation results more accurate. This method can make the system bit error rate reach the minimum lower bound without iteration in the slowly changing channel environment.

The specific examples of the present invention have been described in detail above, but the present invention is not limited to the above examples. Within the spirit and principles of the claims of the present invention, any modifications, equivalent replacements, improvements, etc., should be included in the protection scope of the present invention.

Claims (6)

1. a channel estimation methods that is used for the TDS-OFDM system is characterized in that, at the TDS-OFDM system receiving terminal, described channel estimation methods comprises:

Steps A: previous frame is last or channel estimating that the last iteration of present frame obtains is as the initial channel impulse response of this iteration of present frame, remove in the described present frame pseudo random sequence to the interference of data block according to described initial channel impulse response, and carry out the cyclic reconstruction of described data block;

Step B: described cyclic reconstruction result is carried out equilibrium, and the result of described equilibrium is carried out the part judgement;

Step C: result and described initial channel impulse response according to described part judgement are removed the interference of described data block to described pseudo random sequence, and carry out the cyclic reconstruction of described pseudo random sequence;

Step D: the cyclic reconstruction result of the pseudo random sequence of disturbing according to described removal data block reappraises channel impulse response, and with the channel estimating of described channel impulse response as described this iteration of present frame.

2. the channel estimation methods that is used for the TDS-OFDM system according to claim 1, it is characterized in that, remove in the described present frame pseudo random sequence to the interference of data block according to described initial channel impulse response in the described steps A, and the step of carrying out the cyclic reconstruction of described data block comprises specifically:

Calculate the circular convolution of pseudo random sequence and described initial channel impulse response, the result of described circular convolution is the interference of described pseudo random sequence to data block;

Make up the circular convolution of the channel impulse response of described data block and described this iteration according to the next frame of described present frame and described present frame;

In the circular convolution result of described data block and described initial channel impulse response, remove of the interference of described pseudo random sequence, obtain the cyclic reconstruction result of described data block described data block.

3. the channel estimation methods that is used for the TDS-OFDM system according to claim 1 is characterized in that, among the described step B described cyclic reconstruction result is carried out balanced step and specifically comprises:

Described cyclic reconstruction result is carried out the FFT conversion;

Described initial channel impulse response is carried out the FFT conversion, carry out conjugate operation by element then;

The result of described conjugate operation and the described result that described cyclic reconstruction result is carried out the FFT conversion are carried out dot product, obtain balanced result.

4. the channel estimation methods that is used for the TDS-OFDM system according to claim 1 is characterized in that, the step of among the described step B result of described equilibrium being carried out the part judgement specifically comprises:

Described initial channel impulse response is carried out the FFT conversion, and ask for described transformation results absolute value square;

The result of described equilibrium is carried out the part judgement of in-phase component and quadrature component respectively according to the optimal judgement interval parameter in described square of result, constellation symbol and planisphere space.

5. the channel estimation methods that is used for the TDS-OFDM system according to claim 1 is characterized in that, described step C specifically comprises:

Calculate the circular convolution of described data block and described initial channel impulse response, the result of described circular convolution is the interference of described data block to described pseudo random sequence;

Make up the circular convolution of the channel impulse response of described pseudo random sequence and this iteration according to the next frame of described present frame and described present frame;

In the circular convolution result of the channel impulse response of described pseudo random sequence and described this iteration, remove of the interference of described data block, obtain the cyclic reconstruction result of described pseudo random sequence described pseudo random sequence.

6. the channel estimation methods that is used for the TDS-OFDM system according to claim 1 is characterized in that, described step D specifically comprises:

The cyclic reconstruction result of the pseudo random sequence that described removal data block is disturbed carries out the FFT conversion;

Described pseudo random sequence is carried out the FFT conversion;

Described cyclic reconstruction result's the FFT transformation results and the FFT transformation results of described pseudo random sequence are divided by, described phase division result is carried out the IFFT conversion obtain new channel impulse response, and with the channel estimating of described new channel impulse response as described this iteration of present frame.

Images