Power Reallocation and Complexity Enhancement for Beyond 4G Systems

Omar Daoud＊

Communications and Electronics Engineering Department,Philadelphia University,Amman 19392,Jordan.

Abstract:Competitiveness in the modern wireless systems' provided services is a key factor in the development,in addition to the adaptation to/harmonization of user demand.Therefore,this paper discussed the quality of services from the point of view of the need for supporting the needed data rates.For this purpose,a benchmark based on Multi-Antenna Multiband Orthogonal Frequency Division Multiplexing has been proposed to compatible with Ultra-Wideband systems such as the fifth generation based technologies.In order to enhance the system quality of service,the structure of the ultra-wideband system's main stage; namely Orthogonal Frequency Division Multiplexing has been modified by imposing a low complexity designed Haar-wavelets stage instead of the fast Fourier transform stage.This is in addition to reallocate the transmitted power in order to reduce the effect of one of the main drawbacks that is found in the Orthogonal Frequency Division Multiplexing; namely the peak-to-average power ratio problem.A MATLAB simulation has been performed in order to validate the propositions that have been made based on six different performance factors.As a result,the new propositions were achieved our targets by reducing the system's complexity in terms of mathematical operations and by giving promising results in managing the transmitted powers.Furthermore,the effectiveness of such work has been verified and compared with four different work in the literature

Keywords: multi antennas-multiband wireless; PAPR; Haar-DWT; power reallocation;ACLR; EVM.

I.INTRODUCTION

Due to its bandwidth efficiency; orthogonal frequency division multiplexing (OFDM) has become a very interesting topic for the fifth generation (5G) and other wireless communication technologies; this is due to its orthogonality criterion.Thus,the quality of service(QoS) is easily enhanced by supporting huge data rates.Furthermore the downlink rates will be increased to exceed hundred megabits per seconds..A new version of the OFDM systems was found in the literature to support the high QoS resolution,namely multiband-OFDM(MBOFDM).It consists of a parallel conventional OFDM links associated with a frequency hopping benchmark to enable the multiple access technology,which is widely used in the ultra-wideband (UWB) communication systems [1-5].

UWB technologies have a 20% signal fractional bandwidth greater than its carrier frequency.Under such systems,the available spectrum will be divided into sub-bands each with 500 MHz bandwidth in the systems of band between 3.1G Hz and 10.6GHz [6].One of the main deficiencies that is associated with such requirement; high bit rate transmission(the time dispersion effect),is the inter symbol interference (ISI).This isdue to the frequencies channel selectivity [7,8].In order to attain the needed demand of such system's users and to combat the channel selectivity; the traditional single-band UWB systems have been shifted toward the multi-band approach [9-11].Many candidates have been found in the literature to be used for the UWB technologies,such as the MBOFDM and the multicarrier based systems .Thus,a hybrid work could be considered as the principal for a huge number of different wireless communications systems such as the Long term Evolution-Advanced(LTE-A),IEEE 802 wireless networks standard 11 and 16,and others Broadband Radio Access Network (BRAN) systems [2,6,9-15].

MBOFDM technique is used to enhance the usage of the limited bandwidth and to combat the channel effects such as the ISI easily by applying the Fast Fourier Transform (FFT)concepts [2,7,8].Furthermore,the multicarrier techniques such as the multiple-input multiple-output (MIMO) are proposed to allow the effective bandwidth widening based on the multi antennas criterion.Thus,it will maximize the radio spectrum utilization flexibility.Based on the LTE-A systems specifications,from five carriers of 100 MHz bandwidth up to thirty-two carriers could be used [16,17].

This work proposes two techniques in order to enhance the QoS for multi antennas-MBOFDM (MA-MBOFDM) systems.This enhancement is attained based on reducing the Peak to Average Power Ratio (PAPR)problem and reallocating the signal's power.PAPR limits the use of none linear devices such as the mixers,up-conversion stages,and the power amplifiers (PAs).This will leads to distort the transmitted data and to the intermodulation effect,the spectral spreading problems and may change the signal constellation [7,8,18-20].Different solutions have been found in the literature in order to combat the PAPR effect and to keep the average signal power levels in a tolerable value,which prevents the limitations in the wireless systems circuitry [18-35].Thus,the contribution in this work is to provide an insight on the MA-MBOFDM based systems performance,which will be attained by checking both of the complementary cumulative distribution function (CCDF),the symbol error rates (SER)curves,and many other needed practical parameters will described later in section II.This is in addition to enhance both of the systems complexity and to reallocate the power constellations in the transmitted signals.Thus,a proposition of replacing the FFT stages with a low complexity Haar discrete wavelet transforms (DWT) will be described later.This is due to that the DWT is considered as a high presentation digital signal processing technique,which has some advantages compared to the FFT stage such as overcoming the need for cyclic prefix,the huge flexibility and the optimal resolution [36-38].In addition to a proposed work based on reallocate the signal's power by making use of the pulse width modulation (PWM) theory.By imposing the PWM block,the probability of exceeding a certain PAPR value will be 0 (based on the PAPRdBequation in [7,and 8]),this is due to that the new generated amplitudes have no variations and accordingly the average power will equal the maximum value.Thus,the system performance should be enhanced; maximum power amplifier efficiency issues attained in addition to the ability of using the nonlinear devices.

Through the signals simulation experiments based on different evaluation parameters,the proposed Haar structure reduces the system's complexity by a factor of 33.3% in terms of mathematical operations.

II.SYSTEM MODEL

Figure 1 depicts the idea of the MAMBOFDM.It can be defined as a combination between an OFDM stage with multicarrier processes to be transmitted through a MIMO channel.

The OFDM idea has been proposed in [18],which is based on using modulator banks.This allows the transmission of a huge data rate in a parallel way instead of the conventional sequential process.After that,this idea has been modified by Saltzberg in [19]by making use of the FT process.During the last three decades the idea of the conventional OFDM has been modified several times,one of them is to reduce the system complexity toN×log2(N)by proposing the FFT usage [7,8].

In such system,the UWB channels bandwidth is divided into sub-bands each of which has an almost 500MHz range with a period of 312.5ns.Therefore in each MBOFDM stage,there areNconventional OFDM stages to be sent through thelthtransmitting antenna.This is attained by making use of a Time-Frequency Code selection (TFC) criteria Block.TFC uses a combination between either puncturing pattern with time repetition or with a frequency repetition,which makes it a suitable choice to be used for frequency hopping in the data transmission within the band.This will have a direct effect on the instantaneous transmitted power,i.e.it will be enhanced even at the same spectral power density.Space Time-Frequency codes are used in order to provide the combination of those emerging technologies(UWB,OFDM,MIMO) with significantQoS improvements such as the system capacity,the needed data rate,the bit error efficiencies.The scope of this work is to make use of the published work in [39-42],which examined the space TFC's thoroughly.

As shown inFigure 1,the OFDM stage is the key-block in the whole system.It usually contains three different processes,namely coding stage,modulation stage,and the FFT process stage.The transmitted data OFDM symbol throughlthtransmitting antenna could be represented as in [7,8]:

Fig.1 MA-MBOFDM communication system block diagram.

wherecstands for the carrier number,Aclis the modulated symbol of thecthcarrier and thelthantenna for an independent and identically distributed data,φnis imposed to overcome the channel effect on the time domain signal's envelope distribution,Nstands for the size of the FFT stage,and the oversampling factor denoted byJ.

The OFDM signal will be filtered,up-converted,and passed through the PA stage before entering the MIMO channel.The MIMO channel is defined as in [31]as a Rayleigh channel with a wide sense stationary uncorrelated scattering model.The model for anLtransmitting antennas andMreceiving antennas is clearly shown below in (2) as:

At this stage,and due to the PA limitations,the transmitted signal will be affected and distorted.At the receiver side,the received signal at themthreceiving antenna will be as:

where the added noise to themthreceiving part at thecthcarrier is denoted bywcm.

In order to enhance the MA-MBOFDM system's efficiency,Figure 1 has been improved based on imposing two propositions;replacing the FFT stage by modified Haar wavelets based,and imposing a power transformation stage for the OFDM signal based on the PWM.As mentioned earlier in section 1,the PAPR is considered as a serious problem in the OFDM signal.This is due to that it has a large envelope fluctuation after the IFFT stage in the transmitter,which limits the average power levels.The PAPR,after the IFFT stage,can be defined as:

where,the maximum power that is found in the OFDM symbol is defined asPpeak,andPavgstands for the average power.

Therefore,as shown in Figure 2,a low complexity Haar wavelets replacement stage is proposed to limit the PAPR drawbacks appearance.Daubechies family gives the lowest PAPR among other wavelet families using M-ary signaling.This is due to that the Haar wavelets function has linear time and space complexities.Besides,its window size adjusts itself optimally in two cases; to longer basis function at low frequencies and to shorter basis function at high frequencies.This is in addition to many other advantages as found in[43].

In Figure 2,it is clearly shown that the conventional OFDM block has been modified by imposing the modified Haar wavelets after the serial to parallel (S/P) stage.As well as imposing the power transformation stage before entering the PA stage.

Thus,the first proposition is about replacing the core block of the OFDM systems; IFFT block,by a wavelet transformation process.Therefore,the OFDM systems will be known as orthogonal wavelet division multiplexing(OWDM).It consists of filter banks to fulfill both of the decompositions and the reconstructions stages at the transmission and reception,respectively.In such systems,the transmitted time signals will be mapped into another demission based on the scale factors and the wavelets coefficients.This will improve the systems capabilities and flexibilities.Thus,the wavelets will be defined based on both of the scale factors which is known by approximates(An) and based on the wavelet coefficients which are known by the details (Dn).The whole process of choosing both of theAn's and theDn's depends on the cut off frequency.Therefore,the signals will be divided into a high frequencies block and denoted byDnand a low frequencies block that is denoted byAn.Thus,two types of filters will be used; high pass filters (HPF) and low pass filters (LPF),namelyh[n]andg[n]respectively.Like any communications systems and due to the real devices compatibility,the continuous version will be transformed to the discrete one.Therefore,our scope will cover the discrete wavelet transform (DWT) baby functions only,such as Haar wavelets,wavelets Daubechies 4,Reverse Biorthogonal (RBior 3.3) wavelets,the Biorthogonal (Bior 5.5),Discrete Meyer,Coiflet 3,and Symlets 10[34,35,43,44].As shown in Figure 2,the modulated symbol of thecthcarrier and thelthantenna;Aclwill pass through the IDWT stage to be analyzed at different frequencies based on the cut off frequency.

Fig.2.Modified MBOFDM architecture.

Fig 3.An and Dn decomposition process based on cutoff frequency.

Then,then-th decomposition level ofDn's andAn's will be formed for each symbol as shown in the following flowchart.On the oth-er hand,the shown process in Figure 3 will be inverted to form the reconstruction process.This will be ful filled by making use of the previous level of the approximation part; namelyAn-1; to complete the inversion process [44].After that,the whole work depicted in Figure 1 will be denoted by multi-antenna multiband OWDM (MA-MBOWDM) instead of the MAMBOFDM [39].As found in the literature,the use of DWT will enhance the system's performance either from the complexity point of view or from the data rates point of view.This is due to that the content of the resultant signal will contain both of the time and the frequency information [44-48].

In this work,the author intends to enhance the OWDM based work by proposing a low complexity Haar wavelets construction.In this proposition,the Haar wavelets will be constructed based on some new de fined parameters that were derived from the Haar wavelets matrix parameters (an orthogonal matrix).The analytical model for such proposition has revealed about a relationship between thep-th DWT points (levels),and a constant ofas described below as:

- Stage 1: Defining a squaredN×NHaar matrix (H()N N×):

As defined in [44],the Haar matrix could be fined based on a set of squared signals and the size of such matrix.This is true for the interval of [0,1].Thus the transformation process will be accomplished by multiplyingH()N N×with the transmitted signal.ThusH()N N×withi-th row andq-th column is defined as:

TableI.Haar matrix parameters for different sizes.

From (4),theHmatrix has either a value ofor a value ofHere,Mis defined based on both of the Haar matrix size;Nand the previous row's value;ith-1.Therefore,theHmatrix could be defined based on 5 different parameters;A,B,C,andD.Table1 shows those parameters for different Haar matrix sizes.

- Stage 2: Generating the transformed data:

Based on the previously defined parameters,the transformed signals;T,will be generated by multiplying the modulated symbol of thecthcarrier and thelthantenna;Acl,after dividing it intoNbands with the transformation matrix as:

In the case of the 4-bands (i.e.N=4),the transmitted signal will be generated as expressed in (6) as:

- Stage 3: Generalizing the whole process:

In order to generalize the transformation process forN-Haar wavelets level a transformation parameter; “a” is defined as:

Here,wis any positive integer,p=1,2,…..,N=2w,andis defined earlier as the input modulated symbol of thecthcarrier and thelthantenna to thep-thIDWT point.Thus,this proposition that replaces theh[n]andg[n](bases of the Haar wavelets) by the transformation parameter;ap,will practically facilitate the MBOWDM implementation,such as on the field programmable gate arrays(FPGA's).

- In case forw= 2,thenpwill have values{1,2,3,4}.

- Based on (8),the transformation parameter will be

- Defining a second level transformation parameters as

- The transformed data will be

Accordingly,implementation complexity is reduced at a very promising level based on the mathematical operations.

In order to compare the proposed work that is based on the transformation parameter with the conventional Haar work that is based on theA,B,CandDparameters in case for a transformation matrix of (4×4),the work in Table1 shows that we need 20 mathematical operations (12 multiplications and 8 additions).This mathematical operations have been reduced to 10 (4 multiplications and 6 additions) based on (10).Since the computational complexity for one sample in the conventional Haar work could be calculated by:

Therefore,a total of 50% complexity reduction has been attained by using the proposed work,which is described in Table2.

The reception structure based on the proposed Haar transforming operation is simply attained by inverting the mathematical operation at the transmission stage.In case of four level Haar wavelet transforms,Figure 4 depicts the reception process and shows how this proposition is simple to get the transformation coefficients at the ends.At this case we need 6 addition/subtraction operations and 4 multiplication/division operations as overall mathematical operations,which is the same as the needed mathematical operations at the transmission stage.

In case to check the proposed Haar transforms performance,a comparison has been made with the FFT from the BER point of view.Figure 5,depicts not only the Haar performance compared to the conventional FFT but also depicts the powerfulness of the Haar against DB4 wavelets,RBior3.3,Discrete Meyer,Coiflet 3,and Symlets 10.

Figure 5 concludes that both of the Haar work and the conventional FFT work were diverged quickly compared to the others work;DB,RBior,Discrete Meyer,Coiflet,and Symlets wavelets baby functions.In fact,this result is similar to the conclusions that were found in the literature [44-47]; that the Haar work has the best performance among the other wavelettransform baby functions.In addition,this proposition reduces the system complexity as well.This is in addition to that proposed work enhances the systems performance compared to the FFT work.Thus,the proposed work could enhance the OWDM in comparison with the OFDM from the BER point of view.Figure 2 shows that the MBOFDM has been modified by imposing two different blocks namely Haar wavelets based block and the power transformation based on PWM block respectively.The first target has been accomplished by proposing a low complexity Haar wavelets transform as concluded from both of Table2 and Figure 5.The other challenge is to reduce the nonlinearity limitations effect on the system data rate.Thus,the PAPR problem has been taken into consideration.PAPR,as found in [8],arises due to the coherently addition of numerous signals and is defined as the ratio between the resultant signal to its average.This ratio could affect the nonlinear devices in either the transmission chain or the reception one due to the need of high dynamic ranges.Thus,in this work a proposition has been made based on redistributing the power in the signal based on the slope between two successive samples as depicted in Figure 6.In order to check the performance of such proposition,some factors should be taken into considerations,such as adjacent channel leakage ratio (ACLR),the frame error rate (FER),energy efficiency factor,average mutual information (AMI),and the complementary cumulative distribution function (CCDF).This is in addition to the error vector magnitude (EVM) factor,which is needed for the practical manufacturing issues as found in [41].ACLRfactor importance has been drawn from relating the out-of-band power to the in-band one as shown in equation(12):

TableII.Simple Haar transformation process-based w-level transformation parameters.

Fig.4.Receiver Structure based on 4th-level mathematical operations.

Fig.5.Proposed Haar work performance based BER curves comparison.

For any data transmitted froml-th antenna,c-th carrier andp-th DWT point,Jstands for oversampling factor (determined based on the original data average power;Acl pand its relation to the distorted data average power,

TheAMIfactor gives an indication about the effect of imposing new signal processing blocks on the achieved data rates.Therefore,the consumed energy from the additional processing blocks should be taken into consideration.Thus,de fining the AMI factor could be described as follows:

where,Powerconventionalstands for the conventional consumed power,andαdenotes the percentage needed for combatting the PAPR problem.

Furthermore,from (13) an the AMI could be upper bounded as

While theCCDFfactor could be defined to cover the probability of exceeding the normalized dB power certain threshold as:

The last factor in analyzing the proposed work performance is theEVM,which can be de fined from [50]as (15) shown in the bottom at this page,wheregtstands for the gain imbalance andφtis the quadrature error.

As shown in Figure 6,the proposed work consists of two stages,namely a preprocessing stage and a transforming stage,respectively.

The first stage consists of two operations in order to simplify the processing procedure without any reductions on the system's performance.The first operation deals with pre fixing the resulted signal after Haar wavelets work;with a distinguishing sample.In our case the zero magnitude has been chosen.Furthermore and in order to enhance the conversion ef ficiency,a new sampling rate has been used.This sampling rate can be defined based on the effective bandwidth and on the set of all signals that have energies.Thus,the sampled signal,x(n) at the sample point can be de fined as:

whereMis an integer,fois the signal's frequency andMfoshould satisfy the effective bandwidth of the Nyquist-Landau rate.Therefore,those samples at a rate ofMfocan be easily split into a train ofMsamples at rate offomaking use of the z-transformation process.Furthermore,the optimum sampling position should be carefully chosen where the average power of the received signals should be the maximum.This can be evaluated as [51]:

Fig.6.Proposed Power Transformation block based PWM.

Fig.7.A flowchart of the conversion process.

wherekis number of OFDM subcarrier,zis the Nyquist time or frequency,ζis the sampling offset which is related to the perfect sampling position,ckis the modulated data,Qis the number of adjacent samples.

Based on the mean power algorithm,average the power over small number of samples leads to the optimum position for the sampling process.

The output from the preprocessing stage is ready to be processed in the transforming stage,which is depicted in Figure 8.

This proposition could be inserted either after the guard interval stage to ensure having a free of ISI channel or before it to offer a perfect circular convolution.Therefore,the insertion choice will be optimized to fit the cost and complexity equation criteria.

Fig.8.Transforming stage flowchart.

In the transforming stage,three main processing levels will be fulfilled; calculating the slope between any two points,decision process based on a new generated formula as shown in Figure 8,and the replacement process.Figure 8 shows the flowchart of the transforming process,whileTpw(i) is thei-th sample in the resulted Haar wavelets signal after the preprocessing stage,TRwp(i) stands for thei-th transformed sample,Nmis the new sampling rate after the preprocessing stage.Thus,the power reallocation stages could be summarized as:

· Attach a zero sample to lead the OFDM symbol as a distinguishing one.Therefore,the start point of each symbol will be known.

· Enhance the conversion accuracy by increasing the number of samples between the adjacent OFDM samples;Tpw(n,m) with anew sampling rate;(m).

· ConvertTpw(n,m) from vector into a matrix with three different vectors as

wheren,mand( m ) were defined earlier,while I has values from 1 to

· Each sample insideTpw(i) will be compared withTpw(i-1) and modified to have constant amplitude using the following formula:

In this work,the signal to noise ratio (SNR)expression has been modified in order to attain the power efficiency enhancement.This modification will take extra energy consumption into consideration and has the conventional system's energy.Thus,a modification element has been defined asαto check the effect of the system's energy on the actual datrate and to relate both of signal processing power to the actual system's energy.Furthermore,Figure 9 depicts the power transformation stages that were defined in both Figure 6 and Figure 8.

Figure 9 depicts the transformation process from the original form into a constant envelope signal.The top part of Figure 9 shows the oversampling process in addition to the comparison with the received signal through channel.The black curve shows the original samples of the original signal,while the blue curve shows the oversampled version of the signal which is almost identical to the received signal that is depicted in yellow circles.The middle part shows the transformed version by making use of the depicted procedure in Figure 9.After that,in order to check the proposition performance,a sample error rate (SER)has been calculated in the stage of reconstructing the original signal from the received transformed signal.The SER shows that they are almost identical.Furthermore and in order to check our proposition's performance,theCCDFcurves have been calculated and compared to our previously published work and to some conventional work in the literature as shown in Figure 10.These curves depict the probability of exceeding a normalized dB power any certain threshold.

Figure 10 illustrates the proposed work performance.A comparison has been made among the power transformed based work,previously published work in [49]and some work in the literature such as clipping technique and the partial transmit sequence based work.It is clearly shown that our proposition gives a promising result in combatting the PAPR values compared to the literature.This is due to the simplicity in transforming the OFDM signals into a constant envelope ones and the recovering process that shown in Figure 9.After checking the proposed work performance; based on both ofSERandCCDFcurves,the modified MIMO-OFDM (imposing a hybrid combination between the low complexity Haar work and the power transformed work) system's performance will be checked based on theACLR,FER,CCDFandEVMfactors in the next section.

Fig.9.Power transformation-based process.

Fig.10.Proposed work performance based CCDF curves.

Fig.11.System’s performance based ACLR factor comparison.

III.SIMULATION RESULTS

In this section,simulation results are presented in order to evaluate the proposed work performance as discussed before; low complexity Haar work and the power reallocation work.

The simulations have also been compared with two work categories; our previously published work and the work in the literature to validate the achieved results.Specifically,the performance has been checked based on theACLR,theFER,the energy efficiency factor,theAMI,theCCDFand theEVMfactors.The used simulation parameters are as follows:

- orthogonal subcarriers of 512,

- oversampling factor of 8,

-J; Power reallocation oversampling factor of 9,

- TFC of 3,and 5 carriers,

- 3rdorder,Identical-behavioral-characteristics nonlinear HPAs (class A) with 3 dB and 5 dB input back off (IBO) values,

- DBPSK and 16-QAM modulation techniques,

- Wavelets baby functions (DB4,RBior 3.3,Symlet 10,Coiflet 3,Discrete Meyer and WPT,2×2 MIMO specifications (4 delay taps),

- CM3 multiband channel specifications14.18 ns means excess delay,

o 14.28 ns rms delay spread,

o 0 dB channel energy mean,and

o 3.1 channel energy standard deviation.

Figure 11 is divided into two parts,the top one relates the power amplifier's efficiency to the range of (-20 to -50)ACLRdB values,while the bottom one depicts the relationship between the PA gain with differentACLRdB values.From the depicted results in the top part,the efficiency of the proposed hybrid work is better than the achieved results of using the PAPR reduction work based on the SLM.The blue curve which represents the use of both of the proposed Haar work in addition to power reallocation work shows extra 5% efficiency enhancements over both of the conventional OFDM work and the SLM-based work.Furthermore,the PA gain of the different PAPR reduction techniques has been compared under the sameACLRrange.Once again,the hybrid work proposition achieves the highest gain among them.This means that the proposed work has an excellent capability for combatting the PAPR values.

TheCCDFcomparison work has been de-picted in Figure 12 based on two modulation techniques; 16-QAM and DBPSK.This figure compares theCCDFof the hybrid work with the conventional system in addition to some work in the literature.At 10dB threshold and comparing to the depicted results in Figure 9,the use of the hybrid combination between the proposed Haar work and the power reallocation block gives 89.6% extra performance enhancement.It reduces the probability of exceeding the 10dB from 1.91×10-6to 1.986×10-7.

The same result has been attained even the modulation technique has been changed to DBPSK.From the DBPSK modulation technique point of view,if the system has been designed to not exceed the 10% probability,it is shown that the proposed hybrid work reduces the threshold dB value from 11dB to 9dB.This means that an extra 18.18% performance enhancement has been attained over the proposed power reallocation technique alone,while this performance has been increased to be 25% over our previously published work in[49].

In order to check the system's performance based on both of the AMI parameter and the energy efficiency factor,Figure 13 depicts both values in term of the modified SNR that includes the processing needed energy; the relationship between the data rates have been included.Furthermore,a limit of 45 dB is used in order to be adequate to the new generation of the wireless systems.From the depicted results in Figure 13,the hybrid work gives the best results compared either to the proposed work individually or to the conventional OFDM.Focusing on a 30 dB threshold,the hybrid proposition gives extra 15% enhancement over the conventional work.

Finally,the systems performance has been checked from both of theFERandEVMpoint of view as shown in Figure 14.The top part shows the EVM values for two antennas model,while the bottom part depicts the SER results comparisons.

Fig.12.Proposed Systems performance based CCDF curves comparisons.

Fig.13.System performance based on the AMI and the Energy efficiency factors.

In Figure 14,the EVM performance has been evaluated for a 2×2 MA-MBOFDM system under the assumption of two scenarios; the first for a TFC of 3 and two IBO values (3dB and 5 dB) while the other one is for a TFC of 5.From the depicted results in the top part,it is obvious that the EVM values are high at low SNR due to the effect of the nonlinear noises and to the channel effect.Furthermore,there is no need to increase the TFC complexity by increasing the number of carriers due to a slight enhancement for the EVM values.Thus,these values could be optimized for the system's requirements.Furthermore,the bottom part depicts the FER and once again the hybrid proposition shows promising results at the reception part; at 25 dB threshold it gives an extra 15% enhancement over the conventional work.

Fig.14.Simulated results for 2X2 MA-MBOFDM with two different TFC values under IBO equals 3 and 5.

IV.CONCLUSION

In order to enhance the QoS for a MAMBOFDM,two main propositions have been made to modify the system's structure.Because of the PAPR effect on the system's performance,a power reallocation method based on PWM technique has been proposed.This is in addition to propose a new low complexity Haar structure to enhance the transceiver's structure.Through the signals simulation experiments based on different evaluation parameters,the proposed Haar structure reduces the system's complexity by a factor of 33.3%in terms of mathematical operations.This proposition has been compared to different wavelet baby functions such as DB4,RBior 3.3,Discrete Meyer,Coiflet,and Symlets for validation.In the scope of reducing the effect of the processing power and the one of the transmitted signals,six main factors have been evaluated and simulated,such as theACLR,theFER,theCCDFvalues,theEVM,the energy efficiency parameter,and theAMIvalues.Thus,the verifications have been made for the two propositions and the target has been fulfilled.Thus,a conclusion could be drawn as the MA-MBOFDM efficiency can be enhanced by replacing the FFT block with the low complexity Haar work.Furthermore,a promising enhancement from the power management point of view was achieved for the proposed power reallocation stage.