GB2513839A

GB2513839A - Transmitter and method of transmitting

Info

Publication number: GB2513839A
Application number: GB1305799.7A
Authority: GB
Inventors: Samuel Asangbeng Atungsiri
Original assignee: Sony Corp
Current assignee: Sony Corp
Priority date: 2013-03-28
Filing date: 2013-03-28
Publication date: 2014-11-12
Also published as: WO2014155065A1; GB201305799D0

Abstract

The invention proposes frames of OFDM symbols (see Fig. 5), possibly for use in DVB systems dealing with both static and mobile receivers. The frames commence with a symbol including control signalling and possibly reference symbol sequences for frame synchronisation 406. These control signals may have reduced numbers of sub-carriers compared with the payload data symbols 402/404 and consequently shorter durations in the time domain. Hence a larger guard interval fraction (10c) is selected to ensure that the guard interval is adequate (unlike 10b).

Description

TRANSMITTER AND METHOD OF TRANSMITTING

Field of the Disclosure

The present disclosure relates to transmitters and methods of transmitting payload data using Orthogonal Frequency Division Multiplexed (OFDM) symbols.

BackEround of the Disclosure

There are many examples of radio communication systems in which data is communicatcd using Orthogonal Frequency Division Multiplcxing (OFDM). Systcms which have been arranged to operate in accordance with Digital Video Broadcasting (DYB) 1 0 standards for example, use OFDM. OFDM can be generally described as providing IC narrow band sub-carriers (where IC is an integer) which are modulated in parallel, each sub-carrier communicating a modulated data symbol such as Quadrature Amplitude Modulated (QAM) symbol or Quadrature Phase-shift Keying (QPSK) symbol. The modulation of the sub-carriers is formed in the frequency domain and transformed into the time domain for transmission. Since the data symbols are communicated in parallel on the sub-carriers, the same modulated symbols may be communicatcd on each sub-carrier for an extended period, which can be longer than the coherence time of the radio channel. The sub-carriers are modulated in parallel contemporaneously, so that in combination the modulated carriers form an OFDM symbol. The OFDM symbol therefore comprises a plurality of sub-carriers each of which has been modulated contemporaneously with different modulation symbols. During transmission, a guard interval filled by a cyclic prefix of the OFDM symbol precedes each OFDM symbol. When present, the guard interval is dimensioned to absorb any echoes of the transmitted signal that may arise from multipath propagation or other transmitters transmitting the same signal from a different geographic location.

As indicated above, the numbcr of narrowband carriers IC in an OFDM symbol can be varied depending on operational requirements of a communications system. The guard interval represents overhead and so may be minimized as a fraction of the OFDM symbol duration in order to increase spectral efficiency. For a given guard interval fraction, the ability to cope with increased multipath propagation whilst maintaining a given spectral efficiency can be improved by increasing the number K of sub-carriers thereby increasing the duration of the OFDM symbol. However, there can also be a reduction in robustness in the sense that it may be more difficult for a. receiver to recover data transmitted using a high number of sub-carriers compared to a. smaller number of sub-carriers, because for a fixed transmission bandwidth, increasing the number of sub-carriers K also means reducing the bandwidth of cach sub-carrier. A reduction in the separation between sub-carriers can make demodulation of the data from the sub-carriers more difficult for example, in the presence of Doppler frequency shifts. That is to say that although a larger number of sub-carriers (high order operating mode) can provide a greater spectral efficiency, for some propagation conditions, a target bit error rate of communicated data may require a higher signal to noise ratio to achieve than required for a lower number of sub-carriers.

Summary of Disclosure

According to an aspect of the present disclosure there is provided a transmitter transmits payload data using Orthogonal Frequency Division Multiplexed (OFDM symbols.

The transmifter comprises a frame builder confignred to receive the payload data to be transmitted and to receive signalling data for use in detecting and recovering the payload data.

at a receiver, and to form the payload data with the signalling data into frames for transmission. A modulator is configured to modulate a first OFDM symbol with the signalling data fomiing a part of each of the frames and to modulate one or more second OFDM symbols with the payload data to form one or more of the frames, and a transmission unit for transmitting the first and second OFDM symbols. The first OFDM symbol is a first type having a number of sub-carriers which is less than or equal to the number of sub-carriers of the one or more second OFDM symbols of a second type and a guard inlerval for the first OFDM symbol is selected in dependence upon the longest possible guard interval of the second OFDM symbol. Accordingly an OFDM communications system can he formed iii which data is transmitted using a frame structure in which a guard interval is adapted to allow a mix of different types of OFDM symbols. The mix of OFDM symbols allows the signalling data to be carried by the OFDM symbol of the first type and the payload data to be carried by the OFDM symbol of the second type, and the number of sub-carriers of the OFDM symbols of first type carrying the signalling data is less than or equal to the number of sub-carriers of the OFDM symbol of the second type. If the number of sub-carriers of the OFDM symbol of the first type is less than the number of OFDM symbols of the second type then there is an improved likelihood of the signalling data being detected and recovered before the payload data, so that a more robust communications system can be formed.

According to an arrangement in which embodiments of the present disclosure find application there is a requirement to provide a "preamble" OFDM symbol in a transmission frame. which carries signalling parameters to indicate, for example, at least some of the communications parameters which were used to encode and to modulate payload data onto the data bearing OFDM symbols whereby after detecting the signalling data within the first (preamble) OFDM symbol the receiver can recover the transmission parameters in order to detect the payload data from the data bearing OFDM symbols.

In the following description the first OFDM symbol may he a preamble OFDM symbol or form part of one in a transmission frame and so may be referred to as a preamble OFDM symbol and because this is arranged to carry signalling data, it may be referred to as a signalling OFDM symbol.

According to one embodiment a number of sub-carriers used for the OFDM symbols carrying signalling data may be different from the number of sub-carriers used for the OFDM symbols which are used to carry the payload data. For example in order to improve a likelihood of recovering the signalling data, making it more robust for detection in more challenging radio enviromnents the number of sub-carriers may be smaller than for the OFDM symbols carrying payload data. For example, the payload data bearing OFDM symbols may be requh-ed to have a high spectral efficiency and therefore for example the number of sub-carriers may be 16k (16384) or 32k (32768) whereas hi order to improve a likelihood that a receiver can recover the signalling data from the signalling OFDM symbols, the number of sub-carriers for the first signalling OFDM symbol may be a lower number, for example, a 4k (4096) or 8k (8192).

Various Thither aspects and features of the disclosure are defincd in the appended claims.

Brief Description of the Drawings

Embodiments of the present disclosure will now be described by way of example only with reference to the accompanying drawings wherein like parts arc provided with corresponding reference numerals and in which: Figure 1 is a schematic diagram illustrating an arrangement of a broadcast transmission network; Figure 2 is a schematic block diagram illustrating an example transmission chain for transmitting broadcast data via the transmission network of Figure 1; Figure 3 is a schematic illustration of OFDM symbols in the time domain which include a guard interval; Figure 4 is a schematic block of a typical receiver for receiving data broadcast by the broadcast transmission network of Figure 1 using OFDM; Figure 5 is a schematic illustration of a transmission framc for transmitting broadcast data including payload data and signalling data; Figure 6 is a block diagram showing a transmitter for transmitting signalling data via a signalling or preamble OFDM symbol according to one embodiment; FigtLre 7 is a schematic block diagram of a signature sequence generator according to one embodiment; Figure 8 is a graphicaJ plot of bit error rate with respect to signal to noise ratio in the prcscncc of a.dthtivc white Gaussian noise for coding rates of one half and one quarter; Figure 9 is a graphical plot of bit error rate with respect to a signature sequence back-off power from the power of the modulated signalling data which provides an acceptable performance according to the results of Figure 8; Figure 1 Oa is a schematic representation of OFDM symbols with a guard interval matched to an expected channel delay spread produced for a single frequency transmission network; Figure 1 Oh is a schematic representation of OFDM symbols with diffcrcnt numbers of sub-carriers per OFDM symbol with a guard interval selected as a fixed fraction of the related OFDM symbol duration; and Figure 1 Oc is a schematic representation of OFDM symbols with a diffcrcnt number of sub-carriers per payload data bearing OFDM symbol and a different numbcr of sub-carriers for a signalling OFDM symbol with guard interval selected to have a duration which is matched to both the payload and the signalling OFDM symbols; Figure 1 la is a schematic block diagram of a receiver for detecting and recovering signalling data from a signalling OFDM symbol according to the present technique, Figure 11 h is a schematic block diagram of a frcquency synchronisation detector which forms part of Figure 1 Ia, Figure 1 Ic is a schematic block diagram of a preamble guard interval correlator which forms part of Figure 11 b, Figure lId is an illustrative schematic block diagram of a further example of a coarse frequency offset synchronisation detector which forms part of the receiver of Figure ha, and Figure lIe is an illustrative schematic block diagram of a differential encoder which forms part of Figure lid; Figure 12 is a schematic block diagram of one example of a preamble detection and decoding processor which forms part of the receiver shown in Figure 11 a, which detects and removes the signature sequence in the frequcncy domain; Figure 13 is a schematic block diagram of one example of a preamble detection and decoding processor which forms part of the receiver shown in Figure ha, which dcl.ccts and removes the signature sequence in the time domain; Figure 14 is a schematic block diagram of an example of a signature sequence remover which forms part of the preamble detection and decoding processor shown in Figure 13; Figure iSa is a. schematic block diagram of a. matched filter, which is matched to the signature sequence for which an example generator is shown in Figure 7, and Figure 15b is a schematic block diagram of a signature sequence remover funning part of the receiver shown in Figure 14; Figure 1 6a is a graphical representation of a signal fonncd at the output of the matched filter; Figure 16b is an expanded view of the graphical representation shown in Figure 16b illustrating components of a channel impulse response; Figure 17 is a schematic block diagram illustrating a circuit for detecting a coarse frequency offset hi the receiver of Figure 11 a; Figure 18 is a graphical plot of the correlation output of the circuit shown in Figure 17 for a frequency offset of -88/Tn; Figure 19 provides a graphical plot of bit error rate with respect to signal to noise ratio for different code rates with and without a signature sequcncc added to the signalling OFDM symbol for rate uric half and rate one quarter codes; Figures 20a and 20b provide graphical plots of bit error rate against signal to noise ratio for a 0dB echo channel with two paths as illustrated in Figure 20c respectively with ideal and actual channel estimation.

Description of Example Embodiments

Embodiments of the present disclosure can be arranged to form a transmission network for transmitting signals representing data including video data and audio data so that the transmission network can, for example, form a broadcast network for transmitting television signals to television receiving devices. In some examples the devices for receiving the audio/video of the television signals maybe mobile devices in which the television signals are received while on the tnove Th other examples the audio/video data may he received by conventional television receivers which may be stationaiy and may be connected to a fixed antenna. or antennas.

Television receivers may or may not include an integrated display for television images and may he recorder devices including multiple tuners and demodulators. The antenna(s) may be inbuilt to television receiver devices. The connected or inbuilt antenna(s) may be used to facilitate reception of different signals as well as television signals.

Embodiments of the present disclosure are therefore configured to facilitate the reception of audio/video data representing television programs to different types of devices in different environments.

As will be appreciated, receiving television signals with a mobile device while on the move may be more difficult because radio reception conditions will be considerably different to those of a conventional television receiver whose input comes from a fixed antenna.

An example illustration of a television broadcast system is shown in Figure 1. In Figure 1 broadcast television base stations 1 are shown to be connected to a broadcast transmitter 2. The broadcast transmitter 2 transmits signals from base stations I within a coverage area provided by the broadcast network. The television broadcast network shoin in Figure 1 operates as a so called single frequency network in which each of the television broadcast base stations 1 transmit the ra.dio signals conveying audio/video data contemporaneously so that these can he received by television receivers 4 as well as mobile devices 6 within a coverage area provided by the broadcast network. For the example shown in Figure 1 the signals transmitted by the broadcast base stations 1 are transmitted using Orthogonal Frequency Division Multiplexing (OFDM) which can provide an arrangement for transmitting the same signals from each of the broadcast stations 2 which can be combined by a television receiver even if these signals are transmitted from different base stations 1.

Provided a spacing of the broadcast base stations I is such that the propagation time between the signals transmitted by different broadcast base stations 1 is less than or does not substantially exceed a guard interval that precedes the transmission of each of the OFDM symbols then a receiver device 4, 6 can receive the OFDM symbols and recover data from the OFDM symbols in a way which combines the signals transmitted from the different broadcast base stations 1. Examples of standards for broadcast networks that employ OFDM in this way include DVB-T. D\TB-T2 and TSDB-T.

An example block diagram of a transmitter forming part of the television broadcast base stations 1 for transmitting data from audio/video sources is shown in Figure 2. In Figure 2 audio/video sources 20 generate the audio/video data representing television programmes.

The audio/video data is encoded using forward error correction encoding by an eneoding/interleaver block 22 which generates forward error correction encoded data which is then fed to a modulation unit 24 which maps the encoded data onto modulation symbols which are used to modulate OFDM symbols. Depicted on a separate lower arm, signalling data providing physical layer signalling for indicating for example the format of coding and modulation of the audio/video data is generated by a physical layer signalling unit 30 and after being encoded by an encoding unit 32 the physical layer signalling data is then modulated by a modulation unit 24 as with the audio/video data.

A frame builder 26 is arrangcd to form the data to be transmitted with the physical layer data into a frame for transmission. The frame includes a time divided section having a preamble in which the physical layer signalling is transmitted and one or more data transniission sections which transmit the audio/video data generated by the audio/video sources 20. A symbol interleaver 34 may interleave the data which is formed into symbols for transmission before being modulated by an OFDM symbol builder 36 and an OFDM modulator 38. The OFDM symbol builder 36 receives pilot signals which are generated by a pilot and embedded data generator 40 and fed to the OFDM symbol builder 36 for transmission. An output of the OFDM modulator 38 is passed to a guard insertion unit 42 which inserts a guard interval and the resulting signal is fed to a digital t.o analogne convertor 1 5 44 and then to an RF front end 46 before being transmitted by an antenna 48.

As with a conventional arrangement OFDM is arranged to generate symbols in the frequency domain in which data symbols to be transmitted are mapped onto sub carriers which are then converted into the time domain using an inverse Fourier Transform. Thus the data to be transmitted is formed in the frequency domain and transmitted in the time domain. As shown in Figure 3 each time domain symbol is generated with a useful part of duration Tu and a guard interval of duration Tg. The guard interval is generated by copying a part of the useful part of thc symbol in the time domain. . By correlating the useful part of the burst with the guard interval, a receiver can be arranged to detect the useful part of the OFDM symbol Tu, from which data. can then he recovered fi-om an OFDM symbol by triggering a Fast Fourier I'ransfonn to convert the time domain symbol samples into the frequency domain. Such a receiver is shown in Figure 4.

In Figure 4 a receiver antenna 50 is arranged to detect an RF signal which is passed via a tuner 52 and converted into a digital signal using an analogue to digital converter 54 before the guard interval is removed by a guard interval removal unit 56. After detecting the optimum position for perfonning a fast Fourier Transform (FFT) to convert the time domain samples into the frequency domain, an FFT unit 58 transforms the time domain samples to form the frequency domain samples which are fed to a channel estimation and correction unit 60, The channel estimation and correction unit 60 then estimates the transmission channel for example by using pilot sub-carriers which have been embedded into the OFDM symbols. After excluding the pilot sub-carriers, all the data-bearing sub-carriers are fed to a symbol de-interleaver 64 which de-interleaves the sub-carrier symbols. A dc-mapper unit 62 then extracts the data bits from the sub-carriers of the OFDM symbol.'t'he data bits are fed to a bit de-interleaver 66, which performs the de-interleaving so that the error correction decoder can correct errors in accordance with a conventional operation.

Framing structure Figure 5 shows a schematic of the framing structure according to an example embodiment of the present technique. Figure 5 illustrates different physical layer frames, some targeted for mobile reception whilst others are targeted for fixed roof-top antenna reception.

The system can be expanded in future to incorporate new types of frames, for the current system, these potential new types of frames are simply known as future extension frames (FEEs).

One requirement for fixed reception frames is an improved spectral efficiency which may be assured by such features as adopting a higher order modulation, for example 256QAM, and higher code rates, for example greater than half rate, because of relatively benign channel conditions, and a high number of sub-carriers per 01DM symbol (FF1 size) such as 32K.

This reduces the capacity loss due to the guard interval fraction. However, a higher number of sub-carriers can make such OFDM symbols unsiLitable for mobile reception because of lower tolerance to high Doppler frequency of the received signal. On the othcr hand, the main requirement for mobile reception frames could be robustness in order to ensure a high rate of service availability. This can he improved by adopting such features as a low order modulation for example QPSK or BPSK, low code rates, a low number of sub-can-iers per OFDM symbol (FF1 size) and a high density scattered pilot pattern etc. A low number of sub-carriers for OFDM symbols can be advantageous for mobile reception because a lower number of sub-earricrs can provide a wider sub-carrier spacing and so more resilience to high Dopplcr frequency. Furthermore a high density pilot pattern eascs channel estimation in the presence of Doppler.

The framing structure shown in Figure 5 is therefore characterised by frames which may each include payload data modulated and encoded using different parameters. This may include for example using different OFDM symbol types having different number of sub-earners per symbol, which may be modulated using different modulation schemes, because different frames may he providcd for different types of receiver. however each frame may include at least one OFDM symbol cari-ying signalling data, which may have been modulated dilThrently to the one or more OFDM symbols carrying the payload data. Furthermore the signalling OFDM symbol may be a different type to the OFDM symbol(s) carrying the payload data. The signalling data is required to be recovered so that the payload data may be dc-modulated and decoded.

What characteristics for the preamble? To delimit frame boundaries, a frame preamble symbol such as the P1 symbol in DVB-T2 is required. The preamble symbol would carry signalling that describes how the following frame is built. It is expected that all of the types of receiver mentioned above whether mobile or with a fixed antenna should be able to detect and decode the preamble in order to determine whether or not they should decode the payload in the following frame.

Desirable characteristics for such a preamble include: 1. High Capacity of Signalling; The prcamble should have a high signalling capacity unlike the P1 preamble in DVB.-T2 with capacity of 7 signalling bits, a preambic more like in DVB-C2 with I OOs of signalling bits is desirable. This suggests that the preamble symbol should be an OFDM symbol with enough sub-carriers to eany all the signalling information.

2. Common Macro-structure; All frame preambles should have a common pre-defined macro-structure that is understood by all receiver types. This means that the preamble symbol should have for example a constant duration, constant number of sub-carriers and guard interval for all frame types. This forces a constraint that the guard interval must be similar in duration to the longest guard interval that may be used in fixed antenna reception, otherwise when the network uses this longest guard interval, the preamble symbol will suffer from excessive inter-symbol interference (1ST) and perhaps suffer decoding failure.

3. Low complexity detection and decoding: The preamble symbol detection and decoding complexity should be low enough to easily implement in battciy powered mobile receivers so as to make efficient use of limited stored power. This constrains the maximum FFT size and maximum FEC block length.

4. The preamble should be easily detected in the time domain; in DVB-C2, all 01DM symbols within the frame structure usc 4K subcarrier spacing. This means that the receiver can start with OFDM symbol time synchronisation followed by frequency domain frame synchronisation (preamble detection). In an embodiment of the present disclosure frames can he arranged such that OFDM symbols in different physical layer frames may have difference subcarrier spacing. Frequency domain frame synchronisation (preamble detection) is thus not readily possible. The preamble symbol must therefore be detected in the time domain. It is only after the preamble is iS decoded and its signalling payload interpreted that frequency domain processing of the frame can proceed because only then would the receiver have knowledge of the OFDM parameters (number of sub-carriers, guard interval) etc of the data payload bearing OFDM symbols in the body of the frame.

5. Robustness; The preamble should he detectable and decodeable by all receiver types under all channel conditions where such receivers are expected to work. This means that the preamble should be robust to both high levels of noise, low signal to noise ratios and high levels of Doppler shift as experienced during reception on the move.

Robustness to high levels of noise constrains the maximum transmission parameters for coding and modulation (MODCOD) that can be used for carrying the signalling payload of the preamble whilst robustness to Doppler constrains the minimum sub-carrier spacing of the preamble OFDM symbol. The preamble OFDM symbol must use a sub-carrier spacing that is large enough to be reasonably resilient to a high Doppler spread. Furthermore, the preamble OFDM symbol should also allow decoding in the presence of frequency shift, common phase error, maximum expected multipath delay spreads etc. As explained above the preamble 01DM symbol conveys signalling data whilst the OFDM symbols within the body of the transmission frame convey payload data as shown in Figure 5. Each transmission frame shown in Figure 5 has particular characteristics. A data bearing frame 100 carries a frame of data, which may use a higher operathig mode providing a higher number of sub-carriers per 01DM symbol, for example, approximately 32 thousand sub-carriers (32k mode) thereby providing a relatively high spectral efficiency, but requiring a relatively high siial to noise ratio to achieve an acceptable data integrity in the form of the bit error rate. The higher order operating mode would therefore be most suitable to communicate to stationary television receivers which have sensitive detection capabilities including well positioned fixed antenna for recovering audio/video data from the 32k OFDM symbols. In contrast, the frame structure also includes a second frame 102 which is generated to be received by mobile television receivers in a more hostile radio communications environment.

The frame 102 may therefore he arranged to form payload bearing 01DM symbols with a lower order modulation scheme such as BPSK or QPSK and a small or lower number of sub-carriers per 01DM symbol (FFT size) such as 4K or 8K to improve the likelihood that a mobile receiver may be able to receive and recover the audio/video data in a relatively hostile environment. In both the first frame 100 and the second frame 102 a preamble symbol 104,1 06 is provided which provides signalling parameters for detecting the audio/video data transmitted in the payload part of the transmission frame 100, 102. Similarly, a preamble symbol 108, 110 is provided for a future extension frame 112.

Design of New Preamble Symbol Some example embodiments can provide an arrangement for forming a preamble symbol for usc for cxample with the transmission frames shown in Figure 5 in which there is an improved likelihood of detecting the preamble symbol particularly in harsh radio enviromnents. Furthermore, the framing structure shown in Figure 5 can be devised such that the number of sub-carriers of the payload bearing OFDM symbols is different from frame to frame and furthermore, these sub-carriers may use different modulation schemes. Thus the OFDM symbols which cany the payload data may be of a different type to the OFDM symbols carrying the signalling data. An example block diagram of a part of the transmitter shown in Figure 2 for transmitting the signalling data is shown in Figure 6.

In Figure 6 the signalling data is first fed to a scrambling unit 200 which scrambles the signalling data which is then fed to a forward error correction (FEC) and modulator unit 202 which encodes the signalling data with a forward error correcting code and then interleaves it before mapping the encoded data onto a low order modulation constellation such as BPSK, DBPSK, it/4-BPSK and QPSK. A pilot insertion unit 204 then inserts pilots between modulation symbols to form one of the OFDM symbols of the preamble 104, 106, 108, 110.

The OFDM symbol formning the preamble is then scaled by a scaling unit 206 in accordance with a predetermined factor (I -0). The sealing unit 206 adapts the transmission power of the preamble with respect 1.0 a. signature sequence which is combined with the OFDM symbol of the preamble before transmission so that the total transmission power of the preamble remains the same as it would have been without the signature sequence.

According to the present the technique a signature sequence generator 208 is configured to generate a signature scquenec which is fed to a second sealing unit 210 which scales the signature sequence by a predetermined factor G before the scaled signature sequence is combined with the OFDM symbol of the preamble by a combining units 212. Thus the signature sequence Wtlç) is combined with the OFDM symbol in the frequency domain so that each of the coefficients of the signature sequence is added to one of the subcarricr signals of the OFDM symhol. The combined preamble OFDM symbol and signature sequence are then transformed from the frequency domain to the time domain by an inverse Fourier transform processor (IFFT) 214 before a guard interval insertion unit inserts a time domain guard interval. At an output of the guard insertion unit 216 the preamble symbol is formed on output channel 218.

As can be seen for the example shown in Figure 6 the signature sequence is combined with the OFDM symbol carrying signalling data in the frequency domain so that a spectrum of the preamble symbol after combining remains within a spectral mask for the tramsmission channel. As will he appreciated for some examples the signature sequence may be combined with the OFDM symbol in the time domain. 1-lowever other bandwidth limiting processes must then be introduced after the combination of the signature sequence with the preamble OFDM symbol in the time domain which may affect the correlation properties of the signature sequence at the receiver.

In the example illustration in Figure 6, the scrambling of the signalling data by the scrambling unit 200 ensures that the peak-to-average power ratio (PAPR) of the preamble symbol will not be excessive due to many similarly modulated OFDM sub-carriers. The scrambled signalling bits are then forward error correction encoded by the FEC and BPSIC unit 202 with a code such as 4K LDPC code at a low code rate (1/4 or 1/5) prior to mapping a low order modulation constellation such as BPSK, DBPSK, it/4-BPSK and QPSK within the unit 202. The pilots inserted at this stage by the pilot insertion unit 204 are not for chmmel estimation, but for frequency offset estimation as will be explained shortly. At this stage, a complex preamble signature sequence also composed of the same number of complex samples as the useful sub-carriers as the OFDM symbol is added to the samples of the signalling OFDM symbol by the combiner 212. When generated, each preamble signature sequence sample is a point on the unit circle but before addition to the preamble OFDM symbol, each sample is scaled by a predetermined factor 0, by a scaler 210 and the corresponding OFDM symbol sample is scaled by (I-C) by a scaler 206 so that the power of the composite preamble symbol should he the same as the power of the signalling OFDM symbol at point A in Figure 6.

The IFFT 214 then forms the OFDM symbol in the time domain, which is then followed by the insertion of the guard interval by the guard insertion unit 216 which prepends the Ng samples of the preamble OFDM symbol at the start of the preamble OFDM symbol -also known as the cyclic prefix of the preamble OFDM symbol. After guard interval insertion, a preamble OFDM time domain symbol of duration Ts = lu + Tg made up of Ns = Nu + Ng complex samples where Tu is the useful symbol period with Nu samples and Tg is the guard interval duration with Ng samples is formed.

The Sianature Sequence Generator As explained above, the preamble symbol generator of Figure 6 generates a signature sequence which is combined with the signalling OFDM symbol (first OFDM symbol), which forms the preamble symbol of the frame, in order to allow a receiver to detect the preamble at lower signal to noise ratios compared to signal to noise ratios which are required to detect and recover data from OFDM symbols carrying the payload data. The signature sequence generated by the signature sequence generator 208 can be formed using two pseudo random sequence generators, one for the in-phase and other for the quadrature phase component.

In one example the signature sequence is a constant amplitude zero autoconelation (CAZAC) or Zadoff and Chu sequence. In other examples the signature sequence is formed from a pair of Gold code sequences chosen because of their good auto-correlation properties, or other example signature sequence could be used such as from a pair of M-sequenees.

One example of the signature sequence generator 208 shoxvn in Figure 6 is shown in more detail in Figure 7. Figure 7 is arranged to generate a complex signature sequence which is added to the complex samples of the signalling OFDM symbol by the combiner 212 shown in Figure 6.

In Figure 7 two linear feedback shift registers are used in each case to generate a pair of pseudo random bit sequences for the in-phase 300.1 and 300.2 and quadrature 302.1 and 302.2 components. Tn each case, the pseudo-random bit sequence pair is combined using exclusive-OR circuits 310, 312 to produce the Gold sequences for the in-phase (300.1 and 300.2) and quadrature (302.1 and 302.2) part of the signature sequence, respectively. A binary to bipolar mapper unit 3 14, 316 then forms respectively a sample for the in-phase 318 mid quadrature (imaginary) 320 components of the signature sequence. Effectively, the arrangement shown in Figure 7 generates Gold codes formed by XORing two rn-sequences.

The rn-sequences are generated by the linear feedback shift registers 300, 302. A table 1 below shows the generator polynomials for the linear feedback shift registers according to the example shown in Figure 7: Sequence Name Generator polynomial Req1 x'3 + x" +x + 1 Rseq2 K13 +x9+x5+ 1 1_seqi + + 1 Tseq2 ______________________ Table 1: Generator polynomials for complex signature sequence.

Determining an Optimum Value for the Scaling Factor G As shown in Figure 6, the scaler 210 multiplies the signature sequence by a factor G and the scaler 206 multiplies the signalling OFDM symbol by a. factor 1-G. As such, if the time domain signalling OFDM symbol signal is c(n) while the signature sequence signal is fit), then the composite transmitted preamble symbol s(n) is given by: s(n) = (1-G)c(n)+Gf(n) where G is the scaling applied to the signature sequence. The signature signal effectively adds distortion to the signalling OFDM symbol thereby increasing thc bit error rate of the signalling OFDM symbol at thc receiver. Furthermore, with a normaiised power of 1, the composite symbol in effect distributes power between the signature signal and the signalling OFDM symbol signal. With a high value for C, the signature signal has more power and so frame synchronisation (detection of the preamble) at the receiver should be achieved at a lower signal to noise ratio. However, reducing the power of the signalling OFDM symbol (in order to increase the power of the signature signal) also means that error-free decoding of the signalling information itself becomes more difficult at the receiver as the signal-to-noise of the signalling OFDM symbol has fallen. Therefore, an optimum value for G has to be a compromise between these conflicting aims. We can further define P (1 -G)IG which is proportional to the power ratio between the signalling OFDM symbol and the signature signal.

An appropriate value for U can be set by experimenting with this power ratio P. The performance of example error correction codes which may be used for protecting the preamble symbol can be assessed in the presence of Additive White Gaussian Noise, using an appropriate constellation for the signalling information. For example a QPSK modulation scheme can be used with example error correction codes. In the present example 4K LDPC half rate and quarter rate codes were evaluated. Figure 8 provides a graphical illustration of the performance for communicating the signalling data using the signalling OFDM. symbol for these half and quarter rate LDPC codes and shows for each code a hit error rate performance with respect to signal to noise ratios for an additive white Gaussian noise channel. It can be seen that at a signal to noise ratio of -3dB and a signal to noise ratio of 1 dB, the quarter rate and half rate codes respectively each become error free.. These values of signal to noise ratios were then increased to -2dB and 2dB respectively and then the signature signal added with values of P varied until a bit error rate of zero was achieved.

As will be appreciated the error correction code which may be used to protect the signalling data carried in the preamble symbol may have coding rates whch are different to rate one-half and rate one-quarter. In sonic embodiments the coding rate is less than or equal to one-quarter. In one example the coding rate is one-fifth (1/5).

Figure 9 provides a graphicai plot for code rates of one quarter and one half showing a hit error rate for each code rate as the factor p on the x-axis and SNR fixed to -2dB and 2 dB respectively. As can be seen from these results setting P = 8dB will give a bit error rate close to zero, despite the presence of the signature sequence, which has been added to the signalling OFDM symbol. It can also he seen experimentally, that with this value of the factor P, 1.0 preamble detection can be achieved. A value of P = 8dB has, therefore, been adopted for the different half and quarter rate code rates with Q.PSK modulated data subearriers of the signalling O.FDM symbol. As can be seen an optiniising choice for the factor P can he chosen from the results produced.

Determining a Suitable Guard Interval Fraction According to example embodiments of the present technique, the same preamble symbol will dclhnit physical layer frames meant for both fixed and mobile reception. In the following analysis it is assumed that a broadcast transmission system, which has both types of transmission frames will be used. As such one of the principal factors affecting the reception of payload data bearing OFDM symbols transmitted for fixed reception is spectral efficiency.

As explained above, this means the use of large numbers of sub-carriers for the OFDM symbols and correspondingly large FF1' sizes because a smaller guard interval fraction (GIF) can he used to get a large guard interval duration (GD). A large GID can allow a broadcast system to have a greater separation between broadcast transmitters and can operate in environments with a greater delay spread. In other words the broadcast transmission system is configured with a wider spacing between transmitters foniiing a single frequency network (SFN).

Figure 10 illustrates how the selection of the guard intervals can be affected when different operating modes providing different numbers of sub-carriers per OFDM symbol (different FFT sizes) al-c used for different frames in the same transmission. In contrast to the diagram shown in Figure 5, the diagram shown in Figure 10 is in the time domain. Three sets of OFD.M symbols are shown in the time domain illustrative of what may happen at the point where one frame ends and another starts in a single transmission. In Figure lOa the duration of the last OFDM symbol 402 of the ending frame is the same as that of the first OFDM symbol 404 of the starting frame. The unshaded area 405 between the two OFDM symbols 402 and 404 represents the guard interval that precedes symbol 404. In Figure lOb an example of a preamble symbol shown as the light grey area 406 is inserted to delimit the two frames. As can be seen, this example preamble symbol 406 has a shorter duration than the data hearing symbols 402 and 404 as a consequence of having a different number of sub-carriers per OFDM symbol. Accordingly, if the GIF for the preamble symbol is the same as for the data symbols, tbe guard interval duration for the preamble symbol will not be as long as that of the data bearing symbols. Accordingly, if the delay spread of the channel is as long as the guard interval of the data bearing OFDM symbol 402, then the preamble symbol 406 will suffer inter-symbol interference from the last symbol 402 of the previous frame. Examples shown in Figure 1 Oc can provide an arrangement in which the guard interval fraction for the preamble symbol is selected to the effect that the guard interval duration of the preamble symbol 406 1-I I) matches or may be longer than the guard interval duration of the last data bearing symbol 402 of the previous frame.

According to some example embodiments the largest number of sub-carriers per symbol is subsmntially thirty two thousand (32K). With a 32K FFT size in DVB-T2 for example, the largest G[F is 19/i 28. For 6M.z channel raster, this represents a GID of about 709.33us. When this GTD is used for the frame canying OFDM symbols targeted for fixed receivers, the preamble OFDM symbol GD should at least be of a similar value, otherwise, the preamble symbol will suffer inter-symbol-interference from the last symbol of a previous fixed reception frame.

In a 6MHz channel raster system in which for example DVB-T2 is transmitted, an OFDM symbol having substantially four thousand sub-carriers (4K) OFDM symbol has a duration of only 2*224*816 = 597.33us. Therefore even with a GTF = 1, it is not possible to get a GD of 709.33us with a 4K OFDM symbol. A table below lists possible operating modes that are receivable hi medium to high Doppler frequencies (for the mobile environment) and sonic possible guard intervals.

FFT Tu In 6MHz (us) G[F GID (us) Ts (us) Size 4K 597.33 1 597.33 1194.667 1⁄4 298.67 1493.338 8 19467 V2 597.33 1792.005 K 1. 19/32 709.33 1904.000 896.00 2090.638 Table 2: Mobile FFT modes and their possible guard intervals From the above table it can be seen that only an 8K operating mode for the preamble OFDM symbol has GIF < 1 which matches or exceeds the maximum OlD for a 32K maximum number of sub-carriers of the OFDM symbol. In conclusion therefore, embodiments of the present technique can provide a. number of sub-carriers for the signalling or preamble OFDM symbol of 8192 sub-carriers, which corresponds to all 8K FFT size, for which the 3W will be about 19/32. This means that the total sigualling OFDM symbol will have a duration of l's 1904us. Furthermore an 8K operating mode will have a sub-carrier spacing, which provides a mobile receiver with a reasonable chance of detecting and recovering the signalling data from the preamble OFDM symbol in medium to high Doppler frequencies. It can be understood that in embodiments of this disclosure, the GTE of the preamble symbol has to be chosen to have the same GID that is the same or longer than the longest OlD of the maximum FFT size available in the system.

Channel Estimation Considerations As known in OFDM transmission systems such as DVB-C2, frequency domain preamble pilots may be inserted into a preamble symbol at regular intervals for USC ill channel estimation and equalisation of the preamble symbol. A dcnsity of such pilots Dx, which is the spacing in the frequency is dependent on the maximum delay spread that can be expected on the channel. As explained above, with a single frequency transmission network, it can be advantageous to use a larger CuD. For such single frequency networks, a channel impulse response can have a duration which is equal to the GW. Thus, the delay spread of the channel for preamble equalisation may be as much as the GJD. When using preamble pilots spaced by Dx subearriers, pilot-aided channel estimation is possible for delay spreads as long as Tu/Dx.This means that Dx must be set such that: Since for an 8K preamble in a 6MBz channel, Tu = 1 194.67us, D«=r Tg Substituting Tu 1194.67 and Tg = 70923, D «= 2. This means that more than one in every two sub-carriers of the signalling OFDM symbol would become a pilot sub-carrier. This would have the effect of cutting the capacity of the signalling OFDM symbol by more titan half. As such, this conclusion suggests that an alternative technique should be adopted to estimate the channel impulse response rather than using frequency domain pilots.

Fregueny Offset Considerations At first detection, the signalling or preamble OFDM symbol may have to be decoded in the presence of any tuning frequency offsets introduced by tuner 52, This means that either the signalling data should be modulated unto the preamble OFDM symbol in a manner that reduces the effects of any frequency offsets or resources are inserted into the preamble symbol to allow the frequency offset to be estimated and then removed prior to preamble decoding. In one example the transmission frame may only include one preamble OFDM symbol per frame so the first option is difficult to achieve. For the second option, additional resources can be in the fonn of frequency domain pilot sub-carriers, which are inserted into the OFDM so that these can be used to estimate the frequency offset and common phase error. The frequency offsets are then removed before the symbol is equalised and decoded. In a similar vein to the insertion of pilots into the data payload bearing OFDM symbols, embodiments of the present technique can be arranged to provide within the signalling (preamble) OFDM symbol pilot sub-carriers, which can allow for the estimation of frequency offsets that arc larger than the preamble sub-carrier spacing. These pilots are not spaced regularly in the frequency dimension to avoid instances when multipath propagation may result in regular nulls of the pilots across the full preamble OFDM symbol. Accordingly, 1 80 pilot sub-carriers can he provided across the 8K symbol with the positions defmed apriori. The sub-FFT bin frequency offset is estimated via the detection of the preamble OFDM symbol itself Accordingly embodiments of the present technique can provide a preamble OFDM symbol in which the number of sub-carriers carrying pilot symbols is less than the number which would be required to estimate a channel impulse response through which the preamble O.FDM symbol is transmitted, but sufficient to estimate a coarse frequency offset of the transmitted OFDM symbol.

Freciueney Offset Detection at the Receiver As explained above the preamble is formed by combining an OFDM symbol carrying signalling data with a signature sequence. In order to decode the signalling data, the receiver has to first detect and capture the preamble OFDM symbol. In one example the signature sequence may be detected using a match filter which has impulse response which is matched to the conjugate of the complex samples of the known signature sequence. However any frequency offset in the received signal have an effect of modulating the output of the matched filter and preventing accurate detection of the signature sequence using a match filter. An example receiver for detecting the preamble and recovering the signalling infonnation provided by the preamble in the presence of a frequency offset is shown in Figure 1 la. In Figure 1 la, a signal received from an antenna is converted to a baseband signal, using a conventional arrangement as shown in Figure 4 and fed from an input 420 respectively to a complex number multiplier 422 and a frequency synehroniser 424. [lie frequency synchroniser 424 serves to detect the frequency offset in the received signal r(x) and feed a measure of the offset in respect of a number of subearriers to an oscillator 426. The oscillator 426 generates a complex frequency signal which is fed to a. second input of the multiplier 422 which serves to introduce a reverse of the offset into the received signal r(x). Thus the multiplier 422 multiplies the received signal r(x) with the output from the oscillator 426 thereby compensating or substantially reversing the frequency offset in the received signal so that a preamble detection and decoding unit 430 can detect the preamble OFDM symbol and recover the signalling data conveyed by the preamble which is output on output channel 432.

Figure 1 lb provides an example implementation of the frequency synchroniser 424 which forms part of the. receiver shown in Figure 1 la. In Figure 1 lb the received signal is fed from the input 420 to a preamble guard interval correlator 432 which generates at a first output 434 a signal providing an indication of the start of the useful part of the OFDM symbol. A second output 436 feeds the samples of the OFDM symbol to a Fourier transform processor 43K, but delayed by the number of samples Nn in the useful part. The first output 434 from the preamble guard interval correlator 432 detects the location of the guard interval and serves to provide a trigger signal from a threshold detector 440 to the FF1' 438 through a channel 442 which tnggers the FFT 438 to convert the Nu time domain samples of the useful part of the OFDM symbol into the frequency domain. The output of the Fourier transform processor 438 is fed to a continuous pilot (CP) matched filter unit 444, which correlates the pilot signals in the received OFDM symbol with respect to replicas at the receiver which are used to set an impulse response of the CP matched filter in the frequency domain. The matched filter 444 therefore correlates the regenerated pilots with the received OFDM symbol and feeds a result of the correlation to an input to a detection threshold unit 446. The detection threshold unit 446 detects an offset in the received signa.l in terms of the number of FFT bins on channel 448 which effectively provides the fiequency offset which is fed to the oscillator 426 for correcting the offset in the received signal.

Figure 1 Ic provides an example of implementation of the preamble guard interval correlator 432 and corresponds to a conventional arrangement for detecting the guard interval.

Detection is performed by cross con-elating the samples of the received OFDM symbol with themselves after a delay of Nu samples with the cross con-elation outputs accumulated over consecutive Ng sample intervals. Thus the received signal is fed from an input 420 to a multiplier 450 and a delay unit 452 which feeds an output to a complex conjugator 454 for multiplying by the multiplier 450 with the received signal. A delay unit 456 delays the samples by the number of samples Ng in the guard interval and a single delay unit 458 delays an output of an adder 460. The adder 460 receives from the multiplier 450 the results of multiplying the received signal with a conjugate of the delayed samples corresponding to the useful samples Nu which is then fed to the adder 460. Together adder 460, delay blocks 456 and 458 implement a moving average filter of order Ng whose effect is to accumulate successive outputs of the cross-eorrelator over Ng samples. Thus at a point 434 there is provided an indication of the detection of the useful part of the OFDM symbol by detecting the guard interval period. The output 436 provides the delayed received signal samples which are fed to the FFT in order to trigger the Fourier transform after the guard interval has been detected by the first output 434.

Figure lid provides another example of implementation of the frequency synchroniser 424 and corresponds to a first detection of the preamble symbol by use of a signature sequence matched filter 462. Firstly however, the differential encoder block 461 is used to alter the received signal so as to reduce the modulation of the matched filter output by any frequency offset present in the received signal. [he differential encoder 461 is applied both to the received signal and the time domain signature sequence which is generated by inverse Fourier transform 506 of the output of the frequency domain signature sequence generator 504. The signature sequence matched filter 462 to be described later in Figure iSa. is a finite impulse response filter whose taps are set to the coefficients of the differential encoded time domain signature sequence. The circuit shown in Figure ild therefore fonus an example of the frequency synchroniser 424 in which the signature sequence generator 504 re-generates the signature sequence, the inverse Fourier transformer 506 transforms the signature sequence into the time domain, and the differential encoder 461 compares differentially successive samples of the received signal to reduce a modulating effect of the frequency offset in the radio signal, and correspondingly compares differentially successive samples of the time domain version of the signature sequence. As already explained, the matched filter 462 has an impulse response corresponding to the differentially encoded signature sequence and receives the received signal from the differential encoder 461 and fillers the differentially encoded received signal to generate at an output an estimate of the coarse frequency offset.

Corresponding to output channel 434 in Figure 11 b. output channel 463 in Figure 1 id produces a signal which is fed to the threshold block 440 to generate a trigger for the FFT 438; whilst output channel 436 in Figure 1 lb corresponds to output channel 464 in Figure lId, This channel conveys the preamble OFDM symbol samples to the FFT block 438 which at the right moment is triggered by through channel 442 by the threshold block 440. Figure lIe provides an example of the differential encoding block 46 i. l'he received samples r(r) enter a unit delay element 465 and also a conjugation block 466. The delay element 465 delays each sample for one sample period while the conjugation element 466 changes each input sample to its conjugate at its output whose effect is to convert an input [r1(n) jr4(n)J into an output r(r) -jrq(n)]. This conjugated sample is then subtracted from the output of delay element 465 by the adder 467. For an input signal r1(mi) + jrq(n)] and output [yj(n) + jyq(n)J n0,l,2 the differential encoder 461 acts to implement the equation: [y(n) +jyq(n)1 = [r1(n-l) -r(n)] 1-j[rq(r-l) ± rq(n)] Accordingly before preamble detection and decoding is performed by the preamble detection and decoding unit 430 the frequency offset in the received signal is estimated and corrected by the arrangements shown in Figures 1 la and 1 lb and lie; or lid and lie.

Preamble Detection and Decoding at the Receiver As explained above for the example of the receiver shown in Figure 1 la, a preamble detector and decoder 430 is configured to detect the preamble symbol and to recover the signalling data from the preamble symbol. To this end, the preamble detector and decoder 430 detects the preamble by detecting the signature sequence and then removes the signature sequence before recovering the signalling data from the preamble. Example embodiments of the preamble dctector and decoder 430 are illustrated in Figures 12, 13 and 14.

Embodiments of the present technique can provide a receiver which detects the signature sequence and removes the signature sequence in the frequency domain or in the time domain. Figure 12 provides a first example in which the signature sequence is removed in the frequency domain. Referring to the example receiver shown in Fignre I la, the received base band signal is fed from a receive channel 428 to a matched filter 502 and a demodulator 550.

The match filter 502 receives the signature sequence in the time domain after a signature sequence generator 504, which is the same as the signature sequence generator 212 at the transmitter, re-generates a copy of the signature sequence. The matched filter 502 is configured to have an impulse response which is matched to the time domain signature sequence. As such, it correlates the time domain signature sequence with the received signal fed from the receive channel 428 and the correlation output result can be used to detect the presence of the preamble OFDM symbol when an output of the correlation process exceeds a predetermined threshold. Furthermore, as a result of the presence of the signature sequence in the preamble OFDM symbol, an impulse response of the channel through which the received signal has passed can also be estimated from the correlation output of the matched filter by a channel impulse response estimator 508. [he receiver can therefore include an arrangement for estimating the channel impulse response using the signature sequence without recourse to the traditional scattered pilots.

Having detected the presence of the signature sequence and estimated the channel impulse response, the effect of the channel impulse response can be removed from the received signal within the demodulator 550. Accordingly a Fast Fourier Transformer 518 transforms the channel impulse response estimate into the frequency domain channel transfer function and feeds the channel transfer function to an equaliser 516 within the demodulator 550.

In the receiver shown in Figure 12 the demodulator 550 is arranged to recover the signalling data in a base band form encoded with an error correction code. The demodulator 550 therefore recovers the signalling data from the signalling (preamble) OFDM symbol, which is then decoded using a forward error correction decoder 520 before being descrambled by a descrambling unit 522 which corresponds to the scrambling unit 200 shown in Figure 6 biLt performs a reverse of the scrambling.

The demodulator 550 includes a guard interval remover 512, which removes the guard interval from the signalling OFDM symbols, and an FFT unit 514, which converts the time domain samples into the frequency domain. The equaliser 516 removes the effects of the channel impulse response, which has been converted into the frequency domain to form a channel transfer function by the FFT unit 518 as already explained above. In the frequency domain the equaliser 516 divides each signalling data carrying OFDM sub-carrier by its corresponding channel transfer coefficient to remove, as far as possible, the effect of the transmission channel from the modulation symbols.

A signature sequence remover is formed by an adder unit 519 wlueh receives the signature sequence in the frequency domain generated by the signature sequence generator 504 after this has been scaled by the sealing factor Ci, as explained above by a scaling unit 521.

Thus the signature sequence remover 519 receives at a first input the equaliscd preamble OFDM symbol and on a second input a sealed signature sequence in the frequency domain and subtracts one from the other to fonn at the output estimates of the modulation symbols which were carried by the data bearing subcarriers of the preamble OFDM symbol.

The modulation symbols representing the error correction encoded preamble signalling data are then demodulated and error correction decoded by the demodulator and FEC decoder 520 to form a.t an output the scrambled bits of the Li signalling data which arc then descrambled by the descrambling unit 522 to form as an output 524 the Li signalling data bits.

A further example of the preamble detector and decoder 430 which operates in the time domain to remove the signature sequence is showing in Figures 13 and 14. Figure 13 provides an example of the preamble detector and decoder 430 wlueh corresponds to the example shown in Figure 12 and so only differences with respect to the operation of the example shown in Fignre 13 will be explained. In Figure 13 as with the example in Figure 12 the baseband received signal is fed to a signature sequence matched filter 502 and to a demodulator 550. As with the example shown in Figure 12, the signature sequence matched filter cross-correlates the received signal with an impulse response which is matched to the time domain signature sequence. The signature sequence is received in the time domain form by regenerating the signature sequence in the frequency domain using the signature sequence generator 504 and transforming the signature sequence into the time domain using an inverse Fourier transform processor 506. As with the example shown in Figure 12 a channel impulse response estimator 508 detects the channel impulse response from the output of the signature sequence matched filter 502 and forms this into the frequency domain channel transfer function using an FFT unit 518 to feed the frequency domain channel estimate to an equalli ser 516 within the demodulator 550.

So far the operation of the example shown in Figure 13 corresponds to that shown in Figure 12. As shown in Figure 13 the demodulator 550 includes the signature sequence remover 559 at before the guard remover 512. The time domain signature sequence which is fed from the inverse Fourier transform unit 560 is scaled by the sealing unit 521 by the predetermined factor Ci. The scaled time domain signature sequence is then fed to the signature sequence remover 559 which removes the signature sequence in the time domain from the received baseband signal. Thereafter the guard remover 512, the FFT unit 514 and the equaliser 516 operate in a corresponding way to the elements shown in Figure 12.

The signature sequence remover 559 shown in Figure 13 is shown in more detail in Figure 14. In Figure 14 the signature sequence remover 559 comprises a guard interval inserter 561, a combiner unit 560 and an FIR filter 562. The time domain haseband received signal is received on the input channel 428 at one input of the combiner unit 560. A second input 564 receives the scaled time domain version of the signature sequence, which is fed to the guard interval inserter 561 which prepends a cyclic prefix to the signature sequence in much the same way as the guard interval inserter 561 42 at the transmitter. The output of the guard interval inserter feeds the FIR filter 562 which receives on a second input 566 the estimate of the channel impulse response generated by the channel impulse response extraction block 508.

The FIR filter 562 therefore convolves the channel impulse response estimate with the signature sequence in the time domain which is then subtracted by the combiner 560 from the rcccivcd haseband signal to remove the effect of the signature sequence from the received signal. Figure 15b shows a more detailed example implementation of this signature sequence removal and how the FIR filter 562 is configured.

As will he appreciatedthe operation of the demodulator and FEC decoder 520 and the scrambler 522 perform the same functions as explain with reference to Figurc 12.

1 0 Matched Filter As indicated above the matched filter 502 generates an output signal which represents a correlalion of the received signal with the signature sequence. A block diagram showing an example of the signature sequence matched filter 502 is shown in Figure 1 Sa.

Figure isa shows a sequence of Ns delay elements 600 connected to scaling units 602 which scale each of the samples of the data stored in the delay storing unit 600 by a corresponding component of the signature sequence P(i) but conjugated. The output from each of the scaling units 602 is then fed to an adding unit 604 which forms an output signal representing a correlation of the received signal samples r(n) with the signature sequence at an output 606, The matched filter implements the equation: y(i) = ElP* (m)rØt + 1) for i = -Ns-fl, -Ns+2. .., 0,1.2,.. Ns-l When the filter taps P0) are of form (+l±jl), the multiplier at each tap could simply be done by add and subtract circuits for each of the in-phase and qLtadrature components.

When the signature sequence is a CAZAC sequence, the quadrature components of P0) are not bipolar. TI'he scaling units 602 can use the sign of each quadrature component instead so as to have the form (±l±jl).

Figure 1 6a and Figure 16b provide cxamnples of a. correlation output of the match filter for a multipath environment. In this case the channel is composed of three paths and the preamble is a 4K symbol with GIF of 1⁄4 for illustrative purposes only. As can be seen there is a clear correlation peak when the signature sequence of the received signal coincides with the signature sequence at the receiver. The example shown in Figure 1 Gb shows the output of the match filter but with a more expanded x-a.xis showing an increase in resolution which is expanded from the correlation peak shown in Figure 16a. For this channel, there are three paths as tabulated in the Table below: 1 0 [0] 0 2 10 [68] -10 3 25 [171] -6 Table 3: Multipath profile of example channel Channel Impulse Response Extractor As can be seen from Figure 1 6h, both the amplitudes of the main impulses and their relative delays coincide with the characteristics of the muhipath channel profile through which this particular signal propagated. To detect the actual channel paths, a threshold of energy detection is set to an appropriate multiple of the root mean square (RMS) level of the matched filter output within a window ±Nc of the highest amplitude output sample. The exact multiple of the RMS is chosen experimentally depending on the lowest signal to noise ratio under which the system is to work. Any sample of the matched filter output above this threshold is taken as a channel path, and all other samples are then set to zero in the channel impulse estimator 508. Finally, the channel impulse response (CTR) is normalised by dividing all its samples with the highest amplitude sample. In this way, the relative amplitudes and delays of each of the impulses in the channel through which the received signal has passed can be estimated.

Signature Sequence Remover Having formed an estimate of the channel impulse response, a component of the received signal corresponding to that contributed by the signature sequence in the received signal can be generated by passing the received signal r(i) through the signature sequence remover 559, which is configured with filter taps h11 to reflect the delay and amplitude profile of the channel impulse response. This can he accomplished by suitable scaling, shifting and adding of the signature sequence of length Ns Nu + Ng of the preamble symbol. An example of the filter is shown in Figure ish.

As shown in Figure lSh, the signature sequcnce remover 559 includes a finite impulse response (Fm) filter 562 made up of a delay line comprised of Ns-l delay elements 652.1, 652.2, to 652.Ns-l. The output of these delay elements are connceted to corresponding gain terms 651.1, 651.2, to 651.Ns-1 each of which gain stages feed their output to the adder 653.

The input 654 of the filler is connected both to the input of delay element 652.1 and to the input of gain term 651.0. The output 656 of the FIR filter 650 is connected to the input of an adder 560 whose other input 657 receives the received preamble signal samples r(i). During operation, the gain stages of the FIB. filter are set to the negative values of the samples of the channel inipulse response derived by the channel impulse response estimator 506. The FIR 650 generates at an output 656 a signal representing the convolution of the signature sequence by the channel innpulse response estimate, which effectively provides an estimate of the effect of the channel on the signature sequence imposed upon the signalling OFDM symbol. An adder 560 then subtracts the output signal of the FIR 656 from the received signal from an input 657 to remove the effect of the signature sequence from the received signal to fonn an output 660. Therefore a result (of the signature sequence transiting the channel described by the channel impulse response) is subtracted from the received signal by the signature sequence remover 510 with a delay matched to the point from which the first significant impulse (of the output of the matched filter) occurred. This process can be iterated in that the matched filter 502 can be re-run with the results of the subtraction, the channel impulse response re-estimated by the channel impulse response estimator 508 and the its effect on the signature sequence being extracted again by the signature sequence remover 559. As a result, a more accurate estimate of the effect of the signature sequence on the received signal can be estimated and subtracted from the received signal. Channel impulse responses from all iterations can then he summed and normalised to provide an improved estimate of the channel impulse response from which the channel transfer function (CTF) is derived for preamble symbol equalisation.

Frequency Offset Estimation Figure 17 provides a more detailed schematic block diagram of the preamble pilot matched filter 444 used for detecting a coarse frequency offset in the received signalling OFDM symbol, which may fonn part of the frequency synchroniscr 424 of Figure 11 a. As explained above, the number of pilots introduced into the signalling OFDM symbol is less than the number which would be required in order to estimate the channel. The number of pilot symbols is therefore set to estimate a coarse frequency offset. The block diaam shown in Figure 17 provides an example representation of the coarse frequency remover 513 and is shown with three versions of the received preamble signal 701.

As shown in Figure 17 a sequence of delay elements 700 are used to feed in discrete samples of the signal which are then multiplied by multipliers 702 with the known pilot signal values P('n) and summed by a summing unit 704 to form a correladon output 706. A pulse detector or peak detector 708 is the same one shown as 446 in Figure 1 lb which then generates an output signal on channel 710 showing a peak when there is a coincidence between a relative offset of the received signal with the company of the pilot signals at the receiver. Shaded circles of each received signal 701 show sub-carrier cells that represent preamble pilots whilst the un-shaded cells show non-pilot sub-carrier cells. All sub-carrier cells are shifted into the transversal filter from right to left. The parameter MaxOff is a design parameter that represents the maximum value of the frequency offset in units of sub-carrier spacing 12 that the designer may expect. The output of the pulse detector is only valid between shifts (0.5(Na+Nu)-MaxOfl) and (0.5(Na+Nu) + MaxOffj where Na is the number of sub-carriers (out of a total of Nu) used in the preamble OFDM symbol. If the shifts are numbered from -MaxOff to + MaxOff then the pulse detector output will go high for the shift that corresponds to the observed frequency offset.

Once £2 is detected, this coarse frequency is removed by shifting the subcarriers by -4? i.e. in the opposite direction to the frequency offset. This can also be removed prior to FF1' in common with the fine frequency offset which is estimated from the argument of the peak preamble detection matched filter or guard interval correlation 432 peak sample by modulation with a suitably phased sinusoid generated by the oscillator 426 in Figure 1 la.. The two frequency offsets can be used to start off the carrier correction loop for the rest of the OFDM symbols in the frame.

Figure 18 shows a pilot correlation result of a frequency offset in an example plot of the input of the pulse detector for a frequency offset of £2 -88 in a case where MaxOff is set to 350. The pulse detector might use a threshold to clip this input as a detector of the presence or absence of a substantial pulse.

Preamble Symbol Egualisation After signature sequence removal from the received samples and the coarse frequency offset has been adjusted, OFDM equalisation can begin with the FFT of the received sequence.

The FFT window starts from a (rigger position in the FFT unit 514 corresponding to the relative delay of the first impulse in the channel impulse response estimate, If the channel impulse response estimate duration is longer than the preamble (lID, then the trigger position is altered to ensure that it starts at (he beginning of a Ng (Ng is the number of time domain samples in the guard interval of the preamble symbol) long window under which the maximum of the energy of the channel impulse response estimate falls. The Ni, point FFT produces the preamble OFDM symbol in the frequency domain with the effect of the channel superposed. Before equalisation and decoding, any frequency offsets have to he calculated and removed by the frequency offset remover as explained above with reference to Figures 1 Ia, 1 ib, 1 Ic. This estimation uses correlation with the known preamble pilots to determine how far to the right or left the full symbol is shifted in frequency. Equalisation of the preamble OFDM symbol requires a channel transfer function (CTF). This is derived by executing a Na point FF1 on the channel impulse response estimate by the ITT unit 518. This provides a channel transfer function for all suhearriers in the preamble OFDM symbol allowing subearrier by suhcarrier one-tap equalisation to take place. finally, the equalised data subcarriers are extracted (pilot subcarriers discarded) and de-mapped, forward error correction (FEC) decoded to provide the signalling.

Selected Results Figure 19 provides a graphical plot of bit error rate with respect to signal to noise ratio for different code rates with and without the addition of the signature sequence to the signalling OFDM symbol. Thus, two code rates are shown, rate one half and rate one quarter, each code rate including the example of the presence of the signature sequence and without the signature sequence. As can be seen, the results for rate one quarter show that the signalling OFDM symbol can he detected even at signal to noise ratios of less than -2dBs.

Two further sets of results shown in Figures 20a and 20b provide a graphical plot of bit error rate against signal to noise ratio in which for the results shown in Figure 20a there is a 0dB echo channel with an ideal channel estimation and in Figure 20b a multipath environment with two paths as illustrated in Figure 20e. Thus for Figure 20b in contrast to the result shown in Figure 20a there is a relative degradation in performance resulting from real channel estimation. however, as can be seen, the results are comparable.

Various further aspects and features of the present disclosure are defined in the following numbered clauses: 1. A transmifter for lransmitting payload data using Orthogonal frequency Division Multiplexed (OFDM) symbols, the transmitter comprising a frame builder configured to receive the payload data to be transmitted and to receive signalling data for use in detecting and recovering the payload data at a receiver, and to fonn the payload data with the signalling data into frames for transmission, a modulator configured to modulate a first OFDM symbol with the signalling data forming a part of each of the frames and to niodulate one or more second ()FDM symbols with the payloa.d data to forni one or more of the frames, and a transmission unit for transmitting the first and second OFDM symbols, wherein the first OFDM symbol is a first type having a number of sub-carriers which is less than or equal to the number of sub-carriers of the one or more second OFDM symbols of a second type and a guard interval for the first OFDM symbol is selected in dependence upon the longest possible guard interval of the second OFDM symbol.

2. A transmitter according to clause 1, wherein the number of sub-carriers of the first OFDM symbol is selected in dependence upon the duration of the selected guard interval.

3. A transmitter according to clause 1 or 2, wherein the number of sub-carriers of the first Of DM symbol is selected to increase a likelihood of a receiver being able to detect and recover the signalling data from the first OFDM symbol and the number of sub-carriers of the second OFDM symbol is selected to maximise spectral efficiency.

4. A transmitter according to any of clauses 1, 2 or 3, wherein the number of sub-carricrs in the second OFDM symbol is substantially thirty two thousand, sixteen thousand or eight thousand and the number of sub-carriers of the fir st OFDM symbol is substantially eight thousand.

5. A transmitter according to clause 4, wherein the guard interval is 19/32 or 19/64 of the duration of the first OWM symbol.

(3. A transmitter as claimed in any preceding Claim, wherein the transmitter includes a pilot signal inserter configured to insert pilot symbols on selected sub-carriers of the fir st OFDM symbol, the number of pilot symbol carrying sub-carriers being less than a number which would bc required to estimate a channel impulse response through which the first OFDM symbol is transmitted, and sufficient to estimate a coarse frequency offset of the transmitted OFDM symbol.

7. A transmitter according to any of clauses 1 to 6, wherein the signalling data is encoded with a first error correction code and the payload data is encoded with at least one other error correction code, an encoding rate of the first error correction code being lower than an encoding rate of the at least one other error correction code.

8. A transmitter according to clause 7, wherein the signalling data is encoded with an error correction code, an encoding rate of the error correction code being lower than rate one quarter.

9. A method of transmitting payload data using Orthogonal Frequency Division Multiplexed (OFDM) symbols, the method comprising receiving the payload data to he fran sin itted, receiving signalling data for use in detecting and recovering the payload data to be transmitted at a receiver, forming the payload data with the signalling data into frames for transmission, modulating a first OFIJM symbol with the signalling data forming a part of each of the frames and modulating one or more second OFDM symbols with the payload data to form one or more of the frames, and transmitting the first and second OFDM symbols, wherein the first OFDM symbol is of a first type having a number of sub-carriers which is less than or equal to the number of sub-carriers of the one or more second OFDM symbols of a second type, and the transmitting includes selecting a guard interval for the fir st OFDM symbol depending upon the maximum allowable duration of the guard interval useable for the one or more second OFDM symbols.

10. A method according to clause 9, wherein the selecting the guard interval for the first OFDM symbol includes selecting the number of sub-carriers of the first OFDM symbol in dependence upon the size of the selected guard interval.

11. A method according to clause 9 or 10, wherein the selecting the guard interval for the first OFDM symbol includes selecting the number of sub-carriers of the first OFDM symbol to increase a likelihood of a receiver being able to detect and recover the signalling data from the first OFDM symbol and the method comprises selecting the number of sub-carriers of the second OFDM symbol to maxiinise spectral efficiency.

12. A method according to clause 11, wherein the selecting the number of sub-carriers of the second OFDM symbol includes selecting the number of sub-carriers in the second OFDM symbol to be substantially thirty two thousand, sixteen thousand or eight thousand and the selecting the number of sub-carriers of the first OFDM symbol comprises selecting the number of sub-carriers of the first OFDM symbol lo be substantially eight thousand.

13. A method according to clause 12, wherein the selecting the guard interval comprises selecting the guard interval to be 19/32 or 19/64 of the duration of the first OFDM symbol.

14. A method according to any of clauses 9 to 13, comprising inserting pilot symbols on selected sub-carriers of the first OFDM symbol, the number of pilot symbol carrying sub-carriers being less than a number which would be required to estimate a channel impulse response through which the fir st OFDM symbol is transmitted, and sufficient to estimate a coarse frequency offset of the fransmitted OFDM symbol.

15. A method according to any of clauses 9 to 14, comprising encoding the signalling data with a first error correction code and encoding the payload data with at least one other error correefion code, an encoding rate of the first error correction code being lower than an encoding rate of the at least one other error correction code.

16-A method according to clause 15, comprising encoding the signalling data with an error correction code, an encoding rate of the error correction code being lower than rate one quarter.

1 7. A receiver for detecting and recovering payload data from a received signal, the receiver comprising a detector for detecting the received signal, the received signal comprising the payload data and signalling data for use in detecting and recovering the payload data, the signalling data and the payload data forming frames in the received signal, the signalling data in each frame being carried by a firsi Orthogonal Frequency Division Multiplexed, OFDM, symbol, and the payload data being carried by one or more second OFDM symbols, and the first OFDM symbol has a number of sub-carriers which is less than or equal to the mimber of sub-carriers of the second OFDM symbol and a guard interval for the first OFDM symbol is selected in dependence upon the maximum allowable duration of the guard interval useable for the second OFDM symbol, a demodulator configured to detect the first OFDM symbol and the second OFDM symbol and to recover the signalling data from the first OFDM symbol in the presence of the guard interval using a forward Fourier transform with respect to the number of sub-carriers of the first OFDM symbol and using the signalling data to recover the payload data from the second OFDM symbol.

18. A receiver according to clause 17, wherein the number of sub-carriers of the first OFDM symbol is selected in dependence upon the size of the sciected guard interval.

19. A receiver according to clause 17 or 18, wherein the number of sub-carriers of the first Of DM symbol is selected to increase a likelihood of a receiver being able to detect and recover the payload data from the first OFDM symbol and the number of sub-carriers of the second OFDM symbol is selected to maximise spectral efficiency.

20. A receiver according to any of clauses 17, 1. 19, wherein the number of sub-carriers in the second OFDM symbol is substantially thirty two thousand, sixteen thousand or eight thousand and the number of sub-carriers of the first OFDM symbol is substantially eight thousand.

21. A receiver according to clause 20, wherein the guard interval is 19/32 or 19/64 of thc duration of the first OFDM symbol.

22. A receiver according to any of clauses 17 to 21, wherein the received signal includes pilot symbols inserted on selected sub-carriers of the first OFDM symbol, the number of pilot symbol carrying sub-carriers being less than a number which would be required to estimate a channel impulse response through which the first OFDM symbol has been transmitted, wherein the receiver includes a coarse frequency offset estimator configurcd to estimate a coarse frequency offset of the received first OFDM symbol, and to compensate for the coarse frequency offsct in the received signal.

23. A receiver according to any of clauses 17 to 22, wherein the signalling data has been encoded with a first error correction code and the payload data is encoded with at least one other error correction code, an encoding rate of the first error correction code being lower than an encoding rate of the at least one other error correction code, and the receiver 1 5 comprises an error correction decoder configured to decode the first error correction encoded signalling data to generate an estimate of the signalling data.

24. A receiver according to clause 23, wherein the signalling data. has been encoded with an error correction code, an encoding rate of the error correction code being lower than rate one quarter, and the receiver comprises an error correction decoder configured to decode the error correction encoded signalling data to generate an cstirnatc of the signalling data.

25. A method of detecting and recovering payload data from a received signal, the method comprising detecting the received signal, the received signal comprising the payload data and signalling data. for use in detecting and recovering the payload data, the signalling data and the payload data forming frames in the received signal, the signalling data in each frame being carried by a first Orthogonal Frequency Division Multiplexed, O12DM. symbol, and the payload data being carried by one or more second OFDM symbols and a guard interval for the first OFDM symbol is selected in dependence upon the maximum allowable duration of the guard interval useable for the second OFDM symbol, demodulating the first OFDM symbol and the second OFDM symbol to recover the signalling data from the first OFDM symbol in the presence of the guard interval using a forward Fourier transfonu with respect to the number of sub-carriers of the first OFDM symbol, and using the signalling data to recover the payload data from the second OFDM symbol.

26. A method according to clause 25, wherein the number of sub-carriers of the first OFDM. symbol is selected in dependence upon the size of the selected guard interval.

27. A method according to clause 25 or 26, wherein the number of sub-carriers of the first OFDM symbol is selected to increase a likelihood of a receiver being able to detect and recover data from the first OFDM symbol and the number of sub-carriers of the second OFDM symbol is selected to maxhnise spectral efficiency.

28. A method according to any of clauses 26, 27 or 28, wherein the number of sub-carriers in the second OFDM symbol is substantially thirty two thousand, sixteen thousand or eight thousand, and the number of sub-carriers of the first OFDM symbol is substantially eight thousand.

29. A method according to clause 28, wherein the guard interval is 19/32 or 19/64 of the duration of the first OFDM symbol.

30. A method according to any of clauses 25 to 29, wherein the received signal includes pilot symbols inserted on selected sub-carriers of the first OFDM symbol, the number of pilot symbol carrying sub-carriers being less than a number which would be required to estimate a channel impulse response through which the first OFDM symbol has been transmitted, and the method includes estimating a coarse frequency offset of the received first OFDM symbol, and compensating for the coarse frequency offset in the received signal.

31. A method according I.o any of clauses 25 to 30, wherein the signalling data has been encoded with a first error correction code and the payload data is encoded with at least one other error correction code, an encoding rate of the first error correction code being lower than an encoding rate of thea! least one other error correction code, and the method comprises decoding the first error correction encoded signalling data to generate an estimate of the signalling data.

32. A method according to clause 31, wherein the signalling data has been encoded with an en'or correction code, an encoding rate of the error correction code lower than rate one quarter, and the method comprises decoding the error con'ection encoded signalling data to generate an estimate of the signalling data.

Various further aspects and features of the present disclosure are defined hi the appended claims. Various combinations of features may be made of the features and method steps defined in the dependent claims other than the specific combinations set out in the attached claim dependency. Thus the claim dependencies should not be taken as limiting.

Claims

CLAIMS1. A transmitter for transmitting payload data using Orthogonal Frequency Division Multiplexed (OFDM) symbols, the transmitter comprising a frame builder conflglLred to receive the payload data to he transmitted and to receive signalling data for use hi detecting and recovering die payload data at a receiver, and to form the payload data with the signalling data into frames for transmission, a modulator configured to modulate a first OFDM symbol with the signalling data forming a part of each of the frames and to modulate one or more second OFDM symbols with the payload data to form one or more of the frames, and a transmission unit for transmitting the first and second OFDM symbols, wherein the first OFDM symbol is a first type having a number of sub-carriers which is less than or equal to the number of sub-carriers of the one or more second OFDM symbols of a second type and a guard interval for the fll'st OFDM symbol is selected in dependence upon the longest possible guard interval of the second OFDM symbol.
2. A transmitter as claimed in Claim 1, wherein the number of sub-carriers of the first OFDM symbol is selected in dependence upon the duration of the selected guard interval.
3. A transmitter as claimed in Claim 1, wherein the number of sub-carriers of the first OFDM symbol is selected to increase a likelihood of a receiver being able to detect and recover the signalling data from the first OFDM symbol and the number of sub-carriers of the second OFDM symbol is selected to maximise spectral efficiency.
4. A transmitter as claimed in Claim 1, wherein the number of sub-carriers in the second OFDM symbol is substantially thirty two thousand, sixteen thousand or eight thousand and the number of sub-carriers of the first OFDM symbol is substantially eight thousand.
5. A transmitter as claimed in Claim 4, wherein the guard interval is 19/32 or 19/64 of the duration of the first OFD.M symbol.
6. A transmitter as claimed in Claim 1, wherein the transmitter includes a pilot signal inserter configured to insert pilot symbols on selected sub-carriers of the first OFDM symbol, the number of pilot symbol carrying sub-carriers being less than a number which would he required to estimate a channel impulse response through which the first OFDM symbol is transmitted, and sufficient to estimate a coarse frequency offset of the transmitted OFDM symbol.
7. A transmitter as claimed in Claim 1, wherein the signalling data is encoded with a first error correction code and the payload data is encoded with at least one other error correction code, an encoding rate of the first error correction code being lower than an encoding rate of the at least one other error correction code.
8. A transmitter as claimed in Claim 7, wherein the signalling data is encoded with an error correction code, an encoding rate of the error correction code being lower than rate one quarter.
9. A method of transmitting payload data using Orthogonal Frequency Division Multiplexed (OVDM) symbols, the method comprising receiving the payload data to be transmitted, receiving signalling data for use in detecting and recovering the payload data to be transmitted at a receiver, forming the payload data with the signalling data into frames for transmission, modulating a first OFDM symbol with the signalling data forming a part of each of the frames and modulating one or more second OFDM symbols with the payload data to form one or more of the frames, and transmitting the first and second OFDM symbols, wherein the first OFDM symbol is of a first type having a. number of sub-carriers which is less than or equal to the number of sub-carriers of the one or more second OFDM symbols of a second type, and the transmitting includes selecting a guard interval for the first OF'DM symbol depending upon the maxinnun allowable duration of the guard interval useahie for the one or more second OVUM symbols.
10. A method as claimed in Claim 9, wherein the selecting the guard interval for the first OFDM symbol includes selecting the number of sub-carriers of the first OFDM symbol in dependence upon the size of the selected guard interval.
11. A method as claimed in Claim 9, wherein the selecting the guard inten'al for the first OFDM symbol includes selecting the number of sub-carriers of the first OFDM symbol to increase a likelihood of a receiver being able to detect and recover the sigualling data from the first OF.DM symbol and the method comprises selecting the number of sub-can-iers of the second OFDM symbol to maximise spectral efficiency.
12. A method as claimed in Claim 11, wherein the selecting the number of sub-carriers of the second OFDM symbol includes selecting the number of sub-carriers in the second OFDM symbol to he substantially thirty two thousand, sixteen thousand or eight thousand and the selecting the number of sub-caiTiers of the first OFDM symbol comprises selecting the number of sub-carriers of the first OFDM symbol to he substantially eight thousand.
13. A method as claimed in Claim 12, wherein the selecting the guard interval comprises selecting the guard interval to be 19/32 or 19/64 of the duration of the first OFDM symbol.
14. A method as damned in Claim 9, comprising inserting pilot symbols on selected sub-carriers of the first OFDM symbol, the number of pilot symbol carrying sub-carriers being less than a number which would be required to estimate a channel impulse response through which the first OFDM symbol is transmitted, and sufficient to estimate a coarse frequency offset of the transmitted OFDM symbol.
15. A method as claimed in Claim 9, comprising encoding the signalling data with a first error correction code and encoding the payload data with at least one other error correction code, an encoding rate of the first error correction code being lower than an encoding rate of the at least one other en-or correction code.
16. A method as claimed in Claim 15, comprising encoding the signalling data with an error correction code, an encoding rate of the error correction code being lower than rate one quarter.
17. A receiver for detecting and recovering payload data from a received signal, 1 5 the receiver comprising a detector for detecting the received signal, the received signal comprising the payload data and signalling data for use in detecting and recovering the payload data, the signalling data and the payload data forming frames in the received signal, the signalling data in each frame being carried by a first Orthogona.l Frequency Division Multiplexed, OFDM, symbol, and the payload data being carried by one or more second OFDM symbols, and the first OFDM symbol has a number of sub-carriers which is less than or equal to the number of sub-carriers of the second OITDM symbol and a guard interval for the first OFDM symbol is selected in dependence upon the maximum allowable duration of the guard interval useable for the second OFDM symbol, a demodulator configured to detect the first OFDM symbol and the second OFDM symbol and to recover the signalling data from the first OFDM symbol in the presence of the guard interval using a forward Fourier transform with respect to the number of sub-carriers of the first OFDM symbol and using the signalling data to recover the payload data from the second OFDM symbol.
18. A receiver as claimed in Claim 17, wherein the number of sub-carriers of the first OFDM symbol is selected in dependence upon the size of the selected guard interval.
19. A receiver as claimed in Claim 17, wherein the number of sub-carriers of the first OFDM symbol is selected to increase a likelihood of a receiver being able to detect and recover the payload data from the first OFDM symbol and the number of sub-calTiers of the second OFDM symbol is selected to maximise spectral efficiency.
20. A receiver as claimed in Claim 17, wherein the number of sub-carriers in the second OFDM symbol is substantially thirty two thousand, sixteen thousand or eight thousand and the number of sub-carriers of the first OFDM symbol is substantially eight thousand.
21. A receiver as claimed in Claim 20, wherein the guard interval is 19/32 or 19/64 of the duration of the first OFDM symbol.
22. A receiver as claimed iii Claim 17, wherein the received signal includes pilot symbols inserted on selected sub-carriers of the first OFDM symbol. the number of pilot symbol carrying sob-carriers being less than a number which would be required to estimate a channel impulse response through which the fir st OFDM symbol has been transmitted, wherein the receiver includes a coarse frequency offset estimator configured to estimate a coarse frequency offset of the received first OFDM symbol, and to compensate for the coarse frequency offset in the received signal.
23. A receiver as claimed in Claim 17, wherein the signalling data has been encoded with a first error correction code and the payload data is encoded with at least one other error correction code, an encoding rate of the first error correction code being lower than an encoding rate of the at least one other error correction code, and the receiver comprises an error correction decoder configured to decode the first elTor correction encoded signalling data to generate an estimate of the signalling data.
24. A receiver as claimed in Claim 23, wherein the signalling data has been encoded with an error correction code, an encoding rate of the error correction code being lower than rate one quarter, and the receiver comprises an error correction decoder configured to decode the error correction encoded signalling data to generate an estimate of the signalling data.
25. A method of detecting and recovering payload data from a received signal, the method comprising detecting the received signal, the received signal comprising the payload data and signalling data for use in detecting and recovering the payload data, the signalling data and the payload data funning frames in the received signal, the signalling data in each frame being carried by a first Orthogonal Frequency Division Multiplexed, OFDM, symbol, and the payload data being carried by one or more second OFDM symbols and a guard interval for the first OFDM symbol is selected in dependence upon the maximum allowable duration of the guard interval useable for the second OFDM symbol, demodulating the fir st OFDM symbol and the second OFDM symbol to recover the signalling data from the fir st OFDM symbol in the presence of the guard interval using a forward Fourier transform with respect to the nmnber of sub-carriers of the first OFDM symbol, and using the signalling data to recover the payload data from the second OFDM symbol.
26. A method as claimed in Claim 25, wherein the number of sub-carriers of the first OFDM symbol is selected in dependence upon the size of the selected guard interval.
27. A method as claimed in Claim 25, wherein the number of sub-carriers of the first OFDM symbol is selected to increase a likelihood of a receiver being able to detect and recover data from the fir st OFDM symbol and the number of sub-carriers of the second OFDM symbol is selected to maximise spectral efficiency.
28, A method as claimed in Claim 26, wherein the number of sub-carriers in the second OFDM symbol is substantially thirty two thousand, sixteen thousand or eight thousand, and the number of sub-carriers of the first OFDM symbol is substantially eight thousand.
29. A method as claimed in Claim 28, wherein the guard interva.l is 19/32 or 19/64 of the duration of the first OFDM symbol.
30. A method as claimed in Claim 25, wherein the received signal includes pilot symbols inserted on selected sub-carriers of the first OFDM symbol, the number of pilot symbol carrying sub-carriers being less than a number which would be required to estimate a channel impulse response through which the first 02DM symbol has been transmitted, and the method includes estimating a coarse frequency offset of the received first OFDM symbol, and compensating for the coarse frequency offset in the received signal.
31. A method as claimed in Claim 25, wherein the signalling data has been encoded with a first error correction code and the payload data is encoded with at least one other error correction code, an encoding rate of the first error correction code being lower than an encoding rate of thc at least one other error correction code, and the method comprises decoding the first error correction encoded signalling data to generate an estimate of the signalling data.
32. A method as claimed in Claim 31, wherein the signalling data has been encoded with an error correction code, an encoding rate of the error correction code lower than rate one quarter, and the method comprises decoding the error correction encoded signalling data to generate an estimate of the signalling data.
33. A computer program providing computer executable instructions which when loaded onto a computer causes the computer to perform the method according to Claim 9 or 25.
34. A transmitter substantially as hereinbefore described with reference to the accompanying drawings.
35. A method substantially as hercinbcforc described with reference to the accompanying drawings.