GB2354411A - Transparent-in-tone-band (TTIB) transmitter and receiver - Google Patents

Abstract

A transparent-tone-in-band (TTIB) transmitter comprises separate modulators 33a, 33b for modulating a pair digital data streams onto respective frequency sub-bands. The modulated signals are combined with a pilot tone to produce a resulting signal in TTIB format. The pair of digital data streams may be split from a single data stream or may be from separate sources. A transparent-tone-in-band receiver comprises means for recovering the pilot tone, a pair of demodulators for demodulating the signal in each sub-band into a pair of analogue signals, and means for converting each analogue signal into a respective digital data stream. The system may use m-ary QAM modulation and reference vector equalisation (RVE) of the combined signal.

Description

2354411 DIGITAL DATA COMMUNICATION The present invention relates to

apparatus for, and methods of transmitting and receiving digital data over radio networks.

Providers of wireless telephone services aim to match the services offered by wired (PSTN) networks. Recent growing demand of basic ISDN (64 kbit/s) services on PSTN presents challenges to wireless networks. This has lead researchers all over the world to investigate novel ways of squeezing more and more data into bandwidth-limited mobile radio channels. Among the various techniques considered for increasing throughput across a wireless channel, one of the most common solutions is to use higher level linear modulation schemes that can transmit many bits per symbol.

When compared with the modulation techniques, the QAM family offers the best compromise between bits per second per hertz and performance in signal to noise ration (SNR) (see J G Proakis, "Digital Communications," McGraw Hill, 3' Edition, 1995). However, the very nature of the wireless environment causes amplitude and phase distortion that must be accurately compensated to receive high levels of QAM. Since the 1980s, extensive research has resulted in two well-known methods for compensating this narrowband wireless channel distortion. These methods are based on transmitting a reference along with data. This reference is then used at the receiver to estimate the channel distortion. The first technique inserts pilot symbols within the information sequence (see Sampei S & Sunaga T, "Performance of multi-level QAM with maximal ratio combining space diversity for land mobile radio communications," IEEE Vehi,ular Technology Conference, 1990, pp.459-464), whereas the second technique transmits a pilot tone along with the data spectrum to assist the demodulation process (see Bateman A, "Feedforward transparent tone-in-band: Its implementations and applications," IEEE Transactions on Vehicular Technology, Aug 1990, Vol. 39, No.3, pp.235-243; and Martin P M, "Advanced Linear Modem Design for Narrowband Mobile Communications," PhD Thesis, Department of Electrical & Electronics Engineering, University of Bristol, UK, Feb 1991).

The general technique known as TTIB (transparent-tone-in-band) involves separating the analogue transmission information into two frequency subbands, frequency-translating them apart from one another and inserting a pilot tone at, or close to, the central point of the frequency gap between them. The resulting signal is frequency-translated to transmission frequency, power amplified and broadcast. At the reception side, the pilot tone is used as a reference in processing the received signal to carry out the complementary demodulation which, after frequency-down translation from RF, involves frequency-translating the sub-bands to close the frequency gap formed prior to transmission.

An enhancement of the basic TTEB system for use in transmitting digital data in place of a voice signal is so-called Reference Vector Equalisation (RVE) in which the pilot tone as a reference vector and at the reception side, amplitude and phase corrections necessary to restore the reference vector to a predetermined amplitude and phase are also applied to the two TTIB sub-bands prior to demodulation.

As shown in Fig 2 of the accompanying drawings, existing RVE systems take a single input digital data stream, pass it through a modulator 1 which is generally a QPSK (quadrature phase shift key) or m-ARY QAM (quadrature amplitude modulation) and then split the resulting analogue signal into the two frequency sub- bands for RVE processing by an RVE processor 2 in a similar manner as a standard voice transmission. The resulting signal is then frequency- translated to RF by a frequency-translation section 3, power amplified and broadcast to air.

At the receiver side, the received signal is frequency down-translated from RF at 5 and subject to RVE and FFSR (feed forward signal regeneration) at 6 (Fig.2). The resulting I and Q quadrature signals are phase-locked at 7 to derive a baseband analogue signal and an m-ARY QAM demodulator 8 is then used to recover the digital data stream. One drawback of this system is the need to phase-lock (at 7) the two subbands to ensure that the signal received by the demodulator 8 is correctly reconstituted, which takes both time and processor resources. The present invention seeks to provide an improved method for transmitting and receiving a digital data stream over TTEB in which the need for this phase-locking is avoided.

According to the present invention there is provided a transmitter for a signal in TTIB (transparent-tone-in-band) format, which is to say a format in which a signal is split into two frequency-subbands of equal widths with a frequency gap therebetween, into which gap a pilot tone is inserted, comprising:

means for modulating each one of a pair of inputted digital data streams into a signal occupying a respective one of said frequency-subbands; and means for combining the modulated digital signals with one another and with a pilot tone, to produce a resulting signal in TTIB format.

Since, in effect, two independent digital data streams pass through the two TTIB signal sub-bands, the need for phase-locking of the sub-bands is eliminated.

In this invention the data stream is split at the digital stage before being passed to two separate modulators which operate directly onto each analogue sub-band.

Alternatively, two completely separate data streams could be passed independently to each modulator before passing through the RVE process. At the receiver the signal is passed through the conventional RVE/FFSR process then each separate sub-band is passed to separate demodulator stages which can either produce two independent data streams, or can be digitally combined to produce a single data stream.

The use of RVE/FFSR techniques improves the quality of RF channel so that multi-level modulation schemes may be simply implemented in the real world where multipath, Doppler and noise pose serious problems for more conventional systems which require a significant quantity of the over air data to be reserved for synchronisation, timing recovery etc to be able to extract the data reliably and so reduce the overall dat rate available to the end use. Using RVE/FFSR to correct for the variations in the channel allows for virtually all the over air data to be available to the end user with very little signalling overhead (typically, this would be 50% of a conventional system, this new system should be around 5 %). Removing the need for phase locking improves the speed at which data becomes available at the output of the system and reduces the processor loading required to implement it.

The particular implementation that has been chosen by LMT consists of two 32 QAM sub band modems giving an overall over air data rate of 70kbps within the constraints of a 25kHz radio channel which leaves 64kbps available for the end user.

This scheme is capable of expansion to 140kpbs in a 50kHz channel, or reduction to 35kbps in 12.5kHz channels, or the use of higher (or lower) level modulation schemes to suit the prevailing radio channel conditions.

The incoming digital data streams may be from independent sources or derived from a single input data stream, e.g. by de-interleaving successive blocks of bits into the two streams; likewise, at the receiver side, the two streams may be recombined, e.g. by interleaving, to produce a single output stream.

The invention will be further described by way of non-limitative example with reference to the accompanying drawings, in which:

Figure I is a block diagram of a previously proposed system for transmitting digital data using TTEB.

Figure 2 is a block diagram of a previously proposed system for receiving data from the transmitter of Figure 1.

Figures 3 and 4 are block diagrams of, respectively, a transmitter and receiver embodying the invention.

Figure 3 shows in block form an embodiment of the transmitter 30 according to the present invention for use in fixed or mobile data applications.

The system is intended to provide a gross data rate of 70 Kbits/s in a 20kHz broadcast channel bandwidth. An incoming 70 Kbits/s digital data stream is split into two 36 Kbits/s streams by a serial/parallel converter 31 for processing through a pair of processing channels 32a, 32b. This splitting may be done by feeding successive bits, or groups of N successive bits into alternate ones of the channels 32a, 32b. Data interleaving to reduce the effects of fading and interference may be added at this stage. As noted above, the invention is applicable to two independent digital streams in which case the serial/parallel converter 31 may be omitted and these streams would then be fed directly to the inputs of the channels 32a and 32b, respectively.

The operation of the two channels will be described in more detail below but overall, is to process the two digital data streams into the TTEB/RVE format, which is to say that the two streams are each modulated and processed to occupy a respective one of the two sub-bands of the TTTB signal format.

In the channel 32a, the digital data stream inputted to it is mapped on to a 32 QAM constellation by a 32 QAM modulator 33 a to produce I and Q phase quadrature signals at a band rate of 7 Kbaud/s. These I and Q signals are subject to filtering by a bandpass filter 34a and then passed to a modulator 35a where they are combined to form the upper-frequency signal sub-band of the TTIB signal. The other digital data stream is similarly processed by the components of channel 32b with the difference that the modulator 33d is operative to modulate the signal into the lower-frequency TTIIB sub-band. The frequency bandwidth of the gap between the sub-bands depends upon the maximum frequency offset and Doppler shift to be corrected at the baseband in the system. In this particular system, a gap bandwidth of 70OHz has been used. This results in a relaxed specification for the transmit filters which have an excess bandwidth factor of 0.38. The IF signal at the transmitter can be written mathematically as:

ST(t) = A1(t)cos(wu1.,t+O1(t)) + A,(t)cos(w 3 COS (W,7t + ug + 02(t)) + A where A,(t), A,(t), A3are the amplitudes, (oLqF, %F, (a,Fthe frequencies and C(t), C(t), 4)2 the phases of the upper intermediate, lower intermediate sub-bands and pilot tone respectively.

It will be appreciated from the above that although the resulting signal is 'in TTEB format it is not prepared by the conventional method of splitting the baseband input signal into two sub-bands and shifting these apart in frequency to leave the frequency gap for the pilot tone; rather the operation of the channels 32a,33b is to produce sub-bands occupying the required frequency ranges to either side of the pilot tone frequency.

According to simulation testing the optional ratio of the transmission power of the pilot tone to that of the signal sub-bands is in the range of 10% to 30%.

The outputs of the two channels 32a,32b are added by adder 36 with one another and the output of a source 33 of the TTIB pilot tone. This summed output is then subject to frequency up-translation to the broadcast RF frequency, is power amplified by a linear or linearised power amplifier 38 and applied to an aerial for broadcast.

Figure 4 shows an embodiment of a receiver 50 for a signal received from the transmitter of Figure 3. After frequency down-translation from RF, the received signal SR(t) -is applied to a TTEB-pilot-tone extraction circuit 51 and then to a channel estimator and compensator circuit 52 whose function is to estimate the amplitude and phase errors of the pilot tone and, based on these estimates, to apply amplitude and phase compensation to the two TTEB sub-bands passing through it to reduce or eliminate corresponding amplitude and phase errors in these sub-bands.

The I and Q phase quadrature signals of the two TTEB sub-bands are fed to respective channels 53a, 53b for the upper- and lower-frequency sub-bands respectively.

Within these channels the I and Q signals pass through upper- and lowerfrequency IF carrier generators 54a,54b, and then to adders 55a,55b where they are added to the fed-forward signal SR(t) to carry out feed-forward signal regeneration. They then pass through receive band pass filters 56a,56b. From these filters, the signals are applied to 32 QAM demapping circuits 57a,57b for hard decision and 32 QAM dernapping. The outputs of these demapping circuits 57a,57b are the recovered digital streams, which may then optionally be recombined by a parallel/serial converter 58 to derive a single output 5 digital data stream.

Equation (1) above described the IF signal ST(t) at the transmitter. By the time the signal is processed by the receiver it has been corrupted by amplitude and phase distortion as well as Additive White Gaussian Noise (AWGN). The IF signal at the receiver can be written as:

SR M A I (t)A,4 (t) cos (C,)uFt + q5l (t) + ipcb (t)) + A2(t)Ach(t) COS((OL1Ft+952(t)+Ocb(t)) + A3 Ach (t) cos (c,),,,t + 03 + Ocb (t)) + n (t) (2) where Ach(t) and r h(t) are the amplitude and phase of the distortion introduced by the radio channel and n(t) represents the contribution due to AWGN.

In communication systems, when the received signal is down-converted from RF to IF, most often, the local oscillator drift causes frequency offset in the received signal. This makes the received signal impossible to detect. Equation 2 can be modified after incorporating the effect of frequency drift as shown below:

SR W AI(t)Ah(t) cos(o)ujFt+WDt+Oj(t)+Ob(t)) + AA)Adt) COS(O)UFt+(JDt+02(t)+Ocb(t)) + A3Ach(t) cos(wjFt+wDt+03+,Ocb(t)) + n(t) (3) where WD is the frequency drift introduced by the oscillators. For proper demodulation and detection of the received signal, the frequency drift of the oscillator and the amplitude and the phase distortion of the received signal must be taken into account.

These corrections are performed by filtering the tone from the received signal, as shown in Figure 2.

The tone extraction and channel estimation procedures described in the previous section depend heavily on the quality of the tone signal received in AWGN. A noisy tone produces noisy references and degrades the system performance considerably.

It is therefore important to optimise the tone to data power ratio. The tone to data power ratio is defined as:

R - 2 - A3' ( E[Al' (t)] + E [A22(t)] where E[j represents the mean value.

Large values of R give a clean tone but decrease the data band power and resulting in erroneous detection of the sub-band signals. On the other hand, too low a value of R results in a noisy tone and the sub-bands cannot be demodulated correctly.

R=0.2 offers the minimum error rate and hence this value has been used for evaluation of system performance in AWGN, Rayleigh and Rician channels.

The effect of in Rayleigh fading on different Doppler spreads and local oscillator frequency offsets has been simulated and it has been found that Doppler has no significant effect on the performance of the modem. This is because of the pilot insertion which makes any Doppler spread and frequency offset transparent to the modem performance.

Claims (18)

1. A transmitter for a signal in TTIB (transparent-tone-in-band) format, which is to say a format in which a signal is split into two frequencysubbands of equal widths with a frequency gap therebetween, into which gap a pilot tone is inserted, comprising: means for modulating each one of a pair of inputted digital data streams into a signal occupying a respective one of said frequency-subbands; and means for combining the modulated digital signals with one another and with a pilot tone, to produce a resulting signal in TTIB format.

2. A transmitter according to claim 1 in which there is a respective modulator associate with each of the digital data streams which produces an analogue baseband signal within the respective frequency subband.

3. A transmitter according to claim 2 wherein each modulator is an m-ary quadrature amplitude modulator (QAM).

4. A transmitter according to claim 3 wherein each modulator produces a 32 QAM constellation.

5. A transmitter according to any one of claims 1 to 4, wherein means are provided for RVE (reference vector equalisation) of the combined signal.

6. A transmitter according to any one of claims 1 to 5 and including means for frequency-translating the combined signal from baseband to radio transmission frequency.

7. A transmitter according to any one of claims 1 to 6 and including means for splitting an incoming digital data stream into two digital streams having half the bitrate thereof and for utilising the two digital data streams as said inputted data streams.

8. A receiver for a received signal in TTIB (transparent-tone-in-band) format, which is to say a format in which a signal is split into two frequency-subbands of equal widths with a frequency gap therebetween, into which gap a pilot tone is inserted, comprising: 5 means for recovering said pilot tone from the received signal and a pair of demodulators for demodulating the components of the signal in said sub- bands into a pair of analogue signals; and means operative to convert each of said analogue signals into a respective digital data stream.

9. A receiver according to claim 8, and including means for carrying out reference-vector-equalisation of the signal components in said subbands.

10. A receiver according to claim 9 or 10 wherein each demodulator is an m ary quadrature amplitude demodulator.

11. A receiver according to claim 10 wherein each demodulator produces a 32 QAM constellation.

12. A receiver according to any one of claims 8 to 11 and including means for combining said digital data streams into one.

13. A method of transmitting a signal in TTEB (transparent-tone-in-band) format, which is to say a format in which a signal is split into two frequency-subbands of equal widths with a frequency gap therebetween, into which gap a pilot tone is inserted, comprising:

modulating each one of a pair of inputted digital data streams into a signal occupying a 30 respective one of said frequency-subbands; and combining the modulated digital signals with one another and with a pilot tone, to produce a resulting signal in TTIB format.

14. A method of receiving a signal which has been transmitted in TTIIB (transparent-tone-in-band) format, which is to say a format in which a signal is split into two frequency-subbands of equal widths with a frequency gap therebetween, into which gap a pilot tone is inserted, comprising:

recovering said pilot tone from the received signal and independently demodulating the 10 components of the signal in said sub-bands into a pair of analogue signals; and converting each of said analogue signals into a respective digital data stream.

15. A transmitter constructed and arranged to operate substantially as hereinbefore described with reference to and as illustrated in figure 3 of the accompanying drawings.

16. A receiver constructed and arranged to operate substantially as hereinbefore described with reference to and as illustrated in figure 4 of the accompanying drawings.

17. A method of transmitting a TTIB signal substantially as hereinbefore described with reference to and as illustrated in figure 3 of the accompanying drawings.

18. A method of receiving a TTIB signal substantially as hereinbefore described with reference to and as illustrated in figure 4 of the accompanying drawings.