CN108900461A - A kind of wireless communication system broadband signal design method based on extensive MIMO - Google Patents

Abstract

The wireless communication system broadband signal design method based on extensive MIMO that the invention discloses a kind of, in one embodiment, the transmission design for 100M bandwidth signal include the following steps：The data of 32 subband bandwidth 3MHz are generated on each transmitting antenna in extensive mimo system, and each subband data is that OFDM modulation (OFDM) obtains, the subcarrier spacing of OFDM is 15kHz；The transmitting signal that fast implements 100MHz bandwidth that algorithm synthesized of the time domain data of 32 subband bandwidth 3MHz Jing Guo generalized multi-carrier (GMC) composite filter group；The reception signal of each receiving antenna carries out extensive MIMO signal detection after corresponding GMC analysis filter group.Present invention combination OFDM technology and GMC filter group fast implement algorithm, generate the 100MHz bandwidth signal based on extensive mimo wireless communication private network, have many advantages, such as that algorithm complexity is low, is easy to hardware realization.

Description

A kind of wireless communication system broadband signal design method based on extensive MIMO

Technical field

The present invention relates to the designs of the down link signal of wireless communication system, and in particular to one kind based on extensive MIMO without The design method of line special communication network system middle width strip signal.

Background technique

As 5G technology is ripe day by day, the wireless communication private network application demand based on extensive MIMO is also extremely urgent. However the base station range of private network requires to be much larger than 5G cell mobile communication systems, therefore, is based on direct orthogonalization frequency division multiplexing (OFDM) subcarrier spacing of technology selects 15kHz usually to eliminate the influence of the multidiameter delay of private network application environment, generates The private network signal of 100MHz bandwidth needs to carry out 8192 points of FFT transform, and implementation complexity is higher.So, it is necessary to for based on The wireless communication private network of extensive MIMO studies new, low complex degree broadband signal design scheme.

Summary of the invention

Goal of the invention：The purpose of the present invention is for extensive mimo wireless communication system provide it is a kind of fast implement it is low multiple Miscellaneous degree broadband signal design method.

Technical solution：The present invention adopts the following technical scheme that：

According to the required signal bandwidth W of data transmission_bWith preset subband bandwidth W_sb, obtain number of sub-bands M；

M subband bandwidth W is generated on each transmitting antenna in extensive mimo system_sbData, and every height Band data are all to modulate to obtain by OFDM；

M subband bandwidth W_sbOFDM data fast implement algorithm by generalized multi-carrier (GMC) composite filter group The W synthesized_bThe transmitting signal of bandwidth, composite filter group realize that step mainly includes：Initialization, subband modulation and preceding frequency It moves, IDFT and rear frequency displacement, cyclic extensions, filtering, the cumulative and displacement of signal are exported.

The reception signal of each receiving antenna carries out extensive MIMO signal after corresponding GMC analysis filter group Detection, the realization of analysis filter group mainly include：After initialization, sliding, filtering, summation, DFT and preceding frequency displacement, subband demodulation Frequency displacement and etc..

As a preferred implementation manner, according to signal bandwidth W_bWith preset subband bandwidth W_sbObtain number of sub-bands M's Calculation method is as follows：By signal bandwidth W_bDivided by preset subband bandwidth W_sb, it takes and is less than and closest to the even number of the division result, As number of sub-bands M.

As another preferred embodiment, according to signal bandwidth W_bWith preset subband bandwidth W_sbObtain number of sub-bands M Calculation method it is as follows：By signal bandwidth W_bDivided by preset subband bandwidth W_sb, take and be less than and closest to the surprise of the division result Number, as number of sub-bands M.

The ptototype filter of GMC composite filter group and GMC analysis filter group is low as a preferred implementation manner, Bandpass filter.

As further preferred embodiment, the low-pass filter is root-raised cosine filter.

As another further preferred embodiment, the low-pass filter is by sinusoidal impulse signal adding window And formed, filter coefficient is constructed by following form：

Wherein, w (k) is time window function, using Hanning window, T_wFor window duration, k indicates the discrete form of time；

p_iIt (k) is the sinusoidal impulse signal of filter, U is the quantity of filter tap, is equal to filter order+1, and S is FFT size, and W is the quantity of the data subcarrier of distribution,It is pitch shift amount.

Beneficial effect：The technical solution of the application is suitable for big bandwidth signal (for example, the signal for being greater than 20MHz bandwidth) Transmission design, and the calculation amount for generating OFDM transmitting signal can be significantly reduced.By taking the OFDM of 100M bandwidth emits signal as an example, Existing technology is directly to emit signal, computation complexity 2 by the OFDM that 8192 FFT transform obtain 100MHz bandwidth¹³ × 13, and the application is first to obtain 32 OFDM subband signals obtained by 256 FFT transform, subband signal is again into broad sense excessively The processing of multi-carrier filter group, forms 100MHz bandwidth signal, and the computation complexity of the implementation in the application is 2¹³× 8, the calculation amount of about half is reduced compared with the prior art.

Detailed description of the invention

Fig. 1 is the Transmission system frame of extensive MIMO private network system broadband signal design scheme according to an embodiment of the present invention Figure；

Fig. 2 is that the filter group of GMC modulation /demodulation according to an embodiment of the present invention realizes structure chart；

Fig. 3 is that private network signal GMC filter group according to an embodiment of the present invention fast implements flow chart；

Fig. 4 is that transmitting terminal composite filter group according to an embodiment of the present invention obtains transmitting signal spectrum figure；

Fig. 5 is that schematic diagram is compared in the adjacent channel leakage of transmitting signal according to an embodiment of the present invention；

Fig. 6 is 32 subband signal spectrograms that analysis filter group in receiving end according to an embodiment of the present invention obtains；

Fig. 7 is one of subband signal spectral that analysis filter group in receiving end according to an embodiment of the present invention obtains Figure；

Fig. 8 is planisphere after PDSCH data balancing in receiving end according to an embodiment of the present invention；

Fig. 9 is that PDSCH according to an embodiment of the present invention receives data and emits the Error Vector Magnitude schematic diagram of data.

Specific embodiment

Technical solution of the present invention is described further with reference to the accompanying drawing.

In one embodiment, by taking the transmission design of the wireless communication private network signal of 100M bandwidth as an example, subband takes 3MHz Bandwidth.According to given total signal bandwidth and subband bandwidth, division result is about 33.33, is taken less than the numerical value and closest The odd number or even number of the numerical value, can be used as corresponding number of sub-bands, consider number of sub-bands based on symmetrical in the present embodiment It is 32.Since the interval between subband data is slightly greater than 3MHz, 32 3MHz are approximately equal to 100MHz.Emitting End is modulated by OFDM and generates 32 subband OFDM datas, obtains 100M broadband signal using GMC composite filter.It is connecing Receiving end carries out operation demodulation by GMC analysis filter group, obtains time domain data.

By the subband data of the transmitting terminal 3MHz bandwidth of the OFDM extensive mimo wireless communication private network system modulated It generates according in 3GPP standard the step of physical layer：First to the binary sequence of original data to be sent, that is, load Data carry out DLSCH processing, wherein DLSCH processing include transmission block addition CRC, code character segmentation and CRC addition, channel coding, Code rate matching and code character connect and compose PDSCH input code word.Obtained PDSCH code word first passes through scrambling code, modulation, then exchanges Data processed make layer mapping, and followed by the precoding of wave beam forming, the data after coding are mapped to the resource grid of multiple antennas.Each Frequency domain data in antenna resources grid carries out OFDM and modulates to obtain the time domain data of each subband.

Fast implement algorithm of 32 subband OFDM datas Jing Guo GMC composite filter group obtains 100MHz broadband signal. The discrete-time version of synthesis multi-carrier signal and the discrete-time version of overloading wave analysis of transmission be：

Wherein：L is the discrete representation form of time, and M is subband number, and N is extracting multiple, and a is constant, and j indicates plural number Imaginary part, T be send each sampled point of signal time interval, r (l) be receiving end time domain discrete receive signal, h_s,mFor hair Penetrate the base band shaping pulsed filter of m-th of the subband in end, h_a,mFor the matched filter of m-th of receiving end subband.

P (l) is the pulse shock response of low-pass filter.

In order to verify the correctness of this Design of Signal scheme, need observation signal transmitting terminal and receiving end spectrogram, Adjacent channel leakage than, receive the planisphere and the characteristics such as Error Vector Magnitude of data.It is only at reversed PDSCH receiving end Reason.Fast implement algorithm of the time-domain signal Jing Guo GMC analysis filter group after noisy communication channel obtains resource grid, when For data in frequency resource grid after demapping, solution precoding, solution layer mapping, available planisphere, EVM etc. characterize signal Characteristic, to complete the broadband signal analysis of extensive mimo wireless communication private network.

Step is fast implemented to broadband signal with specific example with reference to the accompanying drawing to be described in detail.

Fig. 1 shows the structure of the wireless communication private network system based on extensive MIMO, refers to wherein defining following technology Mark, these technical indicators are to correspond to proposition to meet the certain particular demands of private network：

(1) frequency point is transmitted：3.4~3.6GHz；

(2) transmission bandwidth：More than or equal to 100MHz；

(3) extensive mimo antenna number is supported：More than or equal to 128；

(4) user's parallel substreams number is supported：More than or equal to 8；

(5) scene is supported：Suburb, island, mountain area；

(6) system peak transmission rate：More than or equal to 1Gbps；

(7) availability of frequency spectrum：Greater than 8bps/Hz；

(8) demodulation threshold：- 1dB (1/3 code rate bit error rate of QPSK is less than 1E-6)

11dB (2/3 code rate bit error rate of 16QAM is less than 1E-6)

17dB (3/4 code rate bit error rate of 64QAM is less than 1E-6)

The downlink transmission end of each antenna in system uses the precoding technique of wave beam forming, generates 32 bandwidth The subband time domain data of 3MHz, each subband has 3584 sampled points in 1ms, and each subband uses OFDM modulation technique, The subcarrier spacing of middle OFDM data is 15kHz, because of technology as defined in the wireless communication private network system based on extensive MIMO It is mentioned in index, highest is modulated using 64QAM, so using 64QAM modulation and demodulation data in embodiment.

1, composite filter group fast implements process

Subband number M=32, extracting multiple N=36, constant a=0.0028 π, sample rate f are set first_s=3.84MHz, Sampled point time interval T=M/f_sN, the length L=2KN+1 of filter, wherein K is a constant, and K value is bigger, and L is longer, at The frequency characteristic of type symbol is better, and intersymbol interference is smaller, but implementation complexity is also higher, and the selection of ptototype filter length is depended on In the requirement of system transmission characteristics and system to performance.System performance is comprehensively considered in the present embodiment and system realizes cost, most K=6 is chosen eventually.Filter coefficient is denoted as p (k), k=0,1,2 ..., L-1.N, k and l is the representation of discrete time, But the discrete-time sample length indicated is different, such as k length is filter coefficient number k=0,1,2 ..., L-1, and N length is the number n=1,2,3 of sampled point in 1ms ..., 3584.Include the following steps：

S11, initialization：The initial value of n and l are set：N=0, l=0；L point data sequence, the effect of this sequence are set It is equivalent to shift register：D (k)=0, k=0,1,2 ..., L-1.

S12, subband modulation：M point sequence is modulated, M point sequence is M subband in some temporal sampled point, Sampled point parallel processing to the same sampling instant of M subband.Data x in m-th of subband time domain_m(n) as the defeated of modulation Enter data, the process of modulation is：Each input data multiplied byAfterwards, it obtains M point sequence y_m(n), 0≤m≤M-1.

S13, preceding frequency displacement, IDFT and rear frequency displacement：Using IFFT to M point sequence y_m(n) it is converted, transformation matrix T= V₂W_M ^*V₁, wherein V₁It is that diagonal line upper value isDiagonal matrix, V₂Be diagonal line upper value be e^-jπk(M-1)/M's Diagonal matrix, W_M=[e^-j2πmn/M]_M×MIt is the IDFT matrix of M × M, obtainsIt is in this way that M baseband signal is defeated Enter and be modulated on different carrier waves, forms multicarrier.

S14, cyclic extensions：M point sequenceIt is extended to 2M point sequence

Then it is extended to L point sequence again, obtains

S15, filtering：It calculates

S16, it adds up：It calculates

D (k)=d (k)+s_n(k), k=0,1 ..., L-1

This is a step accumulating operation, and result of product obtained in the previous step is added with d (k), and d (k) is a continuous renewal Sequence.

S17, displacement output：Preceding M data in L point sequence d (k) are sent to transmission signal sequence, i.e.,：

S (l+k)=d (k), k=0,1,2 ... N-1

Updating L point sequence d (k) is：

S18, update：N=n+1, l=l+N are set, and is transferred to S12.

S19, the transmitting signal finally obtained：

Wherein t_n=1,2 ..., len, len indicate s (t_n) length.

2, analysis filter group fast implements process, includes the following steps：

S21, initialization：The initial value n=1 of n is set；

S22, sliding：L point sequence is taken from receiving in signal r (l)

Wherein discrete time t_nn=N (n-1)+k, k=1,2 ..., L；

S23, filtering：Calculate r_n(k)=p (k) r_n(k), k=0,1 ..., L-1；

S24, summation：The remainder delta of L/ (2M) is sought, it is right2M-delta 0 is mended at end to makeLength Just be 2M integral multiple, thenBecome the matrix that line number is 2M, sums to every data line, obtain the square of 2M × 1 Battle arrayThen it acquires

S25, rear frequency displacement, DFT and preceding frequency displacement：Using FFT to M point sequenceIt is converted, transformation matrix T^H= (V₂W_M ^*V₁)^H, obtain reconstruction signal

S26, subband demodulation：With matrixDot product obtainsFinally obtained again divided by power coefficient

S27, update：N=n+1 is set, S22 is transferred to.

Finally, obtaining the time domain data that length is 3584.After the operations such as demapping, solution precoding, so that it may into The extensive MIMO signal detection of row or signal analysis.

3, in filter group lowpass prototype filter design

The filter group that Fig. 2 and Fig. 3 respectively illustrates GMC modulation /demodulation realizes structure and fast implements process.It is filtered in GMC Wave device is combined into in analysis system, and the impulse response of each composite filter and analysis filter is by low-pass filter Impulse response obtained through ovennodulation, this filter group be known as modulate mode filter group, and low-pass filter be known as prototype filtering Device.Each subband passes through filter again through ovennodulation and can regard subband as passes through low pass, band logical and high-pass filter respectively, realizes Division to entire broadband signal.

There are two types of scheme, one of which uses root-raised cosine filter for the realization of lowpass prototype filter group, can be with It is directly generated using the r cos fir function in MATLAB, parameter is chosen for M=32, N=36, K=6, L=433.It is corresponding MATLAB sentence is：

P=rcosfir (0.15, [- 66], 36,1, ' sqrt')

Wherein, 0.15 is the raised cosine factor, ' sqrt' indicates that the filter designed is root-raised cosine filter.

Another filter implementations is the filter design based on soft truncation.Particularly, it is filtered using prototype The soft truncation of device is as follows：Consider that the filter with rectangular frequency response as ptototype filter, that is, has appropriate bandwidth just String pulse signal p_i(k), this filter is ideal, because it does not generate distortion in the pass-band, while providing comprehensive band Outer inhibition.To p_i(k) impulse response application time window w (k), then by the frequency shifts of filter into required frequency The heart.The both ends of window mask (The windowing mask) are smoothly transitted into zero, can open in this way to avoid truncation filter Unexpected transition at beginning and end, so that the frequency in truncation filter be avoided to overflow；And window is in the arteries and veins of truncation filter Reasonable timi requirement is provided in punching response, the intersymbol interference (ISI) in obtained f-OFDM signal is made to be maintained at and can connect By being limited in range.

The generation process of filter based on soft truncation is as follows：

S31：Generate window function：W (k), it is T that window function, which selects the duration, in embodiment_wHanning window, form is：

S32：Generate sinc function：

Wherein U is the quantity of filter tap, is equal to filter order+1, and S is FFT size, and W is data of distribution The quantity of carrier wave, andIt is pitch shift amount (tone offset).

S33：Generate normalized low-pass filter coefficients：

In one embodiment, according to 15 resource blocks altogether, each resource block has 12 subcarriers, constitutes 3MHz bandwidth System parameter, and subcarrier spacing Δ f is 15kHz, and the size of FFT is 256, and it is as follows to obtain f-OFDM filter parameter：Filter Order 432, pitch shift amount2.5 subcarriers, time window are Wherein L=432.

Every characteristic measurements such as Fig. 4~9 of the signal of the private network system based on GMC finally obtained, are assisted according to 3GPP The technical indicators such as Adjacent Channel Leakage power ratio ACLR, the error vector magnitude EVM provided are discussed to compare.It is mentioned in 3GPP agreement To when each subcarrier data of PDSCH uses 64QAM modulation /demodulation, it is desirable that EVM is less than 9%, and according to Fig. 9,32 OFDM The average EVM of subband data is 2.6%；Provide that adjacent channel leakage ratio should be both less than -44.2dB in 3GPP agreement, according to Fig. 5, ACLR are up to -73.7dB, and measured value is all satisfied index.And with realizing 100MHz band with 8192 point FFT in 5G standard Bandwidth signals computation complexity is compared, and 8192 point FFT computation complexities are 2¹³× 13, the calculating of the implementation in the application is complicated Degree is 2¹³× 8, reduce the calculation amount of about half.

Claims (7)

1. a kind of wireless communication system broadband signal design method based on extensive MIMO, which is characterized in that including walking as follows Suddenly：

According to the signal bandwidth W of data transportation requirements_bWith preset subband bandwidth W_sb, obtain number of sub-bands M；

It is W that M subband bandwidth is generated on each transmitting antenna in extensive mimo system_sbData, and each subband Data are all to modulate to obtain by OFDM；

M subband bandwidth W_sbThe W that fast implements algorithm synthesized of the OFDM data Jing Guo GMC composite filter group_bBandwidth Transmitting signal, composite filter group realize step mainly include：Initialization, subband modulation and preceding frequency displacement, IDFT and rear frequency displacement, Cyclic extensions, filtering, the cumulative and displacement output of signal；

The reception signal of each receiving antenna carries out extensive MIMO signal inspection after corresponding GMC analysis filter group It surveys, analysis filter group realizes that step mainly includes：Initialization, sliding, filtering, summation, DFT and preceding frequency displacement, subband demodulation Frequency displacement afterwards.

2. the wireless communication system broadband signal design method according to claim 1 based on extensive MIMO, feature It is, it is described according to signal bandwidth W_bWith preset subband bandwidth W_sbThe calculation method for obtaining number of sub-bands M is：By signal bandwidth W_bDivided by preset subband bandwidth W_sb, take and be less than and closest to the even number of the division result, as number of sub-bands M.

3. the wireless communication system broadband signal design method according to claim 1 based on extensive MIMO, feature It is, it is described according to signal bandwidth W_bWith preset subband bandwidth W_sbThe calculation method for obtaining number of sub-bands M is：By signal bandwidth W_bDivided by preset subband bandwidth W_sb, take and be less than and closest to the odd number of the division result, as number of sub-bands M.

4. the wireless communication system broadband signal design described in any one of -3 based on extensive MIMO according to claim 1 Method, which is characterized in that the impulse response of the GMC composite filter or GMC analysis filter is all by low-pass filter Impulse response is obtained through ovennodulation.

5. the wireless communication system broadband signal design method according to claim 4 based on extensive MIMO, feature It is, the low-pass filter is root-raised cosine filter.

6. the wireless communication system broadband signal design method according to claim 4 based on extensive MIMO, feature It is, the low-pass filter is formed by sinusoidal impulse signal adding window.

7. the wireless communication system broadband signal design method according to claim 6 based on extensive MIMO, feature It is, the low-pass filter coefficients are constructed by following form：

Wherein, w (k) is time window function, using Hanning window,T_wFor Window duration, k indicate the discrete form of time；

p_iIt (k) is the sinusoidal impulse signal of filter, U is the quantity of filter tap, is equal to filter order+1, and S is FFT size, and W is the quantity of the data subcarrier of distribution,It is pitch shift amount.