WO2011067103A1

WO2011067103A1 - System and arrangement for optical data transmission

Info

Publication number: WO2011067103A1
Application number: PCT/EP2010/067532
Authority: WO
Inventors: Thomas Kamalakis; Georgia Ntogari; Thomas Sphicopoulos; Joachim Walewski
Original assignee: Siemens Aktiengesellschaft
Priority date: 2009-12-01
Filing date: 2010-11-16
Publication date: 2011-06-09
Also published as: KR101486710B1; KR20120099745A; EP2484029A1

Abstract

The invention relates to a wireless visible light communication system based on discrete multitone coding including an optical transmitter fed by a pulse-width modulated signal. Preferably, the pulse-width modulated signal has a variable bandwidth for dimming purposes. According to a first aspect of the invention, the repetition frequency of the pulse-width modulated signal is higher than twice the frequency assigned to the fastest oscillating subcarrier. According to a second aspect of the invention, a decoding system for an optical wireless data transmission system including an optical receiver and including means for decoding the symbols by applying estimation factors considering self - interfering factor of a respective subcarrier frequency. The invention provides means for multiplying the signal dedicated to data transmission and the pulse-width modulated signal for dimming and, beyond that, leaves both signals independent of each other. In contrast to known approaches, the invention provides a decoupled control of dimming and VLC transmission on the transmitter side and compensates for the periodically truncated DMT signal on the receiver side by choosing appropriate demodulation schemes.

Description

System and Arrangement for Optical Data Transmission The Invention relates to a system and arrangement for an op^¬ tical transmission of data. Specifically, the invention re^¬ lates to visible-light communications based on discrete mul- titone coding.

In the field of indoor wireless networks, visible-light com^¬ munications is garnering increasing attention. One of the type of emitters used in this technology are white light- emitting diodes, which can synergistically provide both illu^¬ mination and data transmission.

As the demand for wireless data transmission constantly in^¬ creases, the radio frequency spectrum is becoming increas^¬ ingly congested. For this reason, attention has been drawn towards alternative communication technologies. The abundance of unregulated bandwidth in the optical frequency range makes optical wireless technologies attractive candidates for fu^¬ ture local area networks.

Optical wireless communications using infrared light-emitting diodes, »LEDs«, has been proposed, and is already commer^¬ cially available. Also, optical wireless communications using light-emitting diodes emitting in a visible spectral range has recently received increasing attention and is commonly referred to as visible-light communications or »VLC« .

In some use cases, light-emitting diodes in VLC systems serve a dual role in providing both illumination and wireless connectivity. The IEEE 802.15.7 task group is currently working towards a global VLC standard, in which lightning scenarios are explicitly covered.

Phosphorescent white light-emitting diodes constitute an at^¬ tractive technology for illumination systems. Compared to conventional lighting sources, such as incandescent and fluo^¬ rescent lamps, these light-emitting diodes are advantageous in terms of long life expectancy, high tolerance to humidity, low power consumption, ruggedness, and minimal heat genera- tion.

Another important property of light-emitting diodes is that they can readily be dimmed. It is further known in the art to use white light-emitting diodes lighting systems for communication.

Using discrete multitone modulation (DMT modulation) , VLC systems can provide high transmission rates with commercial high-power lighting light-emitting diodes.

As to dimming, in the case of light-emitting diodes, pulse- width modulation, in the following referred to as »PWM«, is a widely used means for accurately controlling LED illumination while incurring negligible color shifts of the emitted light. In PWM dimming, the brightness of the LED is controlled by square-pulse modulation using a fixed but adjustable repeti^¬ tion frequency of the driving current and by adjusting the duty cycle of the pulse train according to the dimming re- quirement.

A number of approaches relating to dimming methods in the physical layer have been proposed, which either use PWM or rely on changing the modulation depth so that both, bright- ness control and wireless communication can be achieved at the same time.

These approaches, however, are based on a subcarrier pulse position modulation and limited to a bit rates of up to ~ 1 Mb/s. One of the dimming approaches suggested by the standard draft document IEEE 802.15.7 is based on a variable pulse- position modulation, which emulates PWM but encodes the transmitted information in the position of the on-time. Here again the data throughput is limited by the PWM line rate.

It is an object of the present invention to provide means that allow for an independent control of dimming and data transmission while achieving a higher data rate for the transmission .

This object is solved by a coding system according to claim 1 and a decoding system according to claim 9. The object is further solved by a method according to claim 11.

The invention provides means for an optical wireless data transmission system including an optical transmitter fed by a pulse-width modulated signal. Preferably, the pulse-width modulated signal has a variable bandwidth for dimming pur^¬ poses.

Specifically, according to a first aspect of the invention, a coding system for an optical wireless data transmission system is provided. This system includes an optical transmitter, comprising means for segmenting symbols of data to be trans^¬ mitted into a number of parallel sequences, the number of parallel sequences corresponding to a number of orthogonal subcarriers having a proportional subcarrier frequency. Said »proportional« subcarrier frequency for the subcarriers means that the subcarriers have a frequency that can be written as an integer multiplied by a common subcarrier spacing. The system further comprises means for adjusting a repetition frequency for the pulse-width modulated signal to at least twice the frequency assigned to the fastest oscillating sub- carrier. Furthermore, the system comprises means for generat^¬ ing a multicarrier-signal by modulating each parallel se^¬ quence by its assigned subcarrier; means for converting the multicarrier-signal into an analogue multicarrier-signal; means for generating a composite signal by multiplying the analogue multicarrier-signal by a pulse-width modulated sig- nal and feeding the optical transmitter with the composite signal .

According to the first aspect of the invention, discrete mul- titone modulation is applied for visible-light communica^¬ tions. One of the issues to be addressed in these synergetic use cases is how to incorporate light dimming while not cor^¬ rupting the communication link. The invention provides means for multiplying the signal dedi^¬ cated to data transmission and the pulse-width modulated sig^¬ nal for dimming and, beyond that, leaves both signals inde^¬ pendent of each other. In contrast to known approaches, the invention provides a de^¬ coupled control of dimming and VLC transmission on the trans^¬ mitter side and compensates for the periodically truncated DMT signal on the receiver side by applying a suitable de^¬ modulation scheme.

According to the first aspect of the invention, the repeti^¬ tion frequency of the pulse-width modulated signal is equal to or higher than twice the frequency assigned to the fastest oscillating subcarrier.

Also, due to the prevalent high signal-to-noise ratio in in^¬ terior lighting scenarios, higher data rates can be achieved by exploiting spectrally efficient modulation of the DMT sub- carriers. One example is quadrature amplitude modulation.

According to a second aspect of the invention, a decoding system for an optical wireless data transmission system including an optical receiver, comprising an optical receiver for receiving a composite signal; means for decomposing the composite signal and for separating a multicarrier-signal; means for converting the multicarrier-signal into a digital multicarrier-signal; means for decomposing the digital multi^¬ carrier-signal into parallel sequences by demodulating each parallel sequence by a respectively assigned subcarrier hav^¬ ing a respective subcarrier frequency; means for de- segmenting each parallel sequence into symbols and means for decoding the symbols by applying estimation factors consider- ing self-interfering factor of a respective subcarrier frequency .

Preferred embodiments of the invention are set out in depend^¬ ent claims.

According to an embodiment of the invention, the means for generating a multicarrier-signal apply an inverse Fourier transform operation of each subcarrier sequence with its assigned subcarrier frequency.

According to a further embodiment of the invention, the means for generating a multicarrier-signal apply an addition of complex-conjugated subcarriers in order to attain a real- valued signal.

The invention will be described in more detail with reference to the figures.

Fig. 1A depicts a block diagram showing the structure of a visible-light communication system on the transmis^¬ sion side;

Fig. IB depicts a block diagram showing the structure of a visible-light communication system on the receiver side; depicts a timing diagram showing a pulse-width modulated signal; Fig. 3 depicts a timing diagram showing a composite sig^¬ nal ; Fig. 4 depicts a timing diagram showing a pulse-width modulated signal;

Fig. 5A depicts an inverse of SIR values against a ratio R_s in a general case, where the multicarrier signal and the pulse-width modulated signal are unsynchro- nized;

Fig. 5A depicts an inverse of SIR values against a ratio

R_s, where the multicarrier signal and the pulse- width modulated signal are synchronized;

Fig. 6A depicts a block diagram showing the structure of an alternative embodiment of the visible-light commu- nication system on the transmission side;

Fig. 6B depicts a block diagram showing the structure of an alternative embodiment of the visible-light commu^¬ nication system on the receiver side;

In Figure 1A a block diagram showing the structure of a visi^¬ ble-light communication system on the transmission side is depicted . An information source IS of any kind supplies information in the form of a data bit sequence. The data contained in the data bit sequence are to be converted by the coding system for an optical wireless data transmission system according to Figure 1A including an optical transmitter LED.

The data are fed to a conversion unit MP, where data bits are converted into a sequence of symbols, using, for example, a QAM (Quadrature Amplitude Modulation) constellation mapper.

The symbols produced by the conversion unit MP are fed to serial-to-parallel converter SP. This converter SP is seg^¬ menting the symbols into a number M of parallel sequences 1,2,3,...,M. The number of parallel sequences corresponds to a number of orthogonal subcarriers M having a respective subcarrier fre^¬ quency. The sequence number 1 corresponds to a subcarrier 1 having a subcarrier frequency of zero Hertz, viz. DC, the se^¬ quence number M corresponds to a subcarrier M having the maximum subcarrier frequency, or, in other words, this sub- carrier is the fastest oscillating subcarrier. The parallel sequences 2,3,...,M are fed to a modulator IFFT. A first sequence of the parallel sequences 1,2...,M shown in the drawing denoted by »1« is not fed to the modulator IFFT since the first parallel sequence corresponding to the sub- carrier having a subcarrier frequency of zero Hertz, or, DC shape, does not carry any information. This omitted feeding of the first sequence 1 is illustrated in the drawing by a line denoted by 1 disconnected on the side of the modulator IFFT . Further on, in order to generate a real-valued signal, the complex conjugates are fed to the modulator IFFT using taps M+l ... 2M of the modulator IFFT, while a cyclic symmetry is adhered. These details are not shown in the drawing. The description of disconnection of the first sequence 1 and the details of feeding the taps of the modulator IFFT are to be understood as an example of details connecting the modula^¬ tor IFFT und not as essentials of the invention. The parallel sequences 2,3,...,M are fed to a modulator IFFT by which each parallel sequence 2,3,...,M is modulated by its as^¬ signed subcarrier to produce a digital multicarrier-signal , which is converted to an analogue multicarrier-signal x(t) by a digital-analogue converter DAC .

For the modulation into a multicarrier-signal an inverse Fou^¬ rier transform operation of each subcarrier sequence with its assigned subcarrier frequency is carried out by the modulator IFFT. Alternative modulation operations are also within the scope of the invention.

The generated multicarrier signal x(t) is multiplied with a periodic pulse-width-modulated pulse train p(t) in a multi^¬ plier MU. The resulting composite signal y(t) = x(t)p(t) is the driving current of the optical transmitter LED.

This pulse-width-modulated signal p(t) is delivered by a dim- ming unit PWM, which is used to adjust a pulse width accord^¬ ing to a dimming factor for adjusting the radiant intensity of the transmitter LED.

Pulse-Width Modulation, or PWM, is a very efficient means for adjusting the average optical power emitted by the optical transmitter LED over a wide dimming range. The PWM signal consists of a periodic train of pulses, whereby the pulse width in a cycle is adjustable, consequently resulting in the variation of the DC level of the waveform.

It is noted that adjusting the dimming factor is independent of the coding of data to be transmitted, since the coded sig^¬ nal x(t) for transmitting data is simply impinged by a multi^¬ plying operation with the pulse-width-modulated signal p(t).

The pulse-width modulated signal p(t) is characterized by a duty cycle of d = T1/T_PWM, where Tl is the duration of the PWM pulse and T_PWM is the period of the pulse-width modulated sig^¬ nal. A dimming level δ is defined as δ = 1-d.

Turning now to Figure IB, which depicts the receiver side, or decoding system, of the visible light communication system, the transmitted composite signal y(t) is received by a opti^¬ cal receiver PD, e.g. a photodiode. The received optical com- posite signal y(t) is transformed to an electrical signal by the optical receiver PD and the electrical signal is fed to an analogue-digital converter ADC. The outlet of the analogue-digital converter ADC is one-tap equalized and fed to a de-modulator FFT for de-modulating the digital signal into a number of parallel sequences 2,3,...,M by respectively demodulating the signal with a respectively as- signed subcarrier having a respective subcarrier frequency. In this embodiment, the one-tap equalization operation is im^¬ plemented in the de-modulator FFT. In an alternative embodi^¬ ment, see Fig. 6B, this one-tap equalization operation is implemented in a particular device EQ.

For the de-modulation of the multicarrier-signal into a num^¬ ber of parallel sequences 2,3,...,M, or, baseband signals, a Fourier transform operation is carried out. Alternative de^¬ modulation operations are also within the scope of the inven- tion.

The parallel sequences 2,3,...,M are fed to a parallel-to- serial converter PS and de-segmented into a sequence of sym^¬ bols.

The generated symbols are decoded by means - not shown in Figure IB - using appropriate detection schemes as described below by applying estimation factors considering self- interfering factor of a respective subcarrier frequency.

The symbols are then fed to a de-conversion unit DMP, where the sequence of symbols are de-converted into data bits, us^¬ ing, for example, a QAM constellation mapper. The signal-processing chain consisting of the de-modulator

FFT and the parallel-to-serial converter PS is referred to as the demodulator.

In the visible-light communication system according to Fig- ures 1A and IB, intensity modulation with direct detection known as »IM/DD« (intensity modulation, direct detection) is used . Turning now to Figures 6A and 6B, which show an alternative embodiment of the of the visible light communication system on the transmission side and the receiver side, respectively. Most of the devices in these alternative embodiments are equal to the embodiments described in Figure 1A and Figure IB, respectively, so that only differences are discussed.

The transmitter side, according to an alternative embodiment shown in Fig. 6A, comprises a clipping device CLP, which is, by way of example, arranged between the outlets of the modu^¬ lator IFFT and the digital-analogue converter DAC . This clip^¬ ping device CLP sets any signal exceeding an upper threshold Xciipp or going below a lower threshold -x_cii_PP to an upper threshold x_ciip_P or -x_ciip_P, respectively. The advantage of this clipping device CLP is that undipped DMT signals commonly suffer from excessive positive and negative peaks, which would lower the average signal power fed to the optical transmitter LED if left untreated. The clipping device CLP can be combined with a scrambler. A scrambler suppresses the occurrence of repetitive data pat^¬ terns, e.g. blocks of "l"s, in the data to be transmitted. Long repetitions of the same data can lead to very large sig^¬ nal amplitudes after the inverse Fourier transform operation, and thus to recurring, strong clipping, which would result in a large interference noise floor. Such a - not shown - scrambler would be inserted between the information source IS and the conversion unit MP in Fig. 6A. Furthermore, a - not shown - de-scrambler would be needed after the de-conversion unit DMP in Fig. 6B .

Alternatively or additionally, the receiver side, according to an alternative embodiment shown in Fig. 6A, comprises a bias device T which is, by way of example, arranged between the outlets of the digital-analogue converter DAC and the multiplier MU. The bias device adds a bias current DC, which can be advantageously chosen independently of the signal x(t) . In the following section, the detection schemes according to the invention are described in detail.

The pulse-width-modulated signal p(t) consists of a periodic train of pulses, whereby the widths of the signal are adjust^¬ able, consequently resulting in the variation of the DC level of the waveform. The pulse-width-modulated signal p(t) is given by

whereby T_PWM is the period of the pulse-width-modulated signal p(t) and Ti is the »on«-time interval.

Figure 2 shows a time diagram of a pulse-width-modulated sig^¬ nal p(t) with a dimming factor of 20 %, indicating the period Tp_WM of the pulse-width-modulated signal and the »on«-time ΤΊ of the optical transmitter LED. The PWM signal p(t) is plot^¬ ted on the ordinate against time in microseconds on the ab^¬ scissa .

Figure 3 shows a time diagram of a respective composite sig^¬ nal y(t) for the same settings. A normalized radiated power of the composite signal y(t) is plotted on the ordinate against time in microseconds on the abscissa.

The repetition rate f_PWM = 1/T_PWM of the pulse-width-modulated signal p(t) has to be chosen with care. A high value for f_PWM could result in a large part of the driving current spectrum lying outside the 3-dB modulation bandwidth of the LED, leading to inefficient current-to-light conversion efficiency. On the other hand, very low values of f_PWM may result in flicker^¬ ing of the optical transmitter LED. According to the invention, multiple subcarrier schemes such as »Discrete Multitone Modulation«, DMT, are used to compen^¬ sate for the frequency dependence of the optical transmitter LED.

An undipped DMT-signal x(t) comprises of M-l subcarrier sig^¬ nals spaced by Af=l/T in the frequency axis, T being the du^¬ ration of the DMT symbol, and can be expressed as:

whereby s_m is the QAM symbol transmitted in the m subcarrier channel, a_m=2 f_m, f_m=(m-l)/T, and X_Dc is a DC component added to ensure that the current driving the optical transmitter

LED will always be positive. The symbols s jb_m are chosen from QAM constellations.

Hence, the possible values of a_m and b_m are given by

2u- (N^1/2 + 1), where l≤u≤N^1/2 and N is the number of QAM distinct symbols in the constellation, assumed to be a power of 2, i.e. N=2^L.

Note that equation (2) assumes that the DC subcarrier at fo=0 is not modulated.

When the DMT and PWM signals are combined to a composite sig nal y(t), the current y(t) driving the optical transmitter LED is the product of the DMT and PWM waveforms, i.e.: y(t) = x(t)p(t) (3)

In the following, the combination of PWM and DMT is referred to as »PWM-DMT modulations

A ratio R of the DMT symbol duration over the PWM period R = T/Tp_WM should not be made smaller than unity. Indeed, if R < 1 one can choose ΤΊ smaller than T_PWM-T and hence p(t) can be zero inside an entire DMT symbol duration. The inventors have investigated the performance of PWM-DMT both for integer and non integer values of R > 1. Figure 3 shows an example of a PWM-sampled DMT signal with 31 subcar- riers (M = 32) and 16-QAM (N = 16) on all subcarriers . The fastest oscillating subcarrier is positioned at f_M-i = 500 kHz and the symbols s_m are randomly chosen from the 16-QAM constellation. The PWM rate is f_PWM = 1 MHz, corresponding to a value of R = 64.

In the following, the symbol decoding is described.

Assuming a linear optical transmitter LED with a 3-dB bandwidth much larger than the signal bandwidth, the emitted optical power is proportional to the driving current y(t) . In the modulation-frequency range of interest, the free-space channel is flat, and the received decoded symbol s_fflfor sub- carrier m is determined by:

The proportionality constant A incorporates the LED current- to-light conversion efficiency, the gain of the optical chan^¬ nel, and the light-to-current conversion efficiency, as well as the amplifier gain at the receiver. This gain is chosen so that the spacing in the constellation of the received symbols is the same as that of the transmitted symbols.

A non-flat channel and/or a non-flat frequency response of the electro-optical components can readily be incorporated by dividing the right side of Equation (4) with the pertinent frequency response. To derive a formula for the symbol estimates at the receiver, it is assumed for the sake of simplicity that only two sub- carrier channels are active, located at ω = co_m and ω = ω_η, before a general case is described. For two subcarriers one can readily show that by substituting equation (3) into equation (4) and using equation (2), one obtains

where P (ω) is given by:

In (5) the term AX_DCP(co_m) does not depend on the symbols s_m or s_n and therefore affects the symbol decoding deterministi- cally. Equation (5) becomes:

The exponential terms of the DMT signal x(t) are orthogonal under the inner product

and hence, in the absence of PWM, i.e. y(t) = x(t), these signals do not interfere in equation (4) . In the presence of PWM however, the subcarrier signal components will interfere with their complex conjugates in equation (4), giving rise to a self-interference component . One can also ob-

serve the presence of two cross-interference components originating from subcarrier n.

According to equation (7), setting AP(0) = 1 ensures that, in the absence of PWM-induced interference, the received and transmitted QAM symbols will have the same spacing in the constellation. With the same approach as outlined for two subcarriers, one can derive the estimate for a PWM-DMT signal with M-l subcarriers:

The sum term in (8),

accounts for the cross-interference of all other subcarrier channels on the m^th subcarrier. This term becomes substantially zero if the repetition frequency for the pulse-width modulated signal is set to at least twice the frequency as^¬ signed to the fastest oscillating subcarrier, according to the invention.

This term converges even faster to zero if the pulse-width modulated signal p(t) and the multicarrier signal x(t) are synchronously held in phase, according to a further embodi- ment of the invention.

Hence, the inventive means for decoding the symbols suppress the cross-interference factor, which constitutes an interfer^¬ ence of a respective first subcarrier frequency to all other subcarrier frequencies different from the first subcarrier frequency in the case that

the pulse-width modulated signal and the analogue multi- carrier-signal are synchronously held in phase; and; the repetition frequency of a pulse-width modulated sig- nal in the composite signal is adjusted to at least twice the frequency assigned of the fastest oscillating subcarrier .

Suppressing the interference factor may particularly mean estimating this factor according to the equations above and subtracting the estimated factor. Another partial term of equation (8),

herein after referred to as self-interfering factor of a respective subcarrier frequency can be estimated by an estima^¬ tion factor for the purpose of decoding the symbols, accord^¬ ing to a further embodiment of the invention. Note that self- interfering factor in Eq. (8) systematically biases the esti^¬ mate s_m and has thus to be subtracted for a bias-free modu^¬ lation of the transmitted symbol

Also, notice that the partial term

in Equation (8), herein after referred to as »DC component«, systematically biases the estimate s_m and has thus to be subtracted for a bias-free modulation of the transmitted sym bol .

In the following, a synchronization of the pulse-width modulated signal p(t) and the multicarrier signal x(t) synchroni zation according to another aspect of the invention is described .

To properly calculate P (ω) , the relative timing between the multicarrier signal x(t) and the pulse-width modulated signal p(t) has to be taken into account.

If R = T/Tp_WM > 1, then multiple pulses of the pulse-width modulated signal p(t) may occur inside a single symbol of the multicarrier signal x(t). Since R is not generally an integer, there may be a residual duration 0 ≤ Tms - T_PWM for which only part of the PWM pulse resides within the DMT symbol.

Figure 4 shows a pulse-width modulated signal p(t) in the case of a non-integer relationship R between the cycle T (in Figure 4 denoted by T_DMT) of the multicarrier signal x(t) and the cycle T_PWM of the pulse-width modulated signal p(t) .

To calculate P (ω) one must therefore take into account the time offset τ between the positive rising edge of the PWM pulse and the beginning of the DMT symbol. The residual dura^¬ tion can be calculated from Ί_ΒΕ5 = T - T_PWMR' , where

R' = \_R] and |_xj is the largest integer which is smaller than x. In the general case, equation (6) becomes:

where

and

The spectrum Ρ_χ(ω) can be calculated numerically from equa^¬ tions ( 9) -( 11 ) . However, in the case where p(t) is given by (1), it is also possible to derive a closed form for P_x(co). According to equations ( 9) -( 11 ) , the spectrum Ρ_χ(ω) depends on the time offset τ which, for a long data sequence, is assumed to be uniformly distributed in [0 T_PWM] .

After some algebraic manipulation of equation (8), the vari^¬ ances V_mr and V_mi of the real and the imaginary parts of the symbol estimates can be deter^¬

mined through the following equations:

For ergodic symbol streams, averaging over τ readily obtains the averaged variances of the real and the

imaginary parts of the symbol estimates,

With the above results we define the signal-to-interference ratio, in the following referred to as SIR, for the m^th sub- carrier as :

where d_mi_n is the minimum symbol distance for the rectangular QAM constellation assumed in the present analysis. The dependence of the SIR on the ratio R_s is illustrated in Figures 5A and 5B for a dimming level of 80 %. Note that R_s is defined as the ratio of the PRM repetition frequency over the highest DMT subcarrier frequency, i.e. R_s ^_1 = T_PWMf_M.

In Figure 5A minimum of the inverse SIR value is represented by a doted line and by a solid line in Figure 5B, respec^¬ tively. The maximum of the inverse SIR value is represented by a solid line in Figure 5A and by a dotted line in Figure 5B, respectively. In both figures the inverse SIR values are plotted against the ratio R_s . Notice that any SIR degradation is to the impact of PWM sampling, since no additional noise is considered in this analysis. Figure 5A corresponds to the general case where the multicar- rier signal x(t) and the pulse-width modulated signal p(t) are unsynchronized, i.e. there is a random displacement τ be^¬ tween the positive edge of the PWM pulse and the DMT symbol as shown in Fig. 4. Herein, this situation is referred to as »unsynchronized« .

Fig. 5B shows the values of the inverse SIR obtained in the case where τ = 0, which is herein referred to as the »syn- chronized« .

In the unsynchronized case shown in Figure 5A the maximum and the minimum of the inverse SIR values exhibit rapid fluctua^¬ tions with varying R. The inverse SIR is reduced for R_s > 2, implying an improved system performance. As shown in the logarithmic plot the maximum 1/SIR, undergoes strong fluctuations in the same vi^¬ cinity. The minimum 1/SIR undergoes strong fluctuations for R_s > 0.5. Much smoother SIR variations are obtained in the synchronized system according to Figure 5B. For R_s > 2 the SIR becomes infinite, implying the absence of penalties due to PWM dimming.

The Figures verify the conclusion drawn by the invention that the level of interference for R_s > 2 is much smaller than for R_s < 2. In other words, the level of interference for a repe^¬ tition frequency of the pulse-width modulated signal p(t), which is at least twice the frequency assigned to the fastest oscillating subcarrier of the multicarrier-signal , x(t) is much smaller than for smaller repetition frequencies.

It is important to point out that the inventive adjustment of the repetition frequency of the pulse-width modulated signal p(t) to at least twice the frequency assigned to the fastest oscillating subcarrier of the multicarrier-signal is not a simple application of the Nykvist criterion. The Nykvist cri^¬ terion criterion relies on infinitesimally short sampling pulses, so-called Dirac pulses, while the PWM pulses, which serve as the sampling pulses according to the invention, ex- hibit considerable lengths.

The finite length of the sampling pulses destroys the or^¬ thogonality of the DMT subcarriers and leads to an interfer^¬ ence term upon demodulation, which is term two from the left on the right-hand side of equation (8) . However, an analyti^¬ cal expression for this term is provided and by aid of train^¬ ing symbols, i.e. pre-defined bit patterns, that are sent when a certain dimming level is chosen, this static interference term can readily be estimated and numerically subtracted from the demodulated signal, resulting in error-free DMT transmission in an otherwise noise-free channel. The same training symbols can be used to infer the DC term in equation (8) . In case one has access to the undistorted pulse-width modu^¬ lated signal on the transmitter side, synchronizing the mul^¬ ticarrier-signal x(t) can be achieved in a straight-forward manner, e.g. by triggering on the rising edge of the PWM waveform. Therefore, for R_s > 2, the combination of DMT with PWM is both very simple and, due to the lack of interference, very attractive.

Claims

1. A coding system for an optical wireless data transmission system including an optical transmitter, comprising:

means for segmenting symbols of data to be transmitted into a number of parallel sequences, the number of par^¬ allel sequences corresponding to a number of orthogonal subcarriers having a proportional subcarrier frequency; means for adjusting a repetition frequency for a pulse- width modulated signal to at least twice a subcarrier frequency assigned to a fastest oscillating subcarrier; means for generating a multicarrier signal by modulating each parallel sequence by its assigned subcarrier;

means for converting the multicarrier-signal into an analogue multicarrier-signal;

means for generating a composite signal by multiplying the analogue multicarrier-signal with the pulse-width- modulated signal and feeding the optical transmitter with the composite signal.

2. System according to claim 1, wherein said means for generating a multicarrier-signal apply an inverse Fourier trans^¬ form operation of each subcarrier sequence with its assigned subcarrier frequency.

3. System according to claim 2, wherein said means for generating a multicarrier-signal apply an addition of complex- conjugated subcarriers.

4. System according to one of the preceding claims, wherein symbols of data to be transmitted are converted out of data bits by using a quadrature-amplitude constellation mapper.

5. System according to one of the preceding claims, wherein a multiple of a period of the pulse-width modulated signal co^¬ incides with a period of the analogue multicarrier signal.

6. System according to one of the preceding claims, wherein the pulse-width modulated signal and the analogue multicar- rier signal are synchronously held in phase.

7. System according to one of the preceding claims,

including a clipping device (CLP) for limiting upper and lower thresholds of the multicarrier-signal .

8. System according to one of the preceding claims,

including a scrambler for suppressing an occurrence of repetitive data patterns of the multicarrier-signal.

9. A decoding system for an optical wireless data transmis^¬ sion system, comprising:

an optical receiver for receiving an optical signal, the optical signal being a composite signal by a multiplica^¬ tion of an analogue multicarrier-signal and a pulse- width-modulated signal, and for converting the optical signal into an electrical signal;

means for converting the electrical signal into a digi^¬ tal signal;

means for demodulating the digital signal into a number of parallel sequences by respectively demodulating the signal with a respectively assigned subcarrier having a respective subcarrier frequency;

means for de-segmenting each parallel sequence into sym^¬ bols;

means for decoding the symbols by applying estimation factors considering a self-interfering factor of a respective subcarrier frequency and a DC term, whereby both estimation factors are subtracted from each demodu^¬ lated signal of the individual subcarriers .

10. System according to claim 9, wherein the means for decoding the symbols suppress a cross-interference factor which constitutes an interference of a respective first subcarrier frequency to all other subcarrier frequencies different from the first subcarrier frequency in the case that the pulse-width modulated signal and the analogue multi- carrier-signal are synchronously held in phase; and;

a repetition frequency of the pulse-width modulated sig^¬ nal in the composite signal is adjusted to at least twice a subcarrier frequency assigned to a fastest os^¬ cillating subcarrier.

11. An optical wireless data transmission method including an optical transmitter fed by a pulse-width modulated signal, the method comprising the steps of:

segmenting symbols of data to be transmitted into a num^¬ ber of parallel sequences, the number of parallel se^¬ quences corresponding to a number of orthogonal subcar- riers having a respective subcarrier frequency;

- choosing a repetition frequency for a pulse-width modulated signal that is at least twice the subcarrier fre^¬ quency assigned to the fastest oscillating subcarrier; generating a multicarrier-signal by modulating each parallel sequence by its assigned subcarrier;

- converting the multicarrier-signal into an analogue mul^¬ ticarrier-signal;

generating a composite signal by multiplying the ana^¬ logue multicarrier-signal by the pulse-width modulated signal; and;

- feeding the composite signal to the optical transmitter.

12. Method according to claim 11, comprising the step of

receiving the composite signal by an optical receiver; converting the composite signal into a digital composite signal;

demodulating the digital signal into a number of paral^¬ lel sequences by respectively demodulating the signal with a respectively assigned subcarrier having a respec^¬ tive subcarrier frequency;

- de-segmenting each parallel sequence into symbols;

decoding the symbols by applying estimation factors considering a self-interfering factor of a respective sub- carrier frequency and a DC term, whereby both estimation factors are subtracted from each demodulated signal of the individual subcarriers

13. Method according to claim 12, wherein for decoding the symbols a cross-interference factor, which constitutes an in^¬ terference of a respective first subcarrier frequency to all other subcarrier frequencies different from the first subcar^¬ rier frequency, is neglected in the case that the repetition frequency of a pulse-width modulated signal in the composite signal is adjusted to at least twice the frequency assigned of the fastest oscillating subcarrier.