GB2547444A

GB2547444A - Universal filtered multi-carrier systems and methods

Info

Publication number: GB2547444A
Application number: GB1602834.2A
Authority: GB
Inventors: Zhang Lei; Xiao Pei; Ijaz Ayesha; Tafazolli Rahim
Original assignee: University of Surrey
Current assignee: University of Surrey
Priority date: 2016-02-18
Filing date: 2016-02-18
Publication date: 2017-08-23
Anticipated expiration: 2036-02-18
Also published as: GB2547444B; GB201602834D0

Abstract

Embodiments of the invention provide methods and apparatus for transmitting and receiving Universal Filtered Multi-Carrier (UFMC) signals. A multicarrier signal is generated by transforming a plurality of data symbols to be transmitted on a plurality of subcarrier frequencies within a sub-band into a time domain multicarrier signal, and filtering the time domain multicarrier signal using a sub-band filter. The method further comprises a step of adding samples to the time domain multicarrier signal before transmission by applying a cyclic prefix or zero padding, or further comprises a step of removing samples from the filtered time domain multicarrier signal before transmission by performing tail cutting. A method of processing a received multicarrier signal is also disclosed, comprising the steps of: obtaining a plurality of time domain samples of the multicarrier signal; adding a plurality of the samples from a first signal portion at the beginning of the multicarrier signal to a plurality of the samples from a second signal portion at the end of the multicarrier signal; converting the multicarrier signal from the time domain to the frequency domain; and equalizing the frequency-domain multicarrier signal to obtain estimated data symbols. In some embodiments, the equalizer takes into account transmitter imperfections such as a carrier frequency offset and timing offset.

Description

Universal Filtered Multi-Carrier Systems and Methods Technical Field

The present invention relates to Universal Filtered Multi-Carrier (UFMC) systems and methods.

Background UFMC is a promising air-interface waveform candidate solution for the next generation of wireless communications systems. Figure 1 is a block diagram schematically illustrating a prior art UFMC transmitter. In a UFMC transmitter, data symbols in an subband, an, are converted into the time domain by an Inverse Discrete Fourier Transform (IDFT) spreader 102. The time-domain signal for the nth subband is then filtered in a dedicated subband filter Fn 104. The filtered subband signals are then summed and transmitted via an antenna 106.

In comparison with alternatives such as filtered Orthogonal Frequency Division Multiplexing (OFDM) systems which filter the whole bandwidth, and Filter Bank Multi-Carrier (FBMC) systems which perform filtering on a per-subcarrier basis, UFMC provides greater flexibility by filtering a subband that consists of an arbitrary number of consecutive subcarriers. Filtering an arbitrary number of subcarriers allows a UFMC system to suppress out of band (OoB) emission while maintaining the orthogonality between subbands and subcarriers within one subband.

Figure 2 is a graph of filter gain versus subcarrier index for a subband filter in the UFMC transmitter of Fig. 1, for filter lengths L = 20,40 and 60 with the Fast Fourier Transform (FFT) size = 240. As shown in Fig. 2, the OoB emission can be reduced by increasing the filter length, which also improves the robustness to transceiver imperfections. However, drawbacks of using a longer filter include the larger filter gain selectivity along the subcarriers in one subband, as shown in Fig. 2, as well as larger overhead and reduced spectrum efficiency. UFMC is no longer an orthogonal system in multipath environments since there is no guard interval between symbols. It has been claimed that a subband filter with a length comparable to that of the channel will incur negligible performance loss, however, this claim has not been proved analytically and may not be true for some scenarios, for example in harsh channel conditions. To solve this problem, it has been proposed to add a cyclic prefix (CP) after the subband filtering to avoid inter-symbol interference (ISI). However, this system cannot achieve interference-free one-tap equalization, since the circular convolution property is destroyed. A further drawback of adding a CP after the subband filter is increased receiver complexity, since zero padding, down-sampling and high-order DFT operation at the receiver is required.

The invention is made in this context.

Summary of the Invention

According to a first aspect of the present invention, there is provided a method of generating a multicarrier signal, the method comprising: transforming a plurality of data symbols to be transmitted on a plurality of subcarrier frequencies within a subband into a time domain multicarrier signal; and filtering the time domain multicarrier signal using a sub-band filter, wherein the method further comprises: adding samples to the time domain multicarrier signal before transmission by applying a cyclic prefix or zero padding or removing samples from the filtered time domain multicarrier signal before transmission by performing tail cutting.

When performing tail cutting samples are preferably discarded from tail regions on both sides of the sub-band, however, in some embodiments samples may only be discarded from a tail region on one side of the sub-band.

In other embodiments according to the first aspect, the cyclic prefix is added before filtering the time-domain multicarrier signal using the sub-band filter.

In embodiments according to the first aspect, the method further comprises multiplying each one of the plurality of data symbols by a power compensation factor, before transforming a plurality of data symbols into the time domain multicarrier signal. The power compensation factor to be applied to a data symbol can be determined based on the known impulse response of the sub-band filter, and may be determined in advance for different subcarrier indices and stored. Such a method may further comprise normalising the total transmission power of the filtered time domain multicarrier signal to match a transmission power of one or more other sub-bands included in the multicarrier signal.

According to a second aspect of the present invention, there is provided a method of processing a received multicarrier signal, the method comprising: obtaining a plurality of time domain samples of the multicarrier signal; adding a plurality of the samples from a first signal portion at the beginning of the multicarrier signal to a plurality of the samples from a second signal portion at the end of the multicarrier signal; converting the multicarrier signal from the time domain to the frequency domain; and equalizing the frequency-domain multicarrier signal to obtain estimated data symbols.

In some embodiments according to the second aspect, the method further comprises a step of setting a length of the first signal portion and a length of the second signal portion according to whether tail cutting or zero padding was applied at a transmitter. For example, the length of the first and second signal portions can be set as L-i+Lzp, where L is the length of a filter used at the transmitter and Lzp is the length of tail cutting or zero padding applied by the transmitter.

According to a third aspect of the present invention, there is provided apparatus for generating a multicarrier signal, the apparatus comprising: a time domain conversion unit configured to transform a plurality of data symbols to be transmitted on a plurality of subcarrier frequencies within a sub-band into a time domain multicarrier signal; a sub-band filter configured to filter the time domain multicarrier signal; and a sample adding unit configured to add samples to the time domain multicarrier signal before transmission by applying a cyclic prefix or zero padding; or a tail cutting unit configured to remove samples from the filtered time domain multicarrier signal before transmission.

In some embodiments according to the third aspect, the apparatus further comprises a precoding unit configured to multiply each one of the plurality of data symbols by a power compensation factor, before the plurality of data symbols are transformed into the time domain multicarrier signal by the time domain conversion unit. The power compensation factor to be applied to a data symbol can be determined based on a known impulse response of the sub-band filter, and may be calculated in advance for a plurality of subcarrier indices and stored. During operation of the apparatus, the precoding unit may then retrieve the stored power compensation factor associated with a particular subcarrier index for applying power compensation to the data symbol on that subcarrier.

According to a fourth aspect of the present invention, there is provided apparatus for processing a received multicarrier signal, the apparatus comprising: a signal receiver configured to obtain a plurality of time domain samples of the multicarrier signal; a processing unit configured to add a plurality of the samples from a first signal portion at the beginning of the multicarrier signal to a plurality of the samples from a second signal portion at the end of the multicarrier signal; a domain converting unit configured to convert the multicarrier signal from the time domain to the frequency domain; and an equalizing unit configured to equalize the frequency-domain multicarrier signal to obtain estimated data symbols.

According to a fifth aspect of the present invention, there is provided a method of equalizing a frequency-domain multicarrier signal, comprising: obtaining an estimate of one or more parameters relating to transmitter imperfections; and performing one-tap equalization according to the obtained one or more parameters. In some embodiments, a method according to the fifth aspect can be combined with a method of processing a received multicarrier signal according to the second aspect. The one or more parameters relating to transmitter imperfections may, for example, include one or more of: a carrier frequency offset; a timing offset; and a channel autocorrelation matrix.

According to a sixth aspect of the present invention, there is provided apparatus for equalizing a frequency-domain multicarrier signal, the apparatus comprising: an equalizer configured to obtain an estimate of one or more parameters relating to transmitter imperfections, and perform one-tap equalization according to the obtained one or more parameters. In some embodiments, apparatus according to the sixth aspect maybe combined with apparatus according to the third aspect.

Brief Description of the Drawings

Embodiments of the present invention will now be described, byway of example only, with reference to the accompanying drawings, in which:

Figure l schematically illustrates a UFMC transmitter;

Figure 2 is a graph of filter gain versus subcarrier index for a subband filter in the UFMC transmitter of Fig. l;

Figure 3 schematically illustrates apparatus for generating a multicarrier signal, according to a first embodiment of the present invention;

Figure 4 is a flowchart showing a method of generating a multicarrier signal according to the first embodiment of the present invention;

Figure 5 schematically illustrates apparatus for processing a received multicarrier signal, according to the first embodiment of the present invention;

Figure 6 is a flowchart showing a method of processing a received multicarrier signal, according to the first embodiment of the present invention;

Figure 7 schematically illustrates apparatus for generating a multicarrier signal, according to a second embodiment of the present invention;

Figure 8 is a flowchart showing a method of generating a multicarrier signal, according to the second embodiment of the present invention;

Figure 9 schematically illustrates apparatus for processing a received multicarrier signal, according to the second embodiment of the present invention;

Figure 10 is a flowchart showing a method of processing a received multicarrier signal, according to the second embodiment of the present invention;

Figure 11 schematically illustrates a post-data processing method according to the second embodiment of the present invention;

Figure 12 is a flowchart showing an equalization method according to a third embodiment of the present invention;

Figure 13 is a graph comparing the performance of a UFMC system according to an embodiment of the present invention to that of a UFMC system and an OFDM system; Figure 14 is a graph comparing the out-of-band (OOB) emission of a UFMC system according to an embodiment of the present invention to that of a UFMC system and a FBMC system; and

Figure 15 is a graph comparing the peak-to-average power ratio (PAPR) of a UFMC system according to an embodiment of the present invention to that of a UFMC system and an OFDM system.

Detailed Description

Aspects of the invention provide improvements to UFMC systems and methods. Embodiments of the invention which implement the various aspects are described in detail below. Any features of a particular embodiment may be combined with features of any of the other embodiments disclosed herein.

First embodiment: CP-UFMC with power compensation

Apparatus and methods according to a first embodiment of the present invention will now be described with reference to Figs. 3 to 6. Figures 3 and 4 schematically illustrate a UFMC transmitter and a method performed by the UFMC transmitter, according to the first embodiment of the invention. The UFMC transmitter comprises Mbranches, each of which includes apparatus for generating a filtered time-domain multicarrier signal for one of the M subbands within the UFMC signal. The apparatus on each branch comprises a precoding unit 302 for performing power compensation, a time domain conversion unit 304, a cyclic prefix unit 306, and a subband filter 308. The UFMC transmitter further comprises an antenna 310 for transmitting the UFMC signal.

The UFMC transmitter of the present embodiment differs from a known UFMC transmitter in the use of a precoding unit 302 to apply power compensation, and the addition of a cyclic prefix (CP) before the subband filter 308. The precoding unit 302 is configured to multiply each symbol by a non-zero power compensation factor so that all of the symbols within a subband achieve the same gain before channel equalization, substantially eliminating the filter frequency selectivity along the subcarriers in one subband. As shown in Fig. 3, in the present embodiment the power compensation is applied in the frequency domain and the CP is added in the time domain. Adding the CP before the subband filter 308 maintains orthogonality and thereby enables interference-free one-tap channel equalization. Also, adding the CP before the subband filter simplifies the receiver implementation, and provides backwards-compatibility with legacy OFDM receivers. These aspects will be described in more detail below with reference to Fig. 4.

In the present embodiment the time domain conversion unit 304 is configured to apply an IDFT to the power-compensated data symbols received from the precoding unit 302. However, in other embodiments a different type of frequency-to-time domain transform maybe used. The time domain conversion unit 304 is configured to transform a plurality of data symbols to be transmitted on a plurality of subcarrier frequencies within a sub-band into a time domain multicarrier signal. The cyclic prefix unit 306 then adds a CP in the time domain, and the time domain signal with CP is then filtered by the subband filter 308.

The method performed by the CP-UFMC transmitter is illustrated in Fig. 4. First, in step S401 the precoding unit multiplies each one of the plurality of data symbols by a power compensation factor before the plurality of data symbols are transformed into the time domain multicarrier signal by the time domain conversion unit. The power compensation addresses the problem of filter gain selectivity. Referring back to Fig. 2, the filter gain can vary significantly between subcarriers in one UFMC subband, particularly when longer filters are used. The filter gain at subcarriers located near the edge of the subband may be very low compared to the gain at subcarriers near the middle of the subband, resulting in significant performance loss in terms of Bit Error Rate (BER) when all of the subcarriers carry symbols of equal importance. The likelihood of errors occurring in subcarriers at the edge of a subband is much higher due to the lower gain, whilst symbols transmitted on subcarriers in the middle of the subband are allocated unnecessarily high power, leading to inefficient power allocation.

To compensate for the filter gain selectivity, in the present embodiment the precoding unit multiplies each data symbol by a power compensation factor. The use of power compensation factors will now be described in the context of an exemplary multicarrier system comprising N subcarriers, with the index set U= [o, 1, · · ·, N- l]. The N subcarriers are divided into M subbands, with the m* subband comprising Nm consecutive subcarriers from set U. It follows that the subcarrier index set for the mth subband is:

(1)

The Nm long vector-modulated symbols transmitted on the m* subband are given by:

(2)

In the present embodiment, the coefficients are calculated for the precoding matrix as follows. The finite impulse response (FIR) filter for the m1*1 subband is defined as Jm=|/m(o),/m(i),.../m(b^m-i)], where Lp,m is the length of the filter for the m* subband. Taking the IV-point DFT (discrete Fourier transform) of fm to get the filter response gm in frequency domain, it can be assumed that gm= DFY(fm, N) =[gm(o), gm{ 1),... gm(N-i)].

Taking the element index Um of Gm to consist a new vector as:

(3) where cm(N) denotes the power compensation coefficient for the 7Vth subcarrier in the mth subband. The data symbols on the IVth subcarrier in the mth subband are multiplied by the corresponding power compensation coefficient cm(N) to obtain the power compensated transmitter symbol, given by bm =[bm(o), bm( 1),... ,bm(N,„-i)].

Although power compensation is applied in the present embodiment, in other embodiments the pre-filter CP can be applied without the use of power compensation. Similarly, in other embodiments power compensation can be applied without the use of a pre-filter CP.

In the present embodiment the power compensation factors are calculated in advance based on the known impulse response of the sub-band filter, and stored. The precoding unit 302 can subsequently retrieve the relevant power compensation factor for a particular subcarrier based on the subcarrier index. In other embodiments the power compensation factors may be calculated substantially in real-time, for example, if the allocation of subcarriers among subbands changes or if the filter length is changed.

Additionally, in the present embodiment the power compensation unit is configured to apply power normalization to constrain/normalize the total transmission power at the mth subband. Normalization in this context means that the power compensation unit is configured to allocate substantially equal power to each subcarrier within a given subband. Subcarriers in different subbands may have the same power as each other, or may have different powers. Normalizing the power within a subband simplifies the process of deriving the power compensation coefficients. In some embodiments, power normalization per subband may be omitted. For example, instead of normalizing the power within each subband a parameter a; can be defined to represent the power allocated to the Ith subband, and may be adjusted dynamically according to changes in the system configuration or the operating environment.

Next, in step S402 the time domain conversion unit 304 converts the power compensated data symbols into a time domain multicarrier signal, and in step S403 the cyclic prefix unit 306 adds a cyclic prefix to the time domain multicarrier signal.

Finally, in step S404 subband filtering is performed using the subband filter 308. The multi carrier signals for the plurality of subbands can then be summed and transmitted via the antenna 310.

Referring now to Figs. 5 and 6, an apparatus and method for processing the received CP-UFMC signal are illustrated according to the first embodiment. The CP-UFMC receiver comprises an antenna 502 configured to receive the CP-UFMC signal and an RF-to-baseband conversion unit 504 for converting the CP-UFMC signal to baseband. The RF-to-baseband conversion unit 504 outputs time domain samples of the CP-UFMC signal. The apparatus for processing the received CP-UFMC signal comprises a cyclic prefix removal unit 506, a frequency domain conversion unit 508, and a symbol detector 510. The cyclic prefix removal unit 506 is configured to remove the CP that was added to the m4 subband by the cyclic prefix unit 306 of Fig. 1, in step S601. After CP removal, the frequency domain conversion unit 508 converts the signal samples from the time domain to the frequency domain in step S602. In the present embodiment a discrete Fourier transform (DFT) is used, but in other embodiments a different domain transform method maybe used. Then, in step S603 the symbol detector 510 equalizes the frequency-domain signal to obtain the estimated data symbols.

The receiver in the present embodiment is similar to that of a CP-OFDM receiver, and hence the CP-UFMC system of the present embodiment provides backwards compatibility with legacy CP-OFDM receivers. Adding the CP before the subband filter in the transmitter, as in the present embodiment, avoids zero padding, down-sampling and high-order DFT operation which would otherwise be required if the CP was added after the subband filters in a UFMC system.

Second embodiment: ZP/TC-UFMC with power compensation Apparatus and methods according to a second embodiment of the present invention will now be described with reference to Figs. 7 to 11. Figures 7 and 8 schematically illustrate a UFMC transmitter and a method performed by the UFMC transmitter, according to the second embodiment of the invention. Whereas the first embodiment applies a CP before the subband filters, the system of the second embodiment employs an alternative solution to the problem of avoiding ISI. In the second embodiment, the transmitter applies zero-padding (ZP) and/or tail-cutting (TC) after the subband filters, instead of using a CP. The system of the second embodiment may therefore be referred to as a ZP/TC-UFMC system.

Similar to the transmitter of the first embodiment, the transmitter of the second embodiment comprises Mbranches, each of which includes apparatus for generating a filtered time-domain multicarrier signal for one of the M subbands within the UFMC signal. The apparatus on each branch comprises a precoding unit 702 for performing power compensation as described above in the first embodiment, a time domain conversion unit 704, and a subband filter 706. The UFMC transmitter further comprises a zero padding unit 708 configured to apply zero padding and/or tail cutting (ZP/TC) after the subband signals have been filtered and summed, and an antenna 710 for transmitting the UFMC signal.

The length of the zero padding may depend on the length of channel and the overhead budget of the system. In an ideal case, to achieve interference-free one-tap channel equalization, the zero padding length should be longer than the channel length minus one (in UFMC samples: T/Nwith Tbeing the symbol duration in seconds). To put it another way, the zero padding length should be greater than or equal to the channel length. In many real-world scenarios it may not always be possible to meet this ideal ZP length. For example, there may not be sufficient overhead to include zero padding of length L-1 samples. In this case, the filter tails can be cut in order to reduce overhead, albeit at the expense of performance loss. Tail cutting involves discarding samples from a tail region of the filtered multicarrier signal, on either side of the subband. The filter tails are preferably cut from both sides of the subband, but in some embodiments tail-cutting may only be performed on one side of the subband.

As shown in Fig. 8, the method of the present embodiment includes a step S801 of applying power compensation using the precoding unit 702. However, in other embodiments the power compensation step S801 and the precoding unit 702 may be omitted. Next, in step S802 the time domain conversion unit 704 transforms the data symbols for the m* subband into a time domain multicarrier signal, and in step S803 the subband filter 706 filters the time domain multicarrier signal. Then, in step S804, zero padding is added according to the current length L of the communication channel over which the UFMC signal is to be transmitted. As explained above, zero padding of length greater than L-1 is added if possible, and if not, tail cutting can optionally be performed.

As an alternative to zero padding, in some embodiments tail cutting (TC) can be performed to increase system efficiency. For example, as the filter length increases the system overhead also increases as the filter tails become longer. Tail cutting can be performed when a long filter is used, to reduce the overhead and increase system efficiency. Also, although in the present embodiment tail cutting is applied after the subband signals have been added together, in other embodiments tail cutting can be performed after the subband signals have been filtered but before the subband signals are added together.

Referring now to Figs. 9,10 and 11, an apparatus and method for processing the received ZP/TC-UFMC signal are illustrated according to the second embodiment.

The ZP/TC-UFMC receiver comprises an antenna 902 configured to receive the ZP/TC-UFMC signal and an RF-to-baseband conversion unit 904 for converting the ZP/TC-UFMC signal to baseband. The RF-to-baseband conversion unit 904 outputs time domain samples of the ZP/TC-UFMC signal. The apparatus for processing the received ZP/TC-UFMC signal comprises a processing unit 906, a frequency domain conversion unit 908, and a symbol detector 910.

In step S1001 the processing unit 906 performs post-data processing as illustrated in Fig. 11, by adding a plurality of the samples from a first signal portion at the beginning of the multicarrier signal to a plurality of the samples from a second signal portion at the end of the multicarrier signal. The number of samples moved is equal to L-i+LZp, where L is the filter length, and LZP denotes the number of samples added at the transmitter if zero padding was applied, or the number of samples removed at the transmitter if tail cutting was applied. The value of LZP is positive when zero padding is applied, and is negative when tail cutting is performed. The first L-i+LZP samples are then discarded, resulting in a signal of total length N samples being sent to the frequency domain conversion unit 908.

In some embodiments, a system may be configured to always operate as either a ZP-UFMC system or as a TC-UFMC system, and the post-data processing unit in the receiver can be set to always implement the appropriate algorithm as shown in Fig. 11. In other embodiments a system may switch dynamically between the ZP-UFMC and TC-UFMC, for example according to changes in channel conditions. In some embodiments, the transmitter may be configured to signal to the receiver whether ZP-UFMC or TC-UFMC is being used, and the receiver can select the appropriate post-data processing method.

For example, the transmitter and receiver may agree in advance whether tail cutting will be performed. In response to a determination that tail cutting is not being implemented at the transmitter, the receiver can set the length of the first signal portion to L-i. In response to a determination that tail cutting is being implemented at the transmitter, the receiver can set the length of the first signal portion to L-i+Lzp.

After post-data processing is completed, a signal of length N samples is obtained. The frequency domain conversion unit 908 converts the N signal samples from the time domain to the frequency domain in step S1002. In the present embodiment an AT-point DFT is used, but in other embodiments a different domain transform method maybe used. Then, in step S1003 the symbol detector 910 equalizes the frequency-domain signal to obtain the estimated data symbols.

By implementing post-data processing as shown in Fig. 11, reducing the sample length to N samples, an N-point DFT can then be performed. The second embodiment therefore provides a receiver with lower computational complexity receiver than a UFMC receiver with a CP added after subband filtering, which requires down-sampling and a 2iV-point DFT to be implemented at the receiver.

Channel equalization method A method of performing channel equalization at a receiver in a UFMC system will now be described with reference to Fig. 12, according to a third embodiment of the present invention. The method can be implemented in any of the systems described above in relation to the first and second embodiments. The method may also find use in other types of system, for example, in an OFDM receiver.

First, in step S1201 the receiver obtains an estimate of one or more parameters that relate to transmitter imperfections. The number and type of parameters used to describe the transmitter imperfections can vary between embodiments. In the present embodiment a plurality of parameters are estimated, including the channel autocorrelation matrix carrier frequency offset (CFO) and timing offset (TO). In addition, the receiver may also collect information about transmission parameters such as zero padding length, filter length, filter type and so on, to be used during equalization.

In some embodiments the transmission parameters may be communicated to the receiver by the transmitter, or may be estimated by the receiver. For example, in one embodiment a number of predefined categories maybe defined, each including different values for the parameters. The category with parameters which most closely matches the current transmitter configuration maybe signalled to the receiver by means of an index which identifies the particular category among the plurality of predefined categories.

Next, in step S1202 equalization is performed based on the estimated parameters, and/or known transmission parameters such as the zero padding length, filter length and filter type. In the present embodiment the equalizer Wn for the nth subcarrier in the mth subband is defined as:

where v is a parameter defined by:

and

with

and

where Lou is the channel length between the transmitter and receiver of the mth subband, zm is the CFO for the mth user, and Tm is the TO for the mth user. R(ij) is the channel autocorrelation matrix at the ith path and time j, σ2 denotes the noise power, and psym2 denotes the transmitter signal/symbol power. ZF refers to a Zero-Forcing receiver, and MMSE refers to a Minimum Mean Square Error receiver. This is merely one exemplary mathematical expression of the equalizer Wn, and in other embodiments a different form of equalizer can be used depending on the parameters to be taken into consideration.

Simulation Results

Figures 13,14 and 15 illustrate simulation results comparing the performance of the CP-UFMC system of the first embodiment to that of a conventional UFMC system, and an OFDM system utilizing the equalization method described above. Since the performance of the ZP/TC-UFMC system of the second embodiment is similar to that of the CP-UFMC system of the first embodiment, only the results for the CP-UFMC system are shown here. The simulation results were obtained using the following system parameters: turbo-coding (CR=i/3), quadrature phase shift keying (QPSK), fast

Fourier transform (FFT) size = 2048, Long Term Evolution (LTE) extended typical urban (ETU) channel, carrier frequency offset (CFO) = 0.1 & 0.03, timing offset (TO) = 0.05 & 0.02, number of occupied subcarriers = 120, and subband size = 12.

Figure 13 is a graph comparing the performance of a UFMC system according to an embodiment of the present invention to that of a UFMC system and an OFDM system. The graph plots the BER versus the energy per bit to noise power spectral density ratio (Eb/No). As shown in Fig. 13, the CP-UFMC system outperforms both the conventional UFMC system and the OFDM system, in both low and high level interference scenarios.

Figure 14 is a graph comparing the out-of-band (OoB) emission of a UFMC system according to an embodiment of the present invention to that of a UFMC system and a FBMC system. As shown in Fig. 14, the use of precoding to apply power compensation in the CP-UFMC system does not increase the OoB emission in comparison to conventional UFMC.

Figure 15 is a graph comparing the peak-to-average power ratio (PAPR) of a UFMC system according to an embodiment of the present invention to that of a UFMC system and an OFDM system. As shown in Fig. 15, the use of precoding to apply power compensation in the CP-UFMC system does not change PAPR significantly. From Figs. 14 and 15 it can be concluded that the use of precoding is not detrimental to the overall system performance, and provides the added benefit of more efficient power allocation across the subband.

Embodiments of the invention have been described which can combat ICI/ISI, for example by adding a CP before the subband filter as in the first embodiment, or by performing ZP/TC as in the second embodiment. Precoding schemes are also disclosed which compensate for filter gain selectivity that would otherwise lead to performance loss. In addition, new channel equalization methods are disclosed which take into account transceiver imperfections and insufficient ZP/CP length. Simulation results have shown that the proposed methods and systems can improve the system performance while maintaining a similar OoB emission and PAPR performance in comparison to known UFMC systems.

Whilst certain embodiments of the invention have been described herein with reference to the drawings, it will be understood that many variations and modifications will be possible without departing from the scope of the invention as defined in the accompanying claims.

Claims

1. A method of generating a multicarrier signal, the method comprising: transforming a plurality of data symbols to be transmitted on a plurality of subcarrier frequencies within a sub-band into a time domain multicarrier signal; and filtering the time domain multicarrier signal using a sub-band filter, wherein the method further comprises: adding samples to the time domain multicarrier signal before transmission by applying a cyclic prefix or zero padding; or removing samples from the filtered time domain multicarrier signal before transmission by performing tail cutting.

2. The method of claim 1, wherein zero padding is applied according to the channel length by adding a number of zeroes equal to or longer than a length of the current channel.

3. The method of claim 1, wherein tail cutting is performed by discarding samples from tail regions on both sides of the sub-band.

4. The method of claim 1, wherein the cyclic prefix is added before filtering the time-domain multicarrier signal using the sub-band filter.

5. The method of any one of the preceding claims, further comprising: multiplying each one of the plurality of data symbols by a power compensation factor, before transforming a plurality of data symbols into the time domain multi carrier signal, wherein the power compensation factor to be applied to a data symbol is determined based on the subcarrier frequency at which said data symbol is to be transmitted and a known impulse response of the sub-band filter.

6. The method of claim 5, further comprising: normalising the total transmission power of the filtered time domain multicarrier signal to match a transmission power of one or more other sub-bands included in the multicarrier signal.

7. A method of processing a received multicarrier signal, the method comprising: obtaining a plurality of time domain samples of the multicarrier signal; adding a plurality of the samples from a first signal portion at the beginning of the multicarrier signal to a plurality of the samples from a second signal portion at the end of the multicarrier signal; converting the multicarrier signal from the time domain to the frequency domain; and equalizing the frequency-domain multicarrier signal to obtain estimated data symbols.

8. The method of claim 7, further comprising: setting a length of the first signal portion and a length of the second signal portion according to whether tail cutting or zero padding was applied at a transmitter.

9. The method of claim 8, wherein the length of the first and second signal portions is set as L-i+Lzp, where L is the length of a filter used at the transmitter and Lzp is the length of tail cutting or zero padding applied by the transmitter.

10. The method of claim 7, 8 or 9, wherein equalizing the frequency-domain multi carrier signal comprises: obtaining an estimate of one or more parameters relating to transmitter imperfections; and performing one-tap equalization according to the obtained one or more parameters.

11. The method of claim 10, wherein the one or more parameters relating to transmitter imperfections include one or more of: a carrier frequency offset; a timing offset; and a channel autocorrelation matrix.

12. Apparatus for generating a multicarrier signal, the apparatus comprising: a time domain conversion unit configured to transform a plurality of data symbols to be transmitted on a plurality of subcarrier frequencies within a sub-band into a time domain multicarrier signal; a sub-band filter configured to filter the time domain multicarrier signal; and a sample adding unit configured to add samples to the time domain multicarrier signal before transmission by applying a cyclic prefix or zero padding or a tail cutting unit configured to remove samples from the filtered time domain multicarrier signal before transmission.

13. The apparatus of claim 12, further comprising: a precoding unit configured to multiply each one of the plurality of data symbols by a power compensation factor, before the plurality of data symbols are transformed into the time domain multicarrier signal by the time domain conversion unit.

14. Apparatus for processing a received multicarrier signal, the apparatus comprising: a signal receiver configured to obtain a plurality of time domain samples of the multi carrier signal; a processing unit configured to add a plurality of the samples from a first signal portion at the beginning of the multicarrier signal to a plurality of the samples from a second signal portion at the end of the multicarrier signal; a domain converting unit configured to convert the multicarrier signal from the time domain to the frequency domain; and an equalizing unit configured to equalize the frequency-domain multicarrier signal to obtain estimated data symbols.

15. The apparatus of claim 14, further comprising: an equalizer configured to obtain an estimate of one or more parameters relating to transmitter imperfections, and perform one-tap equalization according to the obtained one or more parameters.