JP5393412B2 - Receiving apparatus and demodulation method - Google Patents

Description

  The present invention relates to a receiving apparatus and a demodulation method for estimating a wireless transmission path based on a received signal.

  In a digital wireless communication system, frequency selectivity and time variation of a transmission path occur due to multipath fading caused by reflection of a transmission signal on a building or the like, Doppler fluctuation caused by movement of a terminal, and the like. In such an environment, the receiving apparatus needs to estimate data timing, perform frame synchronization, and perform demodulation.

  For example, Patent Document 1 below proposes a method of estimating a timing based on a correlation value between a replica and a received signal by performing transmission path estimation to create a replica of the received signal. This technique can obtain an accurate transmission path estimation using a small number of pilot symbols, but has a drawback that it cannot follow a transmission path with a fast time variation. Therefore, it is necessary to use a prefilter to absorb the timing error.

  The method of absorbing timing errors using a pre-filter may not be able to cope with the time variation of the transmission path. Therefore, a method of updating the prefilter and the tap of the decision feedback equalizer using a sequential update algorithm has also been proposed (see, for example, Patent Document 2 and Patent Document 3 below).

  Further, for example, in Patent Document 4, Patent Document 5, and Non-Patent Document 1 below, a receiving apparatus corresponding to a changing transmission path is proposed. These receiving apparatuses employ a scheme for estimating fluctuating noise and transmission paths using a Taylor series.

JP 2009-118175 A JP 2005-159467 A JP 2001-177451 A JP 2004-264816 A JP 2008-501272 A

D. K. Borah and B. D. Hart, “Robust receiver structure for time-varying frequency-flat Rayleigh fading channels”, IEEE Trans. On Commun. Vol. 47, no. 3, March 1999, pp. 360-364.

  However, according to the techniques described in Patent Document 2 and Patent Document 3, taps are updated using a sequential update algorithm. For this reason, there is a problem in that tap convergence is slow and processing time is required.

  In addition, Patent Document 4, Patent Document 5, and Non-Patent Document 1 show a technique for estimating fluctuating noise and a transmission path. However, when a timing error occurs due to the influence of time fluctuation of the transmission path, etc. No countermeasure is shown. For this reason, there is a problem that the transmission path cannot be estimated with high accuracy when a timing error occurs.

  The present invention has been made in view of the above, and an object of the present invention is to obtain a receiving apparatus and a demodulation method capable of estimating a transmission path with high accuracy and efficiency even when the transmission path fluctuates over time. To do.

In order to solve the above-described problems and achieve the object, the present invention is a receiving apparatus that demodulates a received signal, and is configured to have N (N is a delay time) set to have different delay times with respect to the received signal. A pre-filter that multiplies tap coefficients for each tap, adds the multiplied signals for each tap and outputs the result as an added signal, and the added signal and its processing result. Using L taps (L is an integer equal to or greater than 1) set to have different delay times with respect to the determination result based on the result, the tap coefficient is multiplied for each tap, and the signal after multiplication for each tap is added. a decision feedback filter to its own processing result Te, based on a transmission symbol corresponding to the received signal and the received signal, before symbol calculates the tap coefficients of the prefilter and the tap was determined coefficient and the Receiving Based on a signal and a transmission symbol corresponding to the received signal, a series expansion coefficient that is a coefficient when the tap coefficient of the decision feedback filter for each of L taps is series-expanded in a predetermined order is obtained, Tap updating means for calculating a tap coefficient of the decision feedback filter based on a series expansion coefficient, and means for making a symbol decision based on the addition signal and the processing result of the decision feedback filter. Features.

  According to the present invention, when performing reception processing using both the prefilter and the decision feedback filter, based on the received signal, when updating the prefilter tap and the decision feedback filter tap, For the decision feedback type filter, the series expanded coefficient is estimated, and the decision feedback type filter tap is updated based on the series expanded coefficient. In addition, there is an effect that it can be estimated efficiently.

FIG. 1 is a diagram illustrating a functional configuration example of the transmission apparatus according to the first embodiment. FIG. 2 is a diagram illustrating an example of a configuration of data and pilot symbols in the first embodiment. FIG. 3 is a diagram illustrating an example of function estimation using Taylor series expansion. FIG. 4 is a diagram illustrating an example of estimation of f between ± 5 samples. FIG. 5 is a diagram illustrating a configuration example of the receiving apparatus according to the first embodiment. FIG. 6 is a flowchart illustrating an example of a tap update procedure according to the first embodiment. FIG. 7 is a diagram illustrating an example of a configuration of data and pilot symbols according to the embodiment. FIG. 8 is a diagram illustrating an example of setting timing of the number M of coefficients. FIG. 9 is a diagram illustrating an example of setting timing of the number M of coefficients. FIG. 10 is a diagram illustrating an example of a tap update processing method when pilot symbols are used. FIG. 11 is a diagram illustrating an example of the tap update processing method according to the second embodiment. FIG. 12 is a flowchart illustrating an example of a tap update process procedure when estimation is repeatedly performed using demodulated data symbols. FIG. 13 is a diagram illustrating an example of the tap update processing procedure according to the third embodiment in the case where the estimation is performed repeatedly.

  Embodiments of a receiving apparatus and a demodulation method according to the present invention will be described below in detail with reference to the drawings. Note that the present invention is not limited to the embodiments.

Embodiment 1 FIG.
FIG. 1 is a diagram illustrating a functional configuration example of a first embodiment of a transmission device according to the present invention. As shown in FIG. 1, a transmission apparatus according to the present invention includes a symbol encoder 1 that performs symbol mapping of information bits to be transmitted based on a modulation scheme such as PSK (Phase Shift Keying) or QAM (Quadrature Amplitude Modulation), and a transmission antenna. And 2. The transmission apparatus according to the present invention is the same as the conventional transmission apparatus. Note that the modulation scheme applied to the transmission signal transmitted by the transmission apparatus is not limited to PSK and QAM, and any modulation scheme may be used.

The Q and a power of two integers, symbols encoder 1 of the transmission apparatus maps the information bits of the log ₂ Q bits to a point of Q signal points, and outputs to the transmission antenna 2. The transmission antenna 2 transmits the mapped signal (symbol) as a radio signal.

Hereinafter, the operation of the present embodiment will be described. First, b _k = [b _{k, 1} , b _{k, 2} ,..., B _{k, lQ} ] (lQ = log ₂ Q) And the symbol output from the symbol encoder 1 is s _k . The symbols used in the description of this embodiment are shown below.
T _s : Symbol time (seconds)
M: Series expansion parameter (number of terms of series expansion)
N _OS : Number of oversamples in one symbol time N _T : Number of pilot symbols L _FB : Number of transmission path taps to be estimated L: Number of multipaths in the transmission path N: Range in which series expansion is performed N _MEM : Estimation during demodulation [A] _{i, j} : (i, j) element of matrix A

  Note that Diag (a) represents a diagonal matrix (the component other than the diagonal component is 0) with the diagonal component as the vector a.

FIG. 2 is a diagram showing an example of the configuration of data and pilot symbols in the present embodiment. s ₀ to s _NT-1 are pilot symbols for transmitting a known bit pattern, and s _NT to s _N-1 are data symbols. The configuration example of FIG. 2 is an example, and the number of pilot symbols, the number of data symbols, and the arrangement of pilot symbols are not limited to this, and _NT pilot symbols are arranged before or after a predetermined number of data symbols. Any configuration may be used as long as the configuration is inserted into the. A ^T indicates the transpose of the matrix A.

Here, series expansion will be described. When the function f (t) is expanded in series using M coefficients g _i (i = 1, 2,..., M−1) and the series θ _i (t), the following equation (1) is obtained. .

Examples of functions used for series expansion include Taylor functions and Fourier functions. For example, when using the Taylor function, θ _i (t) = t ⁱ . FIG. 3 is a diagram illustrating an example of function estimation using Taylor series expansion. FIG. 3 shows an example in the case of M = 3, and shows a concept of approximating f (t) as the sum of a zero-order function (constant), a linear function, and a quadratic function with respect to t. Yes.

Here, if the function f (t) is sampled at N points with different t (N samples) and f (bold) = [f _N−1 , f _N−2 ,..., F ₀ ] ^T , The series expansion of f (bold) can be expressed as the following formula (2).

In the above equation (2), the sampled θ _i (t) is expressed as c _{0: N−1, i} = [c _{N−1, i} , c _{N−2, i} ,..., C _{0, i} ]. ^Shown as ^T.

When estimating f (bold) elements of N samples, N elements are estimated. On the other hand, if series expansion is used, θ _i (t) is known, and the coefficients g _{0 to} g _M-1 may be estimated. Therefore, the number of elements to be estimated is changed from N to M (M <M N) The number can be reduced. Therefore, the estimation accuracy can be improved.

FIG. 4 is a diagram illustrating an example of estimating f (bold) between ± 5 samples. In FIG. 4, when f (bold) is estimated at 10 points from n = −5 to n = + 5, if series expansion (development up to the second order) is used, g ₀ , g ₁ , g ₂ of 3 It shows that one coefficient should be estimated.

Next, a receiving apparatus that receives a transmission signal transmitted from the transmitting apparatus described in FIG. 1 will be described. When the transmission signal s (t) transmitted by the transmission device is received by the reception device, the amplitude and phase are disturbed by the frequency-selective transmission path that varies with time. The reception signal y (t) in which the amplitude and phase are disturbed can be expressed as the following expression (3) using a known transmission path model. Here, the number of transmission paths (paths) is L, and h _l (t) represents the transmission path model of the l-th transmission path.

  The received signal y (t) shown in the above equation (3) can be estimated using series expansion as shown in the following equation (4).

When a shift of τ _D seconds occurs in the sample timing in the synchronization processing in the receiving device, the received signal y _k [m] after the sample processing is defined by the following equation (5). In the sample processing, sampling is performed at time intervals obtained by dividing one symbol time into N _OS . M represents the sample number in the symbol.

Next, transmission path estimation in the receiving apparatus will be described. The operation of the receiving apparatus that performs transmission path estimation by decision feedback using the series decomposition described above will be described. FIG. 5 is a diagram illustrating a configuration example of the receiving apparatus according to the present embodiment. FIG. 5 shows a part related to transmission path estimation in the receiving apparatus, which is the same as the configuration of the conventional receiving apparatus. As shown in FIG. 5, the receiving apparatus of the present embodiment includes delay units 11-1 to 11- (L _FF −1), multipliers 12-1 to 12-L _FF , an adder 13, and a delay. Units 14-1 to 14 -L _FB , multipliers 15-1 to 15 -L _FB , an adder 16, an adder 17, and a Decision Maker (determination unit) 18.

The delay units 11-1 to 11- (L _FF -1), the multipliers 12-1 to 12-L _FF, and the adder 13 are pre-filters (FFF: Feedforward Filter) for absorbing timing shifts. 20 is configured. L _FF is the number of taps of the FFF 20. The delay devices 11-1 to 11- (L _FF -1) give different delays to the input received signals. The multiplier 12-j (j = 2 to L _FF ) multiplies the output of the delay device 11- (j−1) by the tap f ^FF _j set for each. Incidentally, the multiplier 12-1 multiplies the tap FFF20 ₀ set in the received signal without giving a delay. The adder 13 adds the multiplication results of multipliers 12-1 to 12-L _FF, and output to the adder 17.

The adder 17 adds the addition result of the adder 13 and the addition result of the adder 16, and outputs the addition result to a Decision Maker (determination unit) 18. Decision Maker18 performs symbol decision based on the input signal, and outputs the symbol decision result to the delay unit 14-1, also when the determination process ends, outputs the determined symbol s _k.

The delay units 14-1 to 14 -L _FB , the multipliers 15-1 to 15 -L _FB, and the adder 16 constitute a decision feedback filter (FBF: Feedback Filter) 21. The delay units 14-1 to 14 -L _FB give different delays to the determination result input from the Decision Maker 18. The multiplier 15-j (j = 1 to L _FB ) multiplies the output of the delay device 11-j by the tap FBF21 _j set for each. The adder 16 adds the multiplication result of the multiplier 12-1 to 12-L _FF, and output to the adder 17.

  Next, a tap update process using series decomposition according to the present embodiment will be described. The output of the adder 17 of the receiving apparatus shown in FIG. 5 (the sum of the output of the FFF 20 and the output of the FBF 21) can be expressed by the following equation (6).

In this embodiment, f ^FF _j and f ^FB _j are estimated using pilot symbols. At this time, FBF 21 _j is estimated using a coefficient (coefficient of series expansion). A pilot symbol is defined as shown in the following equation (7), and a received signal matrix is defined as in the following equation (8). In this example, it is assumed that the pilot symbol is arranged before the data frame.

The updating of the tap values of the FFF 20 and the FBF 21 is performed individually. Here, the matrix A is defined as shown in the following equation (9). The size of the matrix A is (N _T −L _FB ) × (L _FF + L _FB · M). _{Incidentally, C x: x + m =} [c x: x + m, 0, c x: x + m, 1, ..., c x: x + m, M-1], S D x: x + m = Let Diag {s _{x: x + m} }.

Further, the relationship between the desired signal and the equalizer tap is defined as the following equation (10). The vector t is a vector defined by the following formula (11), where g (bold) _j is a coefficient vector obtained by series expansion of the vector f (bold) _FF indicating the tap for the FFF 20 and the tap f ^FB _{j of the} FBF 21. It is.

  The estimated value of the vector t can be obtained according to the following equation (12).

When set respectively an estimate f (hat) _FF vectors FFT tap obtained by obtaining the estimated value of the vector t to the multipliers 12-1 to 12-L _FF of the receiving device, the output of FFF20 The signal vector z (bold) can be expressed by the following equation (13).

  Then, the tap coefficient (coefficient obtained by series expansion) of the FBF 21 is estimated using the output signal of the FFF 20 represented by Expression (13). Here, the matrix B is defined as the following formula (14).

The size of the matrix B is (N _T −L _FB ) × (M · (L _FB +1)). The relationship shown in the following formula (15) is established between the output signal of the FFF 20 and the coefficient of the FBF 21. The definition of g (bold) _{0: LFB} is shown in the following equation (16).

g (bold) _0: The estimated value of _LFB can be obtained according to the following equation (17).

ここで、ＦＦＦ２０の出力を以下の式（１８）のように定義し、シンボル系列ｓ_k:k-LFB＝［ｓ_k，ｓ_k-1，…，ｓ_k-LFB］^Tとすると、復調（ｓ_kの推定）は、以下の式（１９）に従って行うことができる。

Here, the output of the FFF 20 is defined as in the following equation (18), and the symbol sequence s _{k: k−LFB} = [s _k , s _k−1 ,..., S _k−LFB ] ^T is demodulated ( The estimation of s _k can be performed according to the following equation (19).

Here, the m-th element of the tap for the FBF 21 is expressed by the following equation (20). Each element of the tap for the FBF 21 is updated for each sample. C ^T _km = [c _{km, 0} , c _{km, 1} ,..., C _{km, M−1} ] ^T.

  The estimation of the tap and the coefficient of the FFF 20 using the above equation (12) and the above equation (17) can be performed iteratively. In this embodiment, these estimations are repeatedly performed. FIG. 6 is a flowchart illustrating an example of the tap update procedure according to the present embodiment. Here, although the Decision Maker 18 performs the following tap update procedure, the present invention is not limited to this, and a tap update unit may be separately provided to perform the same processing.

  First, the Decision Maker 18 initializes the FFF 20 tap and the coefficient of the FBF 21 (coefficient obtained by series expansion of the tap value) as an initial setting according to the following expression (22), and based on the result, the FFF 20 tap is determined according to the expression (23). An output signal is obtained (step S11). The m-th updated coefficient and tap are expressed as shown in the following formula (21).

  A coefficient is calculated | required according to the following formula | equation (24) using the result calculated | required by Formula (23) (step S12).

  The counter (i) is initialized (i = 0) (step S13). An estimated value of the received signal is obtained according to the following equation (25) (step S13).

受信信号の推定値を用いて、以下の式（２６）を計算し、計算結果に基づいてＦＦＦ２０のタップと係数の更新を行う（ステップＳ１４）。

The following equation (26) is calculated using the estimated value of the received signal, and the tap and coefficient of the FFF 20 are updated based on the calculation result (step S14).

  The output signal of the FFF 20 is obtained according to the following equation (27) (step S16).

  The coefficient of the FBF 21 is updated according to the following equation (28), and updated to the tap value corresponding to the updated coefficient (step S17).

The counter is incremented (i = i + 1: step S18). It is determined whether or not the counter i is larger than the set number of iterations N _MAX (step S19). If i is greater than N _MAX (Yes in step S19), the process is terminated. If i is equal to or less than N _MAX (No in step S19), the process returns to step S14.

The above tap update process is performed in the pilot symbol period, and in the subsequent data symbol period, demodulation is performed using the tap updated by the tap update process. Note that the transmission path estimation accuracy can be improved by increasing the number of iterations (N _MAX ) in the tap update process.

  In the above description, for simplification, the frame configuration shown in FIG. 2 is assumed. However, the present invention is not limited to this, and a frame configuration in which a pilot symbol period is inserted between data symbol periods may be used. . FIG. 7 is a diagram illustrating a configuration example of a frame in which a pilot symbol period is inserted between data symbol periods. Thus, in the case of a frame configuration in which a pilot symbol period is inserted between data symbol periods, the above tap update processing is performed in the pilot symbol period, and demodulation is performed with the set coefficient in the data symbol period. What should I do?

  Further, the number M of series expansion coefficients can be set adaptively or non-adaptively. When adaptively set, for example, a received signal estimation error and a channel estimation error may be obtained in a pilot section (pilot block), and M may be set based on those errors. Alternatively, a received signal estimation error, a channel estimation error, or the like may be obtained in a data symbol period (data symbol block), and M may be set based on these errors. For example, in order to know the maximum Doppler frequency in advance, an appropriate value of M may be set non-adaptively based on the maximum Doppler frequency. Alternatively, the value of M may be set in advance, and then the above-described adaptive setting may be performed as necessary.

  8 and 9 are diagrams illustrating an example of setting timing of the number M of coefficients. FIG. 8 shows an example in which M is set in the pilot symbol block, and FIG. 9 shows an example in which M is set in both the pilot symbol block and the data symbol block. When M is changed in the data symbol block, the tap update processing performed in the pilot symbol block may be performed again using the received signal in the pilot symbol block to update the tap.

  As described above, in the present embodiment, when reception processing is performed using both FFF 20 and FBF 21, coefficients obtained by series-expanding taps of FFF 20 and taps of FBF 21 based on reception signals corresponding to pilot symbols And the tap of the FBF 21 is obtained based on the series expanded coefficient. Therefore, even when the transmission path fluctuates over time, the transmission path can be estimated with high accuracy and efficiency. Further, it is possible to estimate the transmission path more accurately by repeatedly performing the tap of the FFF 20 and the coefficient obtained by series expansion of the tap of the FBF 21.

Embodiment 2. FIG.
Next, a tap updating method according to the second embodiment of the present invention will be described. The configuration of the communication system of the present embodiment is the same as that of the first embodiment. The configurations of the transmission apparatus and the reception apparatus of this embodiment are the same as those of the first embodiment.

  In the first embodiment, tap update processing is performed using pilot symbols. FIG. 10 is a diagram illustrating an example of a tap update processing method when pilot symbols are used. On the other hand, not only pilot symbols but also tap updates can be performed in the same manner using data symbol demodulation results (demodulated data symbols). In this case, since the tap update timing is not limited to the pilot symbol period, the tap can be updated frequently, and the tracking capability of the transmission path that fluctuates at high speed can be further improved as compared with the first embodiment. Therefore, in this embodiment, tap update processing is performed using both pilot symbols and demodulated data.

  FIG. 11 is a diagram illustrating an example of the tap update processing method according to the present embodiment. When demodulated data is used, in Equations (9) and (14) shown in Embodiment 1, matrix A and matrix B are obtained using demodulated data symbols (demodulated data symbols) instead of pilot symbols. It will be.

  Hereinafter, an example of a process (tap update process) for estimating the series expansion coefficient of the tap of the FFF 20 and the tap of the FBF 21 using the demodulated data symbol will be described. The basic process is the same as that using the pilot symbols described in Embodiment 1 except that demodulated data symbols are used instead of pilot symbols. The description will focus on the differences from the case where pilot symbols are used.

Estimation starts at time k. The length (number of symbols) N _MEM of the estimation memory is set as shown in the following equation (29). At this time, the received signal matrix Y _k is defined as shown in the following equation (30).

The demodulation result up to time k is s (hat) _k, and matrix A is defined as shown in the following equation (31) instead of equation (9) shown in the first embodiment.

The size of the matrix A is N _MEM × (L _FF + L _FB · M). The tap of the FFF 20 is updated according to the following equation (32).

The taps f (hats) _{k, FF} of the FFF 20 obtained according to the above equation (32) are set in the FFF 20, and the output signal of the FFF 20 is defined as shown in the following equation (33).

Then, the FBF 21 coefficient (series expansion coefficient) is estimated using the output signal of the FFF 20 as follows. Here, the matrix B _k is defined as shown in the following equation (34).

The size of the matrix B _k is N _MEM × (M · (L _FB +1)). And the relationship shown in the following formula | equation (35) is formed between the output signal of FFF20, and the coefficient of FBF21.

The estimated value g (hat) _{0, LFB} of the coefficient g _{0, LFB} can be obtained according to the following equation (36).

  According to the above procedure, the tap of the FFF 20 and the coefficient of the FBF 21 can be updated using the demodulated data symbol, but the estimation can be performed repeatedly as in the case of using the pilot symbol.

  FIG. 12 is a flowchart illustrating an example of a tap update process procedure when estimation is repeatedly performed using demodulated data symbols. As shown in FIG. 12, when performing estimation repeatedly, first, the tap of the FFF 20 and the coefficient of the FBF 21 are initialized as shown in the following equation (37), and the following is performed based on the initialized value. The output of the FFF 20 is obtained according to the equation (38) (step S21).

  The coefficient of the FBF 21 is obtained according to the following equation (39) (step S22).

  The counter i is initialized (i = 0) (step S23). Based on the following equation (40), an estimated value of the received signal is obtained (step S24).

  Using the estimated value of the received signal, an estimated value of the tap of the FFF 20 and the coefficient of the FBF 21 is obtained according to the following equation (41), and the tap of the FFF 20 is set corresponding to the estimated value (step S25).

  Based on the estimated value in step S25, the output of the FFF 20 is obtained according to the following equation (42) (step S26).

  According to the following equation (43), the estimated value of the coefficient of the FBF 21 is obtained, and the tap of the FBF 21 is set corresponding to the estimated value (step S27).

The counter i is incremented (i = i + 1: step S28). It is determined whether the counter i is greater than the set number of iterations N _MAX (step S29). If i is greater than N _MAX (Yes in step S29), the process ends. If i is equal to or less than N _MAX (No in step S29), the process returns to step S24. The tap update process using data symbols is the same as in the first embodiment.

  As described above, in this embodiment, the tap of FFF 20 and the tap of FBF 21 are updated using the demodulated data symbol obtained by demodulating the data symbol, as in the first embodiment. Therefore, compared with the first embodiment, the tap update frequency can be improved, the capacity to adapt to the time change of the transmission path is increased, and the transmission path estimation accuracy is further improved. It is also possible to perform tap updating and demodulation using the pilot symbols and demodulated data symbols as in the above embodiment.

Embodiment 3 FIG.
Next, a tap updating method according to the third embodiment of the present invention will be described. The configuration of the communication system of the present embodiment is the same as that of the first embodiment. The configuration of the transmission apparatus of this embodiment is the same as that of Embodiment 1, but the reception apparatus performs transmission path estimation using a maximum likelihood sequence estimator (MLSE).

In MLSE, the coefficient of FFF 20 and the coefficient of FBF 21 (coefficient of series expansion of taps) can be estimated for each series assumption. When performing MLSE estimation for N data symbols (N symbols) from s ₀ to s _N−1 , there is an assumption of Q ^N. The reception signal matrix Y _N-1 for N symbols is defined as shown in the following formula (44).

When the estimation of the N symbols of the sequence assumption i is s (hat) ₀ {i},..., S (hat) _N−1 {i}, the matrix A is used instead of the equation (9) shown in the first embodiment. _N-1 {i} is defined as shown in the following formula (45). Here, for simplification, s (hat) ₋₁ to s (hat) _{−LFB + 1} are assumed to be known as pilot symbols or the like.

The size of the matrix A _N−1 {i} is N × (L _FF + L _FB · M). The update of the tap of the FFF 20 can be performed as follows. First, the estimated values of the taps of the FFF 20 and the coefficients of the FBF 21 are defined as shown in the following formula (46).

T (hat) _N-1 {i} is obtained according to the following equation (47).

The tap f (hat) _{N−1, FF} {i} of the FFF 20 obtained according to the equation (47) is set in the FFF 20, and the output signal of the FFF 20 is defined as shown in the following equation (48).

The coefficient of the FBF 21 is estimated using the output signal of the FFF 20 described above. First, the matrix B _k is defined as shown in the following formula (49).

The size of the matrix B _k is N × (M · (L _FB +1)). The relationship shown in the following formula (50) is established between the output signal of the FFF 20 and the coefficient of the FBF 21.

Coefficient g _{N-1,0: LFB} estimated value g (hat) _{N-1,0: LFB} is defined as shown in the following equation (51).

g (hat) _{N-1,0: LFB} can be obtained based on the following equation (52).

仮定ｉにおけるシンボル系列をｓ（ハット）_k:k-LFBを以下の式（５３）に示すように定義する。最尤仮定は以下の式（５４）に基づいて行なう。

The symbol sequence in assumption i is _defined as s (hat) _{k: k-LFB} as shown in the following equation (53). The maximum likelihood assumption is made based on the following equation (54).

  The m-th element of the tap of the FBF 21 can be expressed as the following formula (55).

  As described above, the estimated values of the taps of the FFF 20 and the FBF 21 can be obtained. Note that the estimation may be performed repeatedly as in the first and second embodiments.

  FIG. 13 is a diagram illustrating an example of a tap update processing procedure according to the present embodiment when iterative estimation is performed. First, j indicating an assumed number is initialized (j = 0) (step S31). As shown in the following formula (56), the tap of the FFF 20 and the coefficient of the FBF 21 are initialized (step S32).

  Based on the initialization value, the output signal of the FFF 20 is obtained according to the following equation (57) (step S33).

  The coefficient of the FBF 21 is obtained according to the following equation (58) (step S34).

  The counter i is initialized (i = 0: step S35). An estimated value of the received signal is obtained according to the following equation (59) (step S36).

  Based on the estimated value of the received signal, the estimated value of the tap of the FFF 20 and the coefficient of the FBF 21 is obtained based on the following equation (60). The tap of the FFF 20 is updated to the estimated value (step S37).

  The output signal of the FFF 20 is obtained according to the following equation (61) (step S38).

  An estimated value of the coefficient of the FBF 21 is obtained according to the following equation (62) (step S39).

The counter i is incremented (i = i + 1: step S40). It is determined whether the counter i is greater than the set number of iterations N _MAX (step S41). If i is N _MAX or less (No in step S41), the process returns to step S36.

When i is larger than N _MAX (Yes in step S41), the tap of the FBF 21 is updated according to the following equation (63), and the estimation error in the assumption j is calculated according to the following equation (64) (step S42).

  j is incremented (j = j + 1: step S43).

It is determined whether or not j ≧ Q ^N (step S44). When j ≧ Q ^N (step S44 Yes), according to the following equation (65), an assumption j (hat) with the smallest estimation error is selected, and FFF20 and FBF21 are set to taps corresponding to the j, A sequence corresponding to j is set as a demodulation result (step S45). If j ≧ Q ^N is not satisfied (No in step S44), the process returns to step S32.

  Thus, by performing iterative estimation for all assumed sequences, data demodulation and transmission path estimation can be performed simultaneously. In addition, since the transmission path is estimated for each sequence, the estimation accuracy is improved.

As an example, a transmission signal is BPSK, the transmission signal s _k is can be expressed by the following equation (66). When N = 5 and 5 symbols are estimated, there are 2 ⁵ = 32 possible sequences. Further, when N = 5 when the transmission signal is 64QAM, there are 64 ⁵ = 1073741824 possible sequences. Thus, when the value of N or Q increases, MLSE requires a large amount of computation. In order to reduce the calculation amount, an algorithm with a reduced calculation amount may be used. For example, “H. Kubo, K. Murakami and T. Fujino,“ Adaptive maximum-likelihood sequence estimation by means of combining equalization and decoding in fading channels ”, IEEE J. Select. Areas of Commun., Vol. , Pp. 102-109, Jan. 1995 ”discloses a demodulation method that has an error rate characteristic close to that of MLSE with a low calculation amount. The tap update processing of this embodiment may be performed by introducing a coefficient obtained by series expansion of the taps of the FBF 21 using an algorithm as shown in this document.

  Further, the update processing of the FFF 20 tap and the coefficient of series expansion of the FBF 21 may be performed using a known technique such as Least Mean Square and the like.

  As described above, in the present embodiment, when performing demodulation using MLSE, the taps of the FBF 21 are estimated using the coefficients obtained by series expansion. Therefore, the same effects as those of the first embodiment can be obtained, and the estimation accuracy can be further improved.

  As described above, the receiving apparatus and the demodulation method according to the present invention are useful for a receiving apparatus that estimates a wireless transmission path based on a received signal, and particularly for a receiving apparatus that uses a prefilter and a decision feedback filter. Is suitable.

1 symbol encoder 2 transmit antennas _{11-1~11- (L FF -1), 14-1~14} -L FB delayer _{12-1~12-L FF, 15-1~15-L} FB multiplier 13, 16, 17 Adder 18 Decision Maker
20 FFF
21 FBF

Claims (11)

A receiving device for demodulating a received signal,
Using N taps (N is an integer of 1 or more) set to have a different delay time for each received signal, multiplying the tap coefficient for each tap, and adding the signal after multiplication for each tap A pre-filter that outputs as a sum signal,
Using L taps (L is an integer of 1 or more) set so that the delay times are different from each other based on the determination result based on the addition signal and its own processing result, the tap coefficient is multiplied for each tap, A decision feedback filter that adds the signal after multiplication for each tap to obtain its own processing result; and
Based on the transmission symbol corresponding to the received signal and the received signal, before Symbol calculates the tap coefficients of the prefilter, also in a transmission symbol corresponding to the tap determined coefficient between the received signal and the received signal Based on the series expansion coefficient, a series expansion coefficient that is a coefficient when the tap coefficient of the determination feedback filter for each of L taps is series-expanded in a predetermined order is obtained, and based on the series expansion coefficient, the determination feedback filter Tap updating means for calculating a tap coefficient;
Means for performing symbol determination based on the addition signal and the processing result of the determination feedback filter;
A receiving apparatus comprising: The received signal is a signal storing a known symbol, and the transmission symbol is a known symbol.
The receiving apparatus according to claim 1. The received signal is a signal storing data symbols, and the transmission symbol is a result of the symbol determination,
The receiving apparatus according to claim 1 or 2, wherein Demodulate by maximum likelihood sequence estimation,
The tap updating means calculates tap coefficients of the pre-filter and the decision feedback filter for each assumed sequence, and estimates a maximum likelihood sequence based on the calculation result for each sequence.
The receiving apparatus according to claim 1. Repetitively calculating the tap coefficients of the pre-filter and the decision feedback filter,
The receiving apparatus according to claim 1, wherein the receiving apparatus includes: The predetermined order is set based on an estimation error of a received signal or a transmission path estimation error obtained based on a known signal storing a known symbol.
The receiving device according to claim 1, wherein Presetting the predetermined order based on a maximum Doppler frequency of the received signal;
The receiving apparatus according to claim 1, wherein The series expansion is a Taylor expansion.
The receiving apparatus according to claim 1, wherein The received signal is a BPSK modulated signal,
The receiving apparatus according to claim 1, wherein the receiving apparatus includes: The received signal is a QAM modulated signal,
The receiving apparatus according to claim 1, wherein the receiving apparatus includes: A demodulation method in a receiving apparatus for demodulating a received signal,
Using N taps (N is an integer of 1 or more) set to have a different delay time with respect to the received signal, the tap coefficient is multiplied for each tap, and the multiplied signal for each tap is added. A pre-step for outputting as an addition signal,
Using L taps (L is an integer of 1 or more) set so that the delay times are different from each other based on the determination result based on the addition signal and its own processing result, the tap coefficient is multiplied for each tap, A decision feedback step in which the signal after multiplication for each tap is added to obtain its own processing result;
Based on the transmission symbol corresponding to the received signal and the received signal, before Symbol calculates the tap coefficients of the prefilter, also in a transmission symbol corresponding to the tap determined coefficient between the received signal and the received signal Based on the series expansion coefficient, a series expansion coefficient that is a coefficient when the tap coefficient of the determination feedback filter for each of L taps is series-expanded in a predetermined order is obtained, and based on the series expansion coefficient, the determination feedback filter A tap update step for calculating a tap coefficient;
A symbol determination step for performing symbol determination based on the addition signal and the processing result of the determination feedback filter;
The demodulation method characterized by including.

Images