Optical transmitter

Abstract

There is disclosed an optical transmitter, comprising: at least a plurality of emitters which are optical emitters, each emitter being configured to emit light according to a state selected between at least a first state, in which the light is emitted at a first power level, and a second state, in which the light is emitted at a second power level different from the first power level, wherein the plurality of emitters are partitioned into a plurality of groups, so that all the emitters of the same group have the same state; an input interface (804), configured to receive a sequence of input values to be transmitted by the plurality of emitters; a symbol mapper (806), to map the sequence of input values onto one or more pulse amplitude modulation (PAM) symbols, each PAM symbol having a power level; and an emitter selector (810), to perform a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol.

Description

[0001] Optical transmiter

[0002] There are disclosed examples of optical transmitters, methods for performing optical transmissions, and non-transitory storage units for controlling optical transmissions. Examples also refer to a wide- beam LiFi transmitter, e.g. using multiple vertical cavity surface emitting laser (VCSEL) arrays e.g. with digital drivers.

[0003] In the following, different inventive embodiments and aspects will be described. Embodiments are also defined by the claims. It should be noted that any embodiments as defined by the claims can optionally be supplemented by any of the details (features and functionalities and details) described in the above mentioned chapters and / or subchapters. Also, the embodiments described in the above mentioned chapters and / or subchapters can be used individually, and can also be supplemented by any of the features in another chapter or subchapter, or by any feature included in the claims. Also, it should be noted that individual aspects described herein can be used individually or in combination. Thus, details can be added to each of said individual aspects without adding details to another one of said aspects. Moreover, features and functionalities disclosed herein relating to a method can also be used in an apparatus (configured to perform such functionality). Furthermore, any features and functionalities disclosed herein with respect to an apparatus can also be used in a corresponding method. In other words, the methods disclosed herein can optionally be supplemented by any of the features and functionalities described with respect to the apparatuses. Also, any of the features and functionalities described herein can be implemented in hardware or in software, or using a combination of hardware and software.

[0004] Some technical problems addressed by the present technical solution:

[0005] In principle, the optical transmission channel offers almost unlimited bandwidth and is therefore suitable for particularly high-bit-rate transmissions. In the field of optical wireless communication however, the actually available bandwidth is limited to some 10s of MHz due to the physical properties of the optical transmitters, more precisely the high-power LEDs that are usually used, which diminishes the advantage over conventional radio systems.

[0006] In addition, the transmission power of the optical components is limited by eye safety requirements, so that only a few meters can be bridged using infrared wavelengths, unless the optical beam is narrowly focused with a small emission angle, which in turn restricts practical applications in terms of user mobility and coverage. There is a trade-off between range or spatial coverage and throughput, as the available optical power can either be focused in a narrow beam to achieve high power at the receiver and thus a high throughput, or distributed over a larger area, but then the throughput is significantly lower due to the lower power at the receiver.

[0007] Prior art:

[0008] Due to the limitations described above, data rates in the range of 1 Gbit / s and more commonly expected by end users today can only be achieved using modulation methods with a high spectral efficiency (e.g. OFDM), which require good channel conditions and cause high energy consumption in the transmitter due to the complex electronics required for modulation.

[0009] Various approaches have been researched to control narrow optical light beams using mechanical or optical methods and actively direct them towards the mobile receiver in order to achieve a high range and data rate at the same time. However, this increases the complexity of the transmission system to an impractical extent.

[0010] To significantly increase the channel bandwidth, various types of lasers have been investigated, including VCSELs. These achieve bandwidths in the range of several GHz, but cannot provide the optical power required for optical wireless communication with user mobility in this bandwidth range (~ 1 W). In addition, lasers generate focused light, which significantly tightens the restrictions due to eye safety concerns.

[0011] Summary

[0012] According to an aspect there is provided an optical transmitter, comprising: at least a plurality of emitters which are optical emitters, each emitter being configured to emit light according to a state selected between at least a first state, in which the light is emitted at a first power level, and a second state, in which the light is emitted at a second power level different from the first power level, wherein the plurality of emitters are partitioned into a plurality of groups, so that all the emitters of the same group have the same state; an input interface, configured to receive a sequence of input values to be transmitted by the plurality of emitters; a symbol mapper, to map the sequence of input values onto one or more pulse amplitude modulation, PAM, symbols, each PAM symbol having a power level; and an emitter selector , to perform a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol.

[0013] The symbol mapper may be configured to generate the M-PAM symbols according to an M-PAM modulation.

[0014] The optical transmitter may assign one or more PAM symbols to a number of bits of at least one input value of the sequence of input values, and the number of states that the number of bits of the at least one input value shall be less or equal to the number of states that the one or more PAM symbols can have.

[0015] The symbol mapper may acquire the number A of groups of emitters connected, and to define the M-PAM modulation according to the relationship M = N + 1 after that A is acquired.

[0016] The emitter selector may perform the selection by choosing among a plurality of possible group combinations which potentially reach the power level of the PAM symbol.

[0017] The emitter selector may, when performing the selection according to a second PAM symbol in a second time slot which immediately follows a first time slot in which the selection according to a first PAM symbol according to a first group combination has been performed, choose a second group combination which maximizes the number of transitions from the first state to the second state and / or from the second state to the first state.

[0018] The emitter selector may, when performing the selection according to a second PAM symbol in a second time slot which immediately follows a first time slot in which the selection according to a first PAM symbol has been performed, evaluate whether the transition from the first PAM symbol to the second PAM symbol implies a required number T of groups transitioning from a state to another one which verifies 0<T<N-l with N being the number of groups, to cause, beyond the T transitionings, at least one first additional group to additionally and redundantly transition from the first state to the second state, and at least one second additional group to additionally and redundantly transition from second state to the first state.

[0019] The emitter selector may select states of different groups randomly. The optical transmitter may select between at least one first configuration mode and at least one second configuration mode, wherein: in the first configuration mode, the plurality of emitters are partitioned into a first plurality of N groups, and in the second configuration mode, the plurality of emitters are partitioned into a second plurality of N’ groups with N’<N.

[0020] The optical transmitter may: in the first configuration mode, each group of the first plurality of N groups is capable of emitting a first highest-power PAM signal at a first highest-power emitting power, and in the second configuration mode, each group of the second plurality of N’ groups is capable of emitting a second highest-power PAM signal at a second highest-power emitting power which is larger than the first emitting power.

[0021] The optical transmitter may: in the first configuration mode, each group of the first plurality of N groups has a first number of emitters, and in the second configuration mode, each group of the second plurality of N’ groups has a second number of emitters greater than in the first configuration mode.

[0022] The optical transmitter may change from the first configuration mode to the second configuration mode in case it is determined that the channel has a quality worse than a predetermined threshold; and / or configured to change from the second configuration mode to the first configuration mode in case it is determined that the channel has a quality better than a predetermined threshold.

[0023] The optical transmitter may, in case it is determined that the channel has a quality worse than a predetermined threshold, reduce the number of the groups for each group and / or in case it is determined that the channel has a quality better than a predetermined threshold, to increase the number of the groups for each group.

[0024] The optical transmitter may select between at least one full-power mode and at least one second, power-saving mode, wherein, in the power-saving configuration mode, each group has at least one emitter which is deactivated, and, in the full-power mode, the at least one emitter is reactivated.

[0025] The optical emitters may be VCSEL emitters or sub-arrays of VCSEL emitters.

[0026] The emitters may be collected in subarrays, all the emitters of one subarrays being selectable to have the same state.

[0027] Each group may be defined using a discrete number of subarrays.

[0028] The emitter selector may define each group of emitters in such a way that light emitted by different groups overlap, so as to reach the power level of the PAM symbol by cumulative power level from all the groups.

[0029] According to an aspect there is provided a method for performing an optical transmission, through at least a plurality of emitters which are preferably optical emitters, each emitter being configured to emit light according to a state selected between at least a first state, in which the light is emitted at a first power level, and a second state, in which the light is emitted at a second power level different from the first power level, wherein the plurality of emitters are partitioned into a plurality of groups, so that all the emitters of the same group have the same state; wherein the method includes: receiving a sequence of input values to be transmitted by the plurality of emitters; mapping the sequence of input values onto one or more pulse amplitude modulation, PAM, symbols, each PAM symbol having a power level; and performing a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol.

[0030] According to an aspect there is provided a non-transitory storage unit storing instruction which, when executed by a processor, cause the processor to control an optical transmission by mapping a sequence of input values onto one or more pulse amplitude modulation, PAM, symbols, each PAM symbol having a power level; and performing a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol. Figures

[0031] Fig. 1 compares the present technique with other techniques.

[0032] Fig. 2 shows a schematic of an experimental setup.

[0033] Fig. 3 shows experimental results of the present technique.

[0034] Figs. 4, 5 A, 5B, 5C, 5D, 6 A, and 6B show examples of operations according to the present technique. Fig. 7 is subdivided between Fig.7(a) which shows a hardware example of the present technique and Fig.7(b) which shows a mapping strategy.

[0035] Figs. 8A, 8B, 8C, 9 A, 9B, and 9C show examples of operations according to the present technique. Fig. 10 shows symbol mapping generation tables according to strategies.

[0036] Examples

[0037] Here below, examples of optical transmitters 800 are provided. An optical transmitter 800 may include a plurality of light emitters 820. Each light emitter may be a light emitting diode, LED. Preferably, each light emitter may be a vertical cavity surface emitting laser (VCSEL) emitter. Each emitter 820 may be one single device, or a plurality (e.g., an array) of devices, all having the same transmitting state. Whether single or array, each emitter 820 may be part of a group (modulation unit). The transmitter 800 may receive an input 820 (or 820’) and transmit one or more PAM symbols. More in detail, the input value 802 (802’) may be mapped onto one or more PAM symbols 808 (in electric version) through a symbol mapper 806. An emitter selector 810 may select (through selections 820bl, 820b2, 820cl, 820c2, 820c3, 820c4, 812a9, etc.) some emitters 820 or groups of emitters to be selected to have a particular state at each time instant. In some examples, the PAM symbols 808 (in electric version) may be converted into OOK vectors 832 (each OOK vector 832 being a vector of selections for each emitter 820 and / or group of emitters for the particular time instant). The OOK vector 832 may control, in a physical layer block 838, AWG (Arbitrary Waveform Generator) channel block(s) 834, which in turn may control OOK drivers 837 (e.g., one driver for each group of emitters), which in turn may control the emitters 820 through groups 820b 1 and 820b2. The summed (cumulated, composed) light signal 822b, generated by the composition (accumulation, sum) of the light signals transmitted by the two groups 820b 1 and 820b2, will generate the PAM symbol 822b (optical signal).

[0038] The optical signal 822b may be received by a photodetector 902 and converted onto an electric form 904. The received signal 904 in electric form may be provided to an electric reader 906 (e.g. oscilloscope or an analog-to-digital converter) which may provide the received signal 908 in digital form, corresponding to the selections (e.g. 820b, 820bl, 820b2, etc.) or sum of the transmitted OOK vectors. Then, a PAM demodulator 910 may provide the decoded signal 912, corresponding to the input values 802, 802’.

[0039] Each emitter 820 emits light [e.g. in the free space] according to a state selected (e.g. based on OOK vector 832) between at least:

[0040] 1) a first state [e.g. first power state] [e.g. indicated with -1], in which the emitter emits light at a first power level, and

[0041] 2) a second state [e.g. second power state] [e.g. indicated with +1], in which the emitter emits light at a second power level different from the first power level,

[0042] The plurality of emitters 820 is partitioned into a plurality of groups [or modulation units] [in some examples, all modulation units (groups of emitters) may have the same (e.g. identical) power level, or are so that they cause identical power differential between the two output states], so that all the emitters 820 of the same group have [e.g. select] the same state [e.g. for at least one time slot], [e.g. the emitters may be modulated using a OOK modulation]. For example:

[0043] 1) In Fig. 8A, there is one single group 820 with eight emitters.

[0044] 2) In Figs. 5A-6B and 8B, there are two groups 820b 1 and 820b2, each with four emitters.

[0045] 3) In Fig. 8C, there are four groups 820cl, 820c2, 820b3, 830b4, each with two emitters.

[0046] It will be shown that in some examples the groups are variable and defined on the fly, but in other examples the groups are fixed and pre-defined.

[0047] In some examples, the first state may be a 0-power state (e.g. in which no light is actually emitted), and the other one may be a non-0-power state (e.g. in which light is actually emitted, e.g. in a particular wavelength or in a particular range of wavelengths). Each emitter may be controlled through on-off-keying (OOK) modulation. Each emitter 820 of a same group may be controlled with the same selection: therefore, all the emitters of the same group may be, simultaneously, selected according to the same selection (OOK vector), so as to have the same state. For example:

[0048] 1) In Fig. 8 A, all the emitters of the single group 820a are selected though same selections 812a (i.e., all the emitters of the single group 820a are simultaneously either all +1 or all -1).

[0049] 2) In Figs. 5A-6B and 8B, all the emitters of the group 820b 1 are selected through same selections 812bl, while all the emitters of the group 820b2 are selected through same selections 812b2, and emiters of different groups are selected independently from each other (i.e., the emitters of the group 820b 1 could have in principle a different state of the emiters of the group 820b2 in case the selection 812b 1 differs from the selection 812b2, or, said in other terms, different groups may be independently selected).

[0050] 3) In Fig. 8C, all the emiters of each of the groups 820cl, 820c2, 820c3, 830c4 are respectively selected through same selections 812cl, 812c2, 812c3, 812c4, and emitters of different groups are selected independently from each other (i.e., the emitters of the group 820cl could have in principle a different state of the emitters of the group 820c2, 820c3, 820c4 in case the selection 812cl differs from the other selections, etc., or, said in other terms, different groups may be independently selected).

[0051] The “same selection” may be understood as implying that the emiters selected according to the same selection have, simultaneously, the same power state (i.e. they are associated with the same element of the OOK vector 832).

[0052] The light emitted by the emitters 820 may be cumulated (composed) with each other (e.g. in free space, or anyway after having been generated), to generate a composed or cumulative PAM power level (e.g. 822a, 822b, 822c, 822a9, 822b9 in Figs. 8A, 8B, 8C, 9A, 9D, respectively). The composed or cumulative PAM power level (822a, 822b, 822c, 822a9, 822b9) may therefore represent an M- PAM symbol. In the case that light from emitters 820 with different state are cumulated with each other, intermediate PAM symbols may be generated. For example:

[0053] 1) Different emiters having the same state +1 will have a cumulative power level which is +1, and the resulting PAM symbol will be +1 (this is the case of Fig. 5B).

[0054] 2) Different emitters having the same state -1 will have a cumulative power level which is -1, and the resulting PAM symbol will be -1 (this is the case of Fig. 5D).

[0055] 3) Different emitters having different states (+1 vs -1) will have a cumulative power level which is 0, and the resulting PAM symbol will be 0 (this is the case of Fig. 5C).

[0056] Therefore, a receiver device acquiring the PAM symbol (e.g., by detecting the composed or cumulative PAM power level 822a, 822b, 822c, 822a9, 822b9, etc.) may detect:

[0057] 1) A PAM symbol which is +1 when detecting a cumulative power level which is +1.

[0058] 2) A PAM symbol which is -1 when detecting a cumulative power level which is -1.

[0059] 3) A PAM symbol which is 0 when detecting a cumulative power level which is 0. Hence, the information will be correctly detected by the receiver device.

[0060] Notably, there may be, in some cases, multiple combinations of selections that may provide the same PAM symbol (i.e., multiple OOK vectors may provide the same PAM symbol): for example, in Fig. 8B, the PAM symbol 0 could be obtained by any of

[0061] 1) selecting -1 for the 1stgroup 820bl (through selection 812bl) and +1 for the 2ndgroup 820b2 (through selection 812b2)

[0062] 2) selecting +1 for the 1stgroup 820b 1 (through selection 812bl) and -1 for the 2ndgroup 820b2 (through selection 812b2).

[0063] In general terms, if there are N groups of emitters 820 that can simultaneously take different states, then there may be N+l (or M, with M=N+1) PAM symbols that can be transmitted simultaneously. In general terms, we write “PAM symbol” or “M-PAM symbol” or “PAM-M symbol” without distinction.

[0064] The cumulative PAM power level (e.g. 822a, 822b, 822c, 822a9, 822b9) represents, in optical form, the electric version 808 of the PAM symbol of Figs. 7, 8A, 8B, 8C, 9A, 9B, and 9C.

[0065] In general terms, the information to be transmitted is not natively in PAM format, but is received in a different format, such as a bit sequence (digital sequence), or more in general a sequence of input values 802 (802’), which may have a different alphabet of the M-PAM symbols. For example, a digital value is written in one bit, i.e. 0 or 1, while Figs. 5A-6B, 8B, and 9B permit an encoding with a 3-PAM modulation (i.e., one of 3 3-PAM symbols may be encoded for each time instant). To increase efficiency, it is possible to map different input sequences 802 (802’) onto different M-PAM symbols.

[0066] For example, in the case of Fig. 4 we have a 4-PAM modulation (i.e. a 4-PAM symbol may be generated with three groups of emitters simultaneously operating); hence, here N=3 and M=N+1=4. Accordingly, each word (sequence) of 2 bits in the sequence of input values 802 (802’) may be simultaneously encoded in (assigned to) one 4-PAM symbol. For example:

[0067] 1) The word 00 in the input values 802 (802’) could be encoded by (assigned to) each group having state -1. 2) The word 01 in the input values 802 (802’) could be encoded by (assigned to) one single group having state -1 and the other two groups having state +1.

[0068] 3) The word 10 in the input values 802 (802’) could be encoded by (assigned to) one single group having state +1 and the other two groups having state -1.

[0069] 4) The word 11 in the input values 802 (802’) could be encoded by (assigned to) each group having state +1.

[0070] In the case of the sequence of input values 802 (802’) being a digital sequence, the example above works easily in the cases in which the number of M-PAM states that can be generated simultaneously is an integer power of 2 (i.e. if M an integer power of 2). However, in the case in which M is not an integer power of 2 (or more in general if the alphabet of the M and the alphabet of the sequence of input values 802 (802’) have not the same number of symbols, or the number of symbols of the sequence of input values is not an integer multiple or integer submultiple of M), then a more sophisticated encoding can be chosen, to reach high efficiency. The more sophisticated encoding may take into consideration the elapsing of time, so that a sequence of more than one PAM symbols are assigned to encode a sequence (word) of input values (802, 802’), for example. Fig. 7, part (b), shows how a word (sequence) of 3 bits of the sequence of input values 802 (802’) may be encoded in two 3 -PAM symbols to be transmitted in two time slots subsequent one to the other. This is because the input word (802, 802’) can take one value among 23=8 values, and two 3 -PAM symbols may take 32=9 values (by virtue of 32=9>8=23, we also have one 3-PAM symbol which is non-used (-1, +1 in Fig. 7 part (b)), or that could be used for providing other information, such as a parity bit for providing redundancy information). Therefore, a timely- defined PAM symbol sequence may efficiently provide information on any sequence of bits. An example is provided by Fig. 6A, which shows a sequence of PAM symbols [0, 1] as transmitted at subsequent time instants (immediately subsequent time slots) [tl, t2], to encode the word 110; and, in Fig. 6B, a sequence of PAM symbols [-1, 0] as transmitted at subsequent time instants (immediately subsequent time slots) [tl, t2], to encode the word 001.

[0071] Some applications are discussed here below.

[0072] Controlling the number of transitions

[0073] The emitter selector 808 may perform selections intelligently (e.g. controlled in a way to work towards / reach a specific goal / effect), so as to maximize the number of transitions. Keeping in mind the example of Figs. 6 A and 6B, and remembering that, for at least some PAM symbols, it is possible to choose different selections of different groups to obtain the same PAM symbol (i.e. different OOK vectors may generate the same PAM symbol), it may be that:

[0074] • in a first time slot tl a first PAM symbol is generated through a first selection which causes a first combination of states for the groups; and

[0075] • in a second (immediately subsequent) time slot t2 a second PAM symbol is generated through a second selection which causes a second combination of states for the groups, but

[0076] • in the transition between the time slots tl and t2, the emitter selector 808 acts so as to change as many as states of the groups of emitters 820 as possible (or at least to vary them randomly), thereby maximizing the transitions.

[0077] For example, with reference to Fig. 4:

[0078] • in a first time slot tl, a PAM symbol with value +1 is transmitted (as an example) through the OOK vector [+1, +1, -1] (i.e.: 1stgroup of emitters: +1; 2ndgroup of emitters: +1; 3rdgroup of emitters: -1)

[0079] • in the second time slot t2, a different PAM symbol with value -1 can be transmitted, which can be indifferently obtained through any of the OOK vectors (selections) [+1, -1, -1] and [- 1, -1, +1] and [-1, +1, -1]

[0080] • in the transition from tl to t2, it could be possible to simply cause one single transition by modifying the OOK vector from [+1, +1, -1] to [+1, -1, -1] (because only the 2ndgroup of emitters would transition to a different selection) of from [+1, +1, -1] to [-1, +1, -1] (because only the first group of emitters would transition to a different selection)

[0081] • however, it may be preferred to transition from tl to t2 by modifying three selections i.e. from OOK vector [+1, +1, -1] to [-1, -1, +1] (notably, the OOK vector [-1, -1, +1] provides the same PAM symbol as the OOK vector [+1, -1, -1]

[0082] • or, it may be alternatively preferred to choose randomly between the OOK vectors [+1, -1, -1] and [-1, -1, +1] and [-1, +1, -1]

[0083] In other examples, it may be possible to have the goal of minimizing or otherwise influencing the number of transitions (globally over all emitters / drivers or per emitter / driver across multiple symbol times). In this way, power efficiency could be increased by avoiding state transitions of individual drivers. The emitter selector 808 may evaluate whether the transition from the first PAM symbol in slot tl to the second PAM symbol in the immediately subsequent slot t2 implies a required number T of groups transitioning from a state (power level) to another state (power level). The evaluation may verify whether 0<T<N-l (N being the number of groups) and cause, beyond the T transitionings, at least one first additional group to additionally and redundantly transition from a first state to a second state, and at least one second additional group to additionally and redundantly transition from the second state to the first state [the number of first additional groups being the same of the number of second additional groups, e.g. between 1 and (N-T) / 2],

[0084] Configuration modes by modifying the groups

[0085] Different configuration modes may be chosen. With different configuration modes, the number of groups of emitters may change. This means that, according to the particular configuration mode chosen, some emitters 820 may change their group. In general terms, configuration modes partitioned in few groups (but each group having many emitters) are highly reliable, but have lower bitrate, while configuration modes partitioned in many groups (but each group having a small number of emitters) are less reliable, but have higher bitrate.

[0086] The optical transmitter 800 may select between at least one first configuration mode and at least one second configuration mode: in the first configuration mode, the plurality of emitters are partitioned into a first plurality of N groups [or modulation units] [e.g., so as to permit an M-PAM, with M = A + 1], and / or in the second configuration mode [e.g. like in Fig. 8B], the plurality of emitters are partitioned into a second plurality of N’ groups [or modulation units] with N’<N [i.e. the second configuration mode has less groups than the first configuration mode] [e.g., so as to permit an M’-PAM, with M' = N' + 1 < M, i.e. each PAM symbol in the second configuration mode can carry less information in than each PAM symbol in the first configuration mode].

[0087] For example:

[0088] • (N’=l) the second configuration mode may be that of Fig. 8A, where the totality of emitters 820 are in one single group (and therefore one single selection 812a controls all the emitters 820).

[0089] • (N=2>N’=1) the first configuration mode may be that of Fig. 8B, where the plurality of emitters 820 is partitioned among a first group 820b 1 and a second group 820b2 (e.g., the two groups 820b 1 and 820b2 may have the same number of emitters each, the two groups 820b 1 and 820b2 may have the same total emitting power).

[0090] • (N=4>N’=1) alternatively, the first configuration mode may be that of Fig. 8C instead of that of Fig. 8B, where the plurality of emitters 820 is partitioned among four groups 820cl-820c4 (e.g., the different groups may have the same number of emitters each, or have the same total emitting power).

[0091] Alternatively,

[0092] • (N’=2) the second configuration mode may be that of Fig. 8B, where the plurality of emitters 820 is partitioned among a first group 820b 1 and a second group 820b2 (e.g., the two groups 820b 1 and 820b2 may have the same number of emitters each, the two groups 820b 1 and 820b2 may have the same total emitting power).

[0093] • (alternatively, the second configuration mode may be that of Fig. 8A with N’=l)

[0094] • (N=4>N’=2) the first configuration mode may be that of Fig. 8C, where the plurality of emitters 820 is partitioned among four groups 820cl-820c4 (e.g., the different groups may have the same number of emitters each, or have the same total emitting power).

[0095] The optical transmitter 800 may choose the configuration mode based, for example, on a manual selection, a predefined selection, or based on an automatic selection. The automatic selection may imply the detection of a feedback from the channel (e.g., an impulse response as measured on the channel), so that the most appropriate configuration mode is chosen. For example, if the impulse response or another measurement is indicative of a high quality of the channel, then a configuration mode with a high number of groups (but each with a small number of emitters, like in Fig. 8C) may be chosen. Otherwise, if the impulse response or the other measurement is indicative of a low quality of the channel, then a configuration mode with a small number of groups (but each with a great number of emitters, like in Fig. 8A) may be chosen. The impulse response or the other measurement indicative of the channel quality may be either measured by a receiver (not shown) associated with (e.g. in the same device as) the transmitter 800 (e.g. by measuring a pilot sequence transmitted by a transmitter associated to the receiver, e.g. 900), or signalled by a further transmitter (associated to the receiver 900), e.g. after that the further receiver 900 has measured a pilot sequence (e.g. transmitted by the transmitter 800 or by another device). Therefore, the optical transmitter 800 can change from the first configuration mode (high number of groups with few emitters 820 each) to the second configuration mode (low number of groups with many emitters 820 each) in case it is determined that the channel has a quality [e.g. either measured by it or after the reception of a signalization that encodes a feedback information (e.g. from the receiver side) e.g. indicating the quality of the channel] worse than a predetermined threshold [e.g. the optical transmitter may perform channel measurement, so as to acquire the channel quality, or may receive the feedback from the feedback signalization in the received side]; and / or the optical transmitter 800 can change from the second configuration mode to the first configuration mode in case it is determined that the channel has a quality [e.g. either measured by it or after the reception of a signalization that encodes a feedback information (e.g. from the receiver side) e.g. indicating the quality of the channel] better (i.e. higher) than a predetermined threshold [e.g. the optical transmitter may perform channel measurement, so as to acquire the channel quality, or may receive the feedback from the feedback signalization in the received side].

[0096] For example, each group of the first plurality of N groups [or modulation units] may be capable of emitting a first highest-power PAM signal at a first highest-power emitting power, and in the second configuration mode, each group of the second plurality of N’ groups [or modulation units] (with N’<N) is capable of emitting a second highest-power PAM signal at a second highest-power emitting power which is larger than the first emitting power [e.g., because in the first configuration mode each group has less emitters than in the second configuration mode, and therefore in the first configuration mode less emissive power is achieved]

[0097] [and / or in the second configuration mode, each of the second plurality of N’ groups [or modulation units] (with N’<N) is capable of emitting the second PAM signal at an average power level which is higher than the average power level of the at which each group of the first plurality of N groups is capable of emitting the first PAM signal]

[0098] [in the second configuration mode, the difference (differential) of power between the closest PAM symbols is higher than in the first configuration mode, because in the second configuration mode the M’-PAM modulation has less power levels than the M-PAM modulation in the first configuration mode, since M’<M, while the extremity values (e.g. 0 power, and maximal power) is the same for both the configuration modes; therefore, the second configuration mode can be expected to be more reliable than the first configuration mode].

[0099] According to an aspect, in the first configuration mode, each group of the first plurality of N groups [or modulation units] has first number of emitters, and in the second configuration mode, each group of the second plurality of N’ groups [or modulation units] has a second number of emitters greater than in the first configuration mode [The power levels (and thus, differentials between the power levels) scales with group size].

[0100] According to an aspect, the optical transmitter may in case it is determined that the channel has a quality [e.g. either measured by it or after the reception of a signalization that encodes a feedback information (e.g. from the receiver side) e.g. indicating the quality of the channel] worse than a predetermined threshold [e.g. the optical transmitter may perform channel measurement], to reduce the number of the groups [and consequently we have a smaller M] for each PAM symbol and to increase the numbers of emitters [and / or the power level of] for each group and / or in case it is determined that the channel has a quality [e.g. either measured by it or after the reception of a signalization that encodes a feedback information (e.g. from the receiver side) e.g. indicating the quality of the channel] better than a predetermined threshold [e.g. the same threshold as above] [e.g. the optical transmitter may perform channel measurement], to increase the number of the groups [and consequently we have a larger M] for each PAM symbol and to reduce the numbers of emitters [and / or the power level of] for each group.

[0101] According to an aspect, the optical transmitter may select between at least one full-power mode and at least one second, power-saving mode, wherein, in the power-saving configuration mode, each group has at least one emitter which is deactivated, and, in the full-power mode, the at least one emitter is reactivated [and / or in the full-power mode all the emitters for each group are activated, and for in the power-saving mode less than all the emitters for each group are activated] [and / or in the full-power mode there are activated more emitters for each group than in the power-saving mode] [it is to be noted that “activated” does not necessarily mean “transmitting”, because it is possible that an activated group of emitters is in the first state, and in some examples the “first state” implies that a 0 power level is implied] [all the groups can have the same numbers of deactivated / reactivated emitters.

[0102] Grouping M-PAM symbol temporal sequences for M-PAM transmission

[0103] It is now explained how to associate an alphabet of M-PAM temporal sequences with an alphabet of input values. Notably, a putative alphabet of M-PAM symbol temporal sequences may be taken into account (e.g. so as each M-PAM symbol temporal sequence could be associated to a corresponding input signal 802, 802’), but some exceeding M-PAM symbol temporal sequences may be unutilized, and therefore the putative alphabet may be shortened (restricted) by excluding the unutilized M-PAM symbol temporal sequences. The exclusion may be based on a particular criterion and may be based by evaluating a metric (such as cumulative power / amplitude per sequence, i.e., the sum of number of active arrays, but other metrics may be evaluated) for all M-PAM temporal sequences of the putative alphabet.

[0104] In cases where the number of input symbols (e.g., binary words) 802, 802’ is not an integer multiple of the number of symbol states M, a number of S individual M-PAM symbols (e.g. each transmitted in S subsequent time slots) can be grouped to enhance the symbol space and increase the code efficiency, i.e., the number of used symbol groups out of the MAS available ones (where MAS means M at the power of S, often represented with M8, and implying for example: if M=3 and S=2, then MAS=3A2=9, if M=5 and S=4, then MAS=5A4=625). Therefore, the putative alphabet of M-PAM symbol temporal sequences includes MAS entries, but we need only 2An entries (because the input value 802, 802’ is considered to have n bits). Given that we need MAS>2An, in case of MAS>2An (i.e. MAS^2An) we find a strategy to restrict the putative alphabet, so as to obtain an alphabet with MAS- 2An entries. Hence, MAS-2An entries (M-PAM symbol temporal sequences) are excluded from the final alphabet to be used.

[0105] The symbol mapper 805 may execute the following steps:

[0106] 1. Mapping of binary words of n bits (i.e. each word having n bits) to PAM symbol sequences of S symbols (M-PAM symbols) with M levels

[0107] 1.1. Selection strategy for 2An used symbol sequences out of the MAS possible sequences (e.g., uniform distribution, lowest average signal power)

[0108] 1.2. Strategy for assignment of binary words to symbol sequences to minimize the impact of symbol errors on binary decoding

[0109] 1.3. Application of a code to facilitate the above criteria

[0110] 1.4. Increasing redundancy by shortening mapped bit words below the theoretical maximum, i. e. use a larger symbol space then necessary

[0111] For step 1.1, the following steps can be followed:

[0112] Procedure to select symbol sequences based on sequence properties: o Generate (or at least take into account) a putative alphabet with all possible MAS symbol sequences o Calculate metric such as cumulative power / amplitude per sequence (i.e., sum of number of active arrays, but other metrics may be evaluated) for all M-PAM temporal sequences o Sort M-PAM temporal sequences by metric o Select first or last 2An sequences for set of sequences to be used

[0113] Procedure to select M-PAM temporal sequences uniformly: o Generate (or at least take into account) a putative alphabet with all possible MAS symbol sequences o Generate (or at least associate) indices for all possible 2An binary words of the input 802, 802’ o Scale indices by MAS / 2An o Round scaled index to assign nearest symbol sequence index

[0114] In the result, the selection of possible symbol states and groups can be influenced to gain favorable properties in the transmitted signal, e.g., with regard to influencing the average signal power or the average power differential between symbols to maximize robustness against detection errors.

[0115] Examples for different mapping tables for the two methods described for step 1.1 above are given in Fig. 10 for two different modes using 3-PAM (M-PAM modulation with M = 3). The left table is generated with the goal of minimizing the average signal amplitude of the optical signal 822, and thus the values are selected so that the symbol sequence [1, 1] is omitted, as it represents the highest average signal amplitude. This can be extended to larger symbol groups and modulation orders by selecting multiple words to omit following the same criteria.

[0116] The right table is generated with the goal of maximizing the average power differential between words. For this purpose, the M-PAM symbol temporal sequence [0, 0] is omitted to maximize the occurrence of steps between +1 and -1 states.

[0117] The assignment of binary words to symbol groups can be rearranged to facilitate step 1.2 above, i.e., minimize the Hamming-distance between similar symbol groups to minimize the probability of bit errors for single symbol detection errors. Therefore, in accordance with step 1.1 it may be provided to associate a putative alphabet of M- PAM temporal sequences to an alphabet of input sequences (802, 802’), e.g. by: o Calculating a metric (such as a cumulative power / amplitude metrics per sequence, i.e., sum of number of active arrays, but different metrics could be used) for each M- PAM temporal sequence of the MAS temporal sequences of the alphabet of M-PAM temporal sequences. o Sorting the MAS M-PAM temporal sequences of the alphabet of M-PAM temporal sequences by metric (e.g., those M-PAM temporal sequences of the alphabet having higher metric being sorted before those M-PAM temporal sequences of the alphabet having low metric, or those M-PAM temporal sequences of the alphabet having lower metric being sorted before those M-PAM temporal sequences of the alphabet having high metric), so as to have a ranking between the M-PAM temporal sequences of the alphabet (the highest ranked M-PAM temporal sequences being those highest in the sorting and therefore having highest (or smallest) metric). o Selecting first 2An M-PAM temporal sequences or (alternative) the last 2An M-PAM temporal sequences to be associated to the alphabet of input sequences (802, 802’). The selecting is according to a chosen criterion. Some criteria are:

[0118] Minimizing average signal amplitude of the optical signal 822 (and thus the values are selected so that the symbol sequence [1, 1] is omitted, as it represents the highest average signal amplitude) (in this case, [1, 1] would be the highest-ranked M-PAM temporal sequence when sorting the alphabet of M-PAM temporal sequences according to the metric of cumulative power / amplitude metrics, and, when adopting the criterion of minimizing average signal amplitude of the optical signal 822, it would be excluded from the used alphabet)

[0119] Maximizing the average power differential between M-PAM symbols, like in the right table of Fig. 10. For this purpose, the M-PAM symbol temporal sequence [0, 0] is omitted to maximize the occurrence of steps between +1 and -1 PAM states (in this case, [0, 0] would be the lowest- ranked M-PAM temporal sequence when sorting the alphabet of M-PAM temporal sequences according to the metric of cumulative power / amplitude metrics, and, when adopting the criterion of maximizing the average power differential between M-PAM symbols, it would be excluded from the used alphabet). o Accordingly, M-PAM temporal sequences which least respond to a selected criterion may be excluded from the alphabet.

[0120] It is also possible to round the alphabet of the input 802, 802’ by MAS / 2An.

[0121] It is possible to select between at least: one first association between the alphabet of input values 802, 802’ and a first alphabet of M-PAM symbol temporal sequences (e.g., left side in Fig. 10) and one second association between the alphabet of input values 802, 802’ and a second alphabet of M-PAM symbol temporal sequences (e.g., right side in Fig. 10).

[0122] The optical transmitter may base the selection e.g. between the first association and the second association, e.g. based on a manual selection and / or on a particular selection criterion established for the optical transmission. For example, the selection criterion may be coherent (or even the same) with the criterion through which the alphabet has been obtained from the putative alphabet: e.g., in order to minimize the average signal amplitude of the optical signal 822, then the first association is chosen (i.e., using the mapping strategy on the left side of Fig. 10); and in order to maximize the average power differential between M-PAM symbols, then the second association is chosen (i.e., using the mapping strategy on the right side of Fig. 10)

[0123] Examples according to the present technical solution:

[0124] Our solution is based on two components. On one hand, lasers, specifically VCSELs, are used, and on the other hand the modulation is simplified as much as possible using on-off- keying (OOK).

[0125] Specifically, multiple small VCSELs each having an optical outputs in the milliwatt range, are combined as arrays to reach the required powers in the range of 100 mW to 1W or even higher. Diffusers are used to increase the spatial emission angle and maintain eye safety.

[0126] This principle is enhanced by placing multiple independent sub-arrays of VCSELs on the receiver, that are each driven by individual drivers. The independent signals for these drivers are superimposed in the analog domain, by electrical linking of the outputs of modulation can be achieved by independent modulation of the VCSEL sub-arrays.

[0127] Furthermore, these independent units of drivers and VCSEL sub-arrays can be grouped into N groups where N equal to the number of independently driven sub-arrays or an integer factor of it. These groups are further referred to as modulation units.

[0128] Logically, the input data is mapped onto PAM symbols with a modulation order M (M- PAM). The M signal levels are converted into a number of modulation units that need to be on and off, respectively. Due to this, M can be as high as the number of independent OOK- modulated drivers N+l . Expressed differently, for the modulation of an M-PAM signal a number of N=M-1 independent modulation units, that are each modulated using at least one OOK driver and VCSEL-subarray is needed.

[0129] The resulting spectral efficiency can be varied between 1 bit per symbol for simultaneous OOK (equivalent to 2 -PAM) modulation on all modulation units and log2(M) = log2(N+l) Values of M that are not equal to powers of 2 can be realized by grouping of M-PAM symbols and assigning longer bit sequences to each group. Example for PAM-3: o A combination of two PAM-3 symbols can represent 9 different states o Using 8 of these states, log2(8) = 3 bits per group of two symbols can be transmitted, resulting in a modulation depth of 1.5bits per symbol o An exemplary code table for this case is included in the annex

[0130] For modulation of the OOK modulation units, N data streams are created as follows: o From the M-PAM symbols, the number of drivers that transmit a “high” state (+1), and the number of modulation units that transmit a “low” state (-1) are derived. These states are then randomly assigned to modulation units. o For the example PAM-3 (M = 3, N +2):

[0131] * PAM-3 symbol +1 : both modulation units outputting +1

[0132] * PAM-3 symbol -1: both modulation units outputting -1

[0133] * PAM-3 symbol 0: one modulation unit outputs +1, the other one -1. The specific assignment of symbols to modulation units is done randomly, in order to avoid long runs of equal states per driver.

[0134] Advantages of the present technical solution:

[0135] The use of OOK allows the deployment of very simple drivers, that offer high bandwidth at high output power and provide high energy efficiency. This allows the maximization of signal power and reach within the limits of eye safety regulations. Furthermore, commercial drivers with matching specification exist for fiber-based applications, already.

[0136] The high bandwidth of the lasers enables reaching a high throughput despite of the low modulation order of OOK. Switching between high order and lower order modulation by superimposing independent modulation units that are each OOK modulated allows balancing reach and throughput according to channel conditions.

[0137] The high efficiency, reach, and low driver complexity of OOK is combined with the high throughput of higher order modulation.

[0138] Channel adaptation can be realized purely by logical changes in the digital domain.

[0139] Some technical applications:

[0140] Indoor OWC or LiFi systems with several Gbit / s throughput with mobile users and overlapping communication cells, i.e., a similar usage scenario to WLAN today, or as a high- bit-rate supplement to these systems

[0141] OWC systems in industrial production facilities

[0142] Wireless communication in areas with high user density and / or sensitivity to electromagnetic interference, such as exhibition halls, airplanes, and hospitals.

[0143] An example (e.g. 7.75 Gbit / s LiFi Transmitter Using High-Power VCSEL Arrays)

[0144] We report on a wide-beam LiFi transmitter using multiple VCSEL arrays with digital drivers. We transmit a 3.75 GBd PAM-3 signal at a gross data rate of 5.625 Gbit / s. Singledriver OOK operation yields 7.75 GBd.

[0145] 1. Introduction

[0146] Future applications like holographic virtual and augmented reality, autonomous driving, ultra-high definition streaming, and cloud gaming place higher demands on the achievable capacity of both fixed and mobile networks. One major building block to meet these high requirements is the exploration of higher frequency bands for mobile communication, including the use of the optical spectrum. Networked optical wireless communication (OWC), also called LiFi, uses light instead of radio frequencies to transmit and receive wireless data using multiple access points. Furthermore, interference between cells is well controllable due to the propagation properties of light, so that the additional spectrum can be reused efficiently and a high area capacity can be achieved. This is especially interesting for off-loading data from radio networks such as Wi-Fi through an optical overlay network. The current generation of LiFi systems can reach up to 1 Gbit / s per link with DC-biased OFDM and adaptive bit-loading together with modern FEC [1-3], However, the low bandwidth of LEDs in the 1 Os of MHz range and poor power efficiency of OFDM limit the achievable data rate and reach. For LED-based LiFi, significant improvements in power efficiency and range of LiFi transmitters have been shown for on-off-keying (OOK) modulation compared to DC-biased OFDM [4, 5],

[0147] To reach higher data rates and maintain sufficient coverage in each LiFi cell, the use of high- power vertical cavity surface emitting laser (VCSEL) arrays instead of LEDs has been proposed [6], The key advantage of VCSELs is a much higher possible modulation bandwidth, while VCSEL arrays can reach similarly high optical output powers as LEDs. This makes OOK as a modulation scheme interesting also for high throughput applications, since power becomes more of a limiting factor than bandwidth. In addition, OOK drivers are available from fiber-optical applications that are capable of providing several 100 mA of drive current at symbol rates of multiple GBd. In [7], we demonstrated up to 4 Gbit / s with OOK, driving four VCSEL sub-arrays yielding a combined optical power of 350 mW by four limiting-amplifier drivers to transmit a single OOK signal. In this work, we analyze the physical limits of the same transmitter prototype by combining it with a higher bandwidth receiver and explore two alternative driving modes to further increase the feasible throughput, reaching 5.625 Gbit / s and 7.75 Gbit / s, respectively. Fig. 1 compares our new throughput results with previous publications, with respect to the optical output power in each case [7-12], While higher data rates can be reached with small VCSEL arrays, suitable e.g. for fixed wireless applications, our results address mobile use cases where a significantly more power illuminates a large area to enable user mobility.

[0148] The new insight is that high-power VCSEL arrays originally developed for LIDAR enable both high speed and high mobility.

[0149] 2. Experimental Setup with Two VCSEL Sub-Arrays

[0150] In our previous setup [7], we modulated all four drivers and all four sub-arrays with identical OOK signals to achieve the highest possible signal power. Here, we use the same Brightlaser VD-0940V- 140M-1C-5A1 10-aperture VCSEL arrays, each with a nominal power of 140 mW and coupled with a diffusor with emission angles of 60° and 45° in x- and y-dimensions, respectively. The lasers operate at a wavelength of 940 nm and are powered by MAX3736 drivers specified for a symbol rate of 3.2 GBd. As shown in Fig. 2, the outputs of the two active drivers are connected electrically and jointly drive two VCSEL arrays. As a receiver, we use a Coherent Solutions matrlQ- O2E 1201 receiver with a bandwidth of 35 GHz, effectively removing any band- limitations on the receiver side. The light is coupled into the receiver using a lensed fiber placed directly in front of the diffusor at one of the VCSEL arrays on the transmitter. Signals are generated with a Keysight M8190A arbitrary waveform generator at symbol rates from 1 to 8 GBd in steps of 250 MBd. The receiver is connected to a LeCroy WavePro 804HD oscilloscope, operating at a sample rate of 20 GS / s. For equalization, a time domain equalizer (TDE) based on a 10-tap FIR filter is applied. Filter coefficients are estimated using the least squares algorithm. Demodulation without an equalizer, using averaging and decimation, is included for comparison.

[0151] To enhance the throughput, the two drivers connected to one of the series of VCSEL arrays on the prototype are modulated separately to create a PAM-3 signal by overlaying two OOK signals. Without any changes to the hardware, this increases the spectral efficiency by 50 %. In a further step, we disconnect one of the drivers on the board to increase signal fidelity by removing impairment effects from the parallel drivers, shown by the orange ‘x’ in Fig. 2. Note that a balanced PAM-2 constellation (+1 / -1) is used throughout to represent OOK, as the drivers are operating in an AC-coupled mode, producing balanced signals that are added to a constant bias. We stick with the OOK nomenclature, nevertheless, due to the resulting optical signal representation. For the PAM-3 measurement, as indicated by the dashed ‘PAM-3 mapping’ box, a PAM-3 signal is first generated and then divided into two OOK vectors that are supplied to the inputs of the two OOK drivers separately. The PAM-3 signal may be then reconstructed through the superposition of the electrical signals at the driver outputs. To map the binary data to PAM-3 symbols, blocks of two PAM-3 symbols may be grouped together to represent 3 bits of input data using 8 out of the possible 9 resulting symbol combinations, providing a spectral efficiency of 1.5 bits / symbol. The OOK driving vectors may be created using the following rules: To create “+1” symbols, both OOK symbols are also set to “+1”; the same applies vice versa for “-1” symbols. For every “0” input symbol, one “+1” and one “-1” symbol are randomly assigned to the two OOK vectors. The goal of this is to avoid long runs of “-1” or “+1” symbols in one of the OOK vectors. For channel estimation, a Gold sequence

[0013] with a length of 1,024 symbols is prepended to the data stream. In the case of PAM-3, two Gold sequences are OOK modulated and added together to form PAM- 3 symbols. The payload data consists of (or more in general comprises) 820,224 symbols, with the same number of bits of random data for OOK and 1,230,336 bits for PAM-3.

[0152] 3. Experimental Results

[0153] BER results over symbol rate are shown in Fig. 3. The displayed BER threshold of2.76e-2 is valid for an LDPC code with a code rate of 5 / 6 and a block length of 4320 bits, with a target block error rate of 10'

[0014] , For comparison, a threshold of 4.5e-3 is also shown, for an LDPC code rate of 20 / 21. Fig. 3 A shows the results for the OOK setup with both drivers active. As a reference (Ref), our corresponding results from [7] are shown in gray. With regard to that reference, the removal of the band limitation at the receiver increases the maximum symbol rate slightly from 4 to GBd within the FEC limit of 4.5e-3. With PAM-3, this limit is crossed after 3.25 GBd, resulting in a throughput of 4.875 Gbit / s, and in the single driver setup after 5 GBd, or 5 Gbit / s. The FEC limit of 2.76e-2 is crossed after 5 GBd for OOK with both drivers, and after 3.75 GBd or 5.625 Gbit / s in the PAM-3 setup. In the single driver setup, the largest increase can be observed: Despite locally peaking at 6 GBd, the BER crosses the higher threshold after 7.75 GBd (7.75 Gbit / s).

[0154] 4. Discussion and Outlook

[0155] Compared to modulating all sub-arrays jointly with one OOK signal, transmitting PAM-3 through two separately modulated VCSEL sub-arrays here yields up to 15% higher data rate at the same power. The highest data rates can be reached by driving the sub-arrays with a single driver using OOK, but then the modulation amplitude is reduced. This shows that the path towards higher data rates is i) to utilize multiple VCSEL arrays each having an individual driver, and ii) allowing high flexibility in logically grouping these sub-arrays on the transmitter, realizing various levels of spectral efficiency. For a transmitter with four independently controllable VCSEL sub-arrays, joint modulation of all sub-arrays realizes OOK at a basic data rate. For PAM-3, two groups of two VCSEL sub-arrays each can be formed, which are modulated with different OOK driving vectors. For PAM-5, all four sub-arrays are fed with different OOK driving vectors. In all cases, the effective SNR remains the same, but the more signal levels are realized, the higher the required SNR for error-free transmission, and vice versa. This allows adapting the rate to channel conditions in a mobile scenario, where users are moving freely. While this is commonly achieved by using OFDM with adaptive bit- loading and linear drivers, our scheme can use digital drivers, which are more energy-efficient.

[0156] 5. Conclusion

[0157] To the best of our knowledge, these are the highest data rates achieved so far with high-power VCSEL arrays that enable a wide optical beam for mobile LiFi applications. Our measurements show that the individual modulation of drivers for two VCSEL sub-arrays allows PAM-3 with symbol and data rates of 3.75 GBd and 5.625 Gbit / s, respectively, at bit error rates below the threshold of common LDPC FEC. With individual drivers and OOK, we achieved symbol rates of up to 7.75 GBd. These results indicate that moving towards further “pixelation” of the VCSEL array, with individual drivers per pixel (where a pixel is a sub-array), is the way to scale up data rates. This both increases the bandwidth for transmission of OOK signals and enables dynamic link adaptation. Higher-order PAM signals are realized by logical grouping of sub-arrays sending OOK driving vectors which are superimposed to form, e.g., OOK, PAM-3 or PAM-5 symbols. This new approach for adaptation to variable channel conditions in a LiFi system with mobile users relies on the use of binary drivers for each sub-array and low complexity logic for mapping the OOK driving vectors onto the pixelated VCSEL array.

[0158] Further characterization of the figures

[0159] Fig. 1. Placement within the prior art

[0160] Fig. 2. Schematic of the experimental setup. The orange ‘x’ marks the disconnection of one of the drivers for the single driver (s. drv.) setup

[0161] Fig. 3. BER for (a) Baseline setup, (b) PAM-3 setup, (c) Single driver setup.

[0162] References

[0163] [1] D. Schulz et al., “Use Cases for Optical Wireless Communication,” in 2018 Optical Fiber Communications Conference and Exposition (OFC), 2018.

[0164] [2] H. Haas, L. Yin, Y. Wang and C. Chen, "What is LiFi?," in Journal of Lightwave Technology, vol. 34, no. 6, pp. 1533-1544, 2016.

[0165] [3] L. Grobe et al., „High-speed visible light communication systems, “ IEEE Comm. Magazine, pp. 60-66, Dec. 2013.

[0166] [4] M. Hinrichs et al., "A Physical Layer for Low Power Optical Wireless Communications," in IEEE Transactions on Green Communications and Networking, vol. 5, no. 1, pp. 4-17, March 2021, doi: 10.1109 / TGCN.2020.3038692.

[0167] [5] M. Hinrichs et al., "A Performance Comparison of OFDM and Pulsed PHY Modulations in Optical Wireless Communications", South American Conference on Visible Light Communication (SACVLC), 2023.

[0168] [6] M. Yoshikawa et al., "High power VCSEL devices for free space optical communications," Proceedings Electronic Components and Technology, 2005. ECTC '05., Lake Buena Vista, FL, USA, 2005, pp. 1353-1358 Vol. 2, doi: 10.1109 / ECTC.2005.1441445.

[0169] [7] M. Hinrichs et al., “4 Gbit / s Optical Wireless Communication with High-Power Transmitter”, ECOC 2023, paper We.A.2.5.

[0170] [8] R. Kirrbach et al., “High Power Eye-Safe Optical Wireless Gigabit Link Employing a Freeform Multipath Lens,” IEEE Communications Letters, pp. 1-1, 2022, doi: 10.1109 / LCOMM.2022.3162642.

[0171] [9] Z. Wei et al., “Full-duplex high-speed indoor optical wireless communication system based on a micro-LED and VCSEL array,” Optics Express, vol. 29, no. 3, pp. 3891-3903, Feb. 2021, doi: 10.1364 / OE.412348.

[0172]

[0010] S. Liverman et a., “VCSEL Array-Based Gigabit Free-Space Optical Femtocell Communication,” Journal of Lightwave Technology, vol. 38, no. 7, pp. 1659-1667, Apr. 2020, doi: 10.1109 / JLT.2019.2958733.

[0173]

[0011] H. L. Minh et al., “A gigabit / s indoor optical wireless system for Home Access Networks,” in 20107th International Symposium on Communication Systems, Networks & Digital Signal Processing (CSNDSP 2010), Jul. 2010, pp. 532-536.

[0174]

[0012] H. Kazemi et al., “Experimental Demonstration of 17 Gb / s Transmission Over Indoor Optical Wireless Communication Links Using A Low- Power Single-Mode VCSEL,” in Optical Wireless Communication Conference, Dec. 2023.

[0175]

[0013] R. Gold, “Optimal Binary Sequences for Spread Spectrum Multiplexing.” IEEE Trans. Inf. Theory, 13, 619-621 (1967).

[0176]

[0014] P. W. Berenguer e al., "The benefit of frequency-selective rate adaptation for optical wireless communications," Proc. CSNDSP 2016.

[0177] Some comments on the discussions above and below

[0178] In examples, above, reference has been made, for example, to an AWG (Arbitrary Waveform Generator) blocks 834. However, another type of device or unit may be used. It may be replaced with a more generic term, e.g., digital-to-analog converter (1 and 2). Actually, the block 834 may be completely omitted, as the concept would be applicable to drivers that accept digital inputs. The AWG blocks 834 can be understood as representing an interface between the logical emitter selector and the physical (OOK) drivers 837 of the VCSELs 820b 1 and 820b2. This applies similarly to the receiver, where the oscilloscope (906) may be particular to a lab setup, while a system “in the field” would use an analog-to-digital-converter combined with digital signal processing to demodulate signals.

[0179] Summarizations on the aspects above

[0180] According to an aspect, there is provided an optical transmitter [e.g. for transmission through free space, e.g. for optical wireless transmission], comprising: at least a plurality of emitters which are optical emitters [e.g. laser emitters, e.g. VCSEL emitters, or more in general laser emitters, or general LED emitters], each emitter being configured to emit light [e.g. in the free space] according to a state selected between at least a first state [e.g. first power state] [e.g. indicated with -1], in which the light is emitted at a first power level, and a second state [e.g. second power state] [e.g. indicated with +1], in which the light is emitted at a second power level different from the first power level, wherein the plurality of emitters are partitioned into a plurality of groups [or modulation units] [in some examples, all modulation units (groups of emitters) may have the same (identical) power, or are so that they cause identical power differential between the two output states], so that all the emitters of the same group have [e.g. select] the same state [e.g. for at least one time slot], [e.g. the emitters may be modulated using a OOK modulation] [in some examples, one of the first state and the second state may be a 0-power state (e.g. in which no light is actually emitted), and the other one may be a non-0-power state (e.g. in which light is actually emitted, e.g. in a particular wavelength or in a particular range of wavelengths); this may be in accordance with the OOK modulation] [in some examples, all the emitters have the same power] [in examples, all emitters of each group have the same selected state]; an input interface, configured to receive a sequence of input values [e.g. in binary, digital form, each input value having, for example, a particular number of possible states (e.g. a bit can have two states, a byte can have 256 states, etc.)] to be transmitted by the plurality of emitters; a symbol mapper, to map the sequence of input values onto one or more pulse amplitude modulation, PAM, symbols, each PAM symbol having a power level [e.g. the power level being associated with the sum or composition of the light waves emitted by the emitters according to their state; for example, in case of two groups of emitters and each emitter having a state selected from two states (-1, +1), the PAM symbol may have a power level which is one of three power levels (indicated with -1, 0, +1), i.e. a first power level indicated with -1 (as resulting from the composition of -Is from both the groups), a second power level (+1, as resulting from the composition of +ls from both the groups), and an intermediate power level 0 (as resulting from the composition between the first state (-1) of the first group and the second state (+1) of the second group)]; and an emitter selector, to perform a selection, for each group of emitters, between the first state and the second state so as to reach [e.g. in the free space] the power level of each PAM symbol [e.g. so that the cumulative power level emitted by the plurality of groups (e.g. in one time slot) represents the power level of the PAM symbol that has been generated in the by the symbol mapper] [and / or so that the light emitted by the plurality of emitters represents, encoded therein according to the PAM modulation, the input values of the sequence of input values] [e.g., the emitter selector selects the groups of emiters in such a way that the PAM symbols are converted onto OOK symbols] [as shown below, in some examples different selections of groups can be performed in different configuration modes]

[0181] [in examples, the PAM symbols are according to an M-PAM modulation, which has order M (M is the number of different power levels that can be represented by an M-PAM symbol); there may be the relationship M = N + 1 where N is the number of groups of emitters (number of modulation units)].

[0182] [the numerical values labelling the power levels are +1 and -1 for the individual OOK emiters and also from -1 to +1 for the resulting PAM signal, as OOK / PAM-2 is our baseline when all emitters are grouped. With this denomination the PAM symbols can be thought of as the average of the individual modulated OOK values, though in the physical sense it may be understood as a superposition of the emitter outputs]

[0183] According to an aspect, the symbol mapper may generate the M-PAM symbols [e.g. at least in one configuration mode] according to an M-PAM modulation, [in some examples, M can be different from 2 [M #= 2] at least in one configuration mode, but in some other examples, M=2 at least in one configuration mode] [or in other examples, M is not an integer power of 2, e.g. M #= 2P, where p is a natural number at least in one configuration mode, but in some other examples, it could be that M = 2Pat least in one configuration mode] [as can be seen below, in the cases in which different configuration modes are used, it may be that different values of M are used, and in some cases M can also be 2],

[0184] The optical transmitter may assign one or more PAM symbols to a number of bits of at least one input value of the sequence of input values, and the number of states that the number of bits of the at least one input value shall be less or equal to the number of states that the one or more PAM symbols can have [e.g. a bit can be an input value, and has 21=2 possible states, while one single 3-PAM symbol has 3 possible states and two 3-PAM symbols have 32=9 possible states, therefore the “number of bits of the at least one input value” may be 3 bits and with 23=8 possible states, so that the 3 bits are encoded in two 3-PAM symbols] [even if we lose some states, we still can encode on average an increased number of input values].

[0185] According to an aspect, the symbol mapper may acquire the number A of groups of emitters are connected, and to define the M-PAM modulation according to the relationship M = A + 1 after that TV is acquired] [or, in some examples, according to define the M-PAM according to the relationship M < N + l][This solution can be applied in the implementations in which the emitters can be externally activated vs deactivated].

[0186] According to an aspect, [e.g. in at least one configuration mode] the emitter selector is configured to perform the selection [e.g. for each time slot] by choosing among a plurality of possible group combinations which potentially (and / or effectively) reach the power level of the PAM symbol [e.g. there can be many possible combinations for obtaining a particular PAM symbol, but the emitter selector will only select one of them].

[0187] According to an aspect, [e.g. in at least one configuration mode] the emitter selector may, when performing the selection according to a second PAM symbol in a second time slot which immediately follows a first time slot in which the selection according to a first PAM symbol according to a first group combination has been performed, choose a second group combination which maximizes [or at least varies, e.g. randomly] the number of transitions from the first state to the second state and / or from the second state to the first state

[0188] According to an aspect, [e.g. in at least one configuration mode] the emitter selector may, when performing the selection according to a second PAM symbol in a second time slot which immediately follows a first time slot in which the selection according to a first PAM symbol has been performed, and evaluate whether the transition from the first PAM symbol to the second PAM symbol implies a required number T of groups transitioning from a state to another one which verifies 0<T<N-l with N being the number of groups, cause, beyond the T transitionings, at least one first additional group to additionally and redundantly transition from the first state to the second state, and at least one second additional group to additionally and redundantly transition from second state to the first state [the number of first additional groups being the same of the number of second additional groups, and being between 1 and (N-T) / 2],

[0189] According to an aspect, [e.g. in at least one configuration mode] the emitter selector may [e.g. for at least one time slot] [e.g. in at least one configuration mode] select states of different groups randomly [e.g. so as to maximize variation of the states].

[0190] According to an aspect the optical transmitter may select between at least one first configuration mode and at least one second configuration mode, wherein: in the first configuration mode, the plurality of emitters are partitioned into a first plurality of N groups [or modulation units] [e.g., so as to permit an M-PAM, with M = TV + 1], and / or in the second configuration mode [e.g. like in Fig. 8B], the plurality of emitters are partitioned into a second plurality of N’ groups [or modulation units] with N’<N [i.e. the second configuration mode has less groups than the first configuration mode] [e.g., so as to permit an M’-PAM, with M' = N' + 1 < M, i.e. each PAM symbol in the second configuration mode can carry less information in than each PAM symbol in the first configuration mode]

[0191] [there can be more than two configuration modes, having less and less numbers of groups],

[0192] [In one example, the first configuration mode may be that of Fig. 8A and the second configuration mode may be that of Fig. 8B or 8C. In another example, the first configuration mode may be that of Fig. 8B and the second configuration mode may be that of Fig. 8C.]

[0193] According to an aspect, in the first configuration mode, each group of the first plurality of N groups [or modulation units] is capable of emitting a first highest-power PAM signal at a first highest-power emitting power, and in the second configuration mode, each group of the second plurality of N’ groups [or modulation units] (with N’<N) is capable of emitting a second highest-power PAM signal at a second highest-power emitting power which is larger than the first emitting power [e.g., because in the first configuration mode each group has less emitters than in the second configuration mode, and therefore in the first configuration mode less emissive power is achieved]

[0194] [and / or in the second configuration mode, each of the second plurality of N’ groups [or modulation units] (with N’<N) is capable of emitting the second PAM signal at an average power level which is higher than the average power level of the at which each group of the first plurality of N groups is capable of emitting the first PAM signal]

[0195] [in the second configuration mode, the difference (differential) of power between the closest PAM symbols is higher than in the first configuration mode, because in the second configuration mode the M’-PAM modulation has less power levels than the M-PAM modulation in the first configuration mode, since M’<M, while the extremity values (e.g. 0 power, and maximal power) is the same for both the configuration modes; therefore, the second configuration mode can be expected to be more reliable than the first configuration mode]. According to an aspect, in the first configuration mode, each group of the first plurality of N groups [or modulation units] has first number of emitters, and in the second configuration mode, each group of the second plurality of N’ groups [or modulation units] has a second number of emitters greater than in the first configuration mode [The power levels (and thus, differentials between the power levels) scales with group size].

[0196] According to an aspect, the optical transmitter may change from the first configuration mode to the second configuration mode in case it is determined that the channel has a quality [e.g. either measured by it or after the reception of a signalization that encodes a feedback information (e.g. from the receiver side) e.g. indicating the quality of the channel] worse than a predetermined threshold [e.g. the optical transmitter may perform channel measurement, so as to acquire the channel quality, or may receive the feedback from the feedback signalization in the received side]; and / or configured to change from the second configuration mode to the first configuration mode in case it is determined that the channel has a quality [e.g. either measured by it or after the reception of a signalization that encodes a feedback information (e.g. from the receiver side) e.g. indicating the quality of the channel] better than a predetermined threshold [e.g. the optical transmitter may perform channel measurement, so as to acquire the channel quality, or may receive the feedback from the feedback signalization in the received side].

[0197] According to an aspect, the optical transmitter may in case it is determined that the channel has a quality [e.g. either measured by it or after the reception of a signalization that encodes a feedback information (e.g. from the receiver side) e.g. indicating the quality of the channel] worse than a predetermined threshold [e.g. the optical transmitter may perform channel measurement], to reduce the number of the groups [and consequently we have a smaller M] for each PAM symbol and to increase the numbers of emitters [and / or the power level of] for each group and / or in case it is determined that the channel has a quality [e.g. either measured by it or after the reception of a signalization that encodes a feedback information (e.g. from the receiver side) e.g. indicating the quality of the channel] better than a predetermined threshold [e.g. the same threshold as above] [e.g. the optical transmitter may perform channel measurement], to increase the number of the groups [and consequently we have a larger M] for each PAM symbol and to reduce the numbers of emitters [and / or the power level of] for each group. According to an aspect, the optical transmitter may select between at least one full-power mode and at least one second, power-saving mode, wherein, in the power-saving configuration mode, each group has at least one emitter which is deactivated, and, in the full-power mode, the at least one emitter is reactivated [and / or in the full-power mode all the emitters for each group are activated, and for in the power-saving mode less than all the emitters for each group are activated] [and / or in the full-power mode there are activated more emitters for each group than in the power-saving mode] [it is to be noted that “activated” does not necessarily mean “transmitting”, because it is possible that an activated group of emitters is in the first state, and in some examples the “first state” implies that a 0 power level is implied] [all the groups can have the same numbers of deactivated / reactivated emitters.

[0198] According to an aspect, the optical emitters are VCSEL emitters or sub-arrays of VCSEL emitters. According to an aspect, the emitters are collected in subarrays, all the emitters of one subarrays being selectable to have the same state.

[0199] According to an aspect, each group is defined using a discrete number of subarrays.

[0200] According to an aspect, the emitter selector is configured to define each group of emitters in such a way that light emitted by different groups overlap, so as to reach the power level of the PAM symbol by cumulative power level from all the groups.

[0201] According to an aspect, there is provided a receiver apparatus comprising: at least one optical receiver for receiving a PAM signal; a signal mapper to convert the PAM signal onto an output signal.

[0202] According to an aspect, there is provided a method for performing an optical transmission, through at least a plurality of emitters which are preferably optical emitters [e.g. laser emitters, e.g. VCSEL emitters, or more in general laser emitters, or general LED emitters], each emitter being configured to emit light according to a state selected between at least a first state [e.g. first power state] [e.g. indicated with -1], in which the light is emitted at a first power level, and a second state [e.g. second power state] [e.g. indicated with +1], in which the light is emitted at a second power level different from the first power level, wherein the plurality of emitters are partitioned into a plurality of groups [or modulation units] [in some examples, all modulation units (groups of emitters) may have the same (identical) power, or are so that they cause identical power differential between the two output states], so that all the emitters of the same group have [e.g. select] the same state [e.g. for at least one time slot], [e.g. the emitters may be modulated using a OOK modulation] [in some examples, one of the first state and the second state may be a 0-power state (e.g. in which no light is actually emitted), and the other one may be a non-0-power state (e.g. in which light is actually emitted, e.g. in a particular wavelength or in a particular range of wavelengths); this may be in accordance with the OOK modulation] [in some examples, all the emitters have the same power] [in examples, all emitters of each group have the same selected state]; wherein the method includes: receiving a sequence of input values [e.g. in binary, digital form, each input value having, for example, a particular number of possible states (e.g. a bit can have two states, a byte can have 256 states, etc.)] to be transmitted by the plurality of emitters; mapping the sequence of input values onto one or more pulse amplitude modulation, PAM, symbols, each PAM symbol having a power level [e.g. the power level being associated with the sum or composition of the light waves emitted by the emitters according to their state; for example, in case of two groups of emitters and each emitter having a state selected from two states (-1, +1), the PAM symbol may have a power level which is one of three power levels (indicated with -1, 0, +1), i.e. a first power level indicated with -1 (as resulting from the composition of -Is from both the groups), a second power level (+1, as resulting from the composition of +ls from both the groups), and an intermediate power level 0 (as resulting from the composition between the first state (-1) of the first group and the second state (+1) of the second group)]; and performing a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol [e.g. so that the cumulative power level emitted by the plurality of groups (e.g. in one time slot) represents the power level of the PAM symbol that has been generated in the by the symbol mapper] [and / or so that the light emitted by the plurality of emitters represents, encoded therein according to the PAM modulation, the input values of the sequence of input values] [e.g., the emitter selector selects the groups of emitters in such a way that the PAM symbols are converted onto OOK symbols] [as shown below, in some examples different selections of groups can be performed in different configuration modes] [e.g. the M-PAM may have an M chosen in real time, e.g. after having knowledge of the quality of the channel].

[0203] According to an aspect, there is provided a non-transitory storage unit storing instruction which, when executed by a processor, cause the processor to control an optical transmission [e.g. through method above, or through the equipment of any of the preceding aspects] by mapping a sequence of input values onto one or more pulse amplitude modulation, PAM, symbols [e.g. M-PAM symbols], each PAM symbol having a power level; and performing a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol.

[0204] According to an aspect, a non-transitory storage unit storing instruction which, when executed by a processor, cause the processor to control the above-mentioned method.

[0205] Other examples

[0206] Here, different inventive examples, embodiments and aspects are described. Also, further embodiments will be defined by the enclosed claims. It should be noted that any embodiments as defined by the claims can be supplemented by any of the details (features and functionalities) described in the following chapters. Also, the embodiments described in the following and / or previous chapters can be used individually, and can also be supplemented by any of the features in another chapter, or by any feature included in the claims. Also, it should be noted that individual aspects described herein can be used individually or in combination. Thus, details can be added to each of said individual aspects without adding details to another one of said aspects. It should also be noted that the present disclosure describes, explicitly or implicitly, features of a mobile communication device and of a receiver and of a mobile communication system. Thus, any of the features described herein can be used in the context of a mobile communication device and in the context of a mobile communication system (e.g. comprising a satellite). Therefore, disclosed techniques are suitable for all fixed satellite services (FSS) and mobile satellite services (MSS). Moreover, features and functionalities disclosed herein relating to a method can also be used in an apparatus. Furthermore, any features and functionalities disclosed herein with respect to an apparatus can also be used in a corresponding method. In other words, the methods disclosed herein can be supplemented by any of the features and functionalities described with respect to the apparatuses. Also, any of the features and functionalities described herein can be implemented in hardware or in software, or using a combination of hardware and software, as described below. Depending on certain implementation requirements, examples may be implemented in hardware. The implementation may be performed using a digital storage medium, for example a floppy disk, a Digital Versatile Disc (DVD), a Blu-Ray (registered trademark) Disc, a Compact Disc (CD), a Read-only Memory (ROM), a Programmable Read-only Memory (PROM), an Erasable and Programmable Read-only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM) or a flash memory, having electronically readable control signals stored thereon, which cooperate (or are capable of cooperating) with a programmable computer system such that the respective method is performed. Therefore, the digital storage medium may be computer readable. Generally, examples may be implemented as a computer program product with program instructions, the program instructions being operative for performing one of the methods when the computer program product runs on a computer. The program instructions may for example be stored on a machine-readable medium. Other examples comprise the computer program for performing one of the methods described herein, stored on a machine-readable carrier. In other words, an example of method is, therefore, a computer program having a program-instructions for performing one of the methods described herein, when the computer program runs on a computer. A further example of the methods is, therefore, a data carrier medium (or a digital storage medium, or a computer-readable medium) comprising, recorded thereon, the computer program for performing one of the methods described herein. The data carrier medium, the digital storage medium or the recorded medium are tangible and / or non-transitionary, rather than signals which are intangible and transitory. A further example comprises a processing unit, for example a computer, or a programmable logic device performing one of the methods described herein. A further example comprises a computer having installed thereon the computer program for performing one of the methods described herein. A further example comprises an apparatus or a system transferring (for example, electronically or optically) a computer program for performing one of the methods described herein to a receiver. The receiver may, for example, be a computer, a mobile device, a memory device or the like. The apparatus or system may, for example, comprise a file server for transferring the computer program to the receiver. In some examples, a programmable logic device (for example, a field programmable gate array) may be used to perform some or all of the functionalities of the methods described herein. In some examples, a field programmable gate array may cooperate with a microprocessor in order to perform one of the methods described herein. Generally, the methods may be performed by any appropriate hardware apparatus. The above described examples are illustrative for the principles discussed above. It is understood that modifications and variations of the arrangements and the details described herein will be apparent. It is the intent, therefore, to be limited by the scope of the impending patent claims and not by the specific details presented by way of description and explanation of the examples herein.

Claims

Claims1. An optical transmitter, comprising: at least a plurality of emitters which are optical emitters, each emitter being configured to emit light according to a state selected between at least a first state, in which the light is emitted at a first power level, and a second state, in which the light is emitted at a second power level different from the first power level, wherein the plurality of emitters are partitioned into a plurality of groups, so that all the emitters of the same group have the same state; an input interface (804), configured to receive a sequence of input values to be transmitted by the plurality of emitters; a symbol mapper (806), to map the sequence of input values onto one or more pulse amplitude modulation, PAM, symbols, each PAM symbol having a power level; and an emitter selector (810), to perform a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol.

2. The optical transmitter of claim 1, wherein the symbol mapper is configured to generate the M-PAM symbols according to an M-PAM modulation.

3. The optical transmitter of any of the preceding claims, configured to assign one or more PAM symbols to a number of bits of at least one input value of the sequence of input values, and the number of states that the number of bits of the at least one input value shall be less or equal to the number of states that the one or more PAM symbols can have.

4. The optical transmitter of any of the preceding claims, wherein the symbol mapper acquires the number A of groups of emitters connected, and to define the M-PAM modulation according to the relationship M = A + 1 after that A is acquired.

5. The optical transmitter of any of the preceding claims, wherein the emitter selector is configured to perform the selection by choosing among a plurality of possible group combinations which potentially reach the power level of the PAM symbol.

6. The optical transmitter of claim 5, wherein the emitter selector is configured, when performing the selection according to a second PAM symbol in a second time slot which immediately follows a first time slot in which the selection according to a first PAM symbol according to a firstgroup combination has been performed, to choose a second group combination which maximizes the number of transitions from the first state to the second state and / or from the second state to the first state.

7. The transmitter of any of the preceding claims, wherein the emitter selector is configured, when performing the selection according to a second PAM symbol in a second time slot which immediately follows a first time slot in which the selection according to a first PAM symbol has been performed, to evaluate whether the transition from the first PAM symbol to the second PAM symbol implies a required number T of groups transitioning from a state to another one which verifies 0<T<N-l with N being the number of groups, to cause, beyond the T transitionings, at least one first additional group to additionally and redundantly transition from the first state to the second state, and at least one second additional group to additionally and redundantly transition from second state to the first state.

8. The optical transmitter of any of the preceding claims, wherein the emitter selector is configured to select states of different groups randomly.

9. The optical transmitter of any of the preceding claims, configured to select between at least one first configuration mode and at least one second configuration mode, wherein: in the first configuration mode, the plurality of emitters are partitioned into a first plurality of N groups, and in the second configuration mode, the plurality of emitters are partitioned into a second plurality of N’ groups with N’<N.

10. The optical transmitter of claim 9, wherein: in the first configuration mode, each group of the first plurality of N groups is capable of emitting a first highest-power PAM signal at a first highest-power emitting power, and in the second configuration mode, each group of the second plurality of N’ groups is capable of emitting a second highest-power PAM signal at a second highest-power emitting power which is larger than the first emitting power.

11. The optical transmitter of claim 9 or 10, wherein: in the first configuration mode, each group of the first plurality of N groups has a first numberof emitters, and in the second configuration mode, each group of the second plurality of N’ groups has a second number of emitters greater than in the first configuration mode.

12. The optical transmitter of any of claims 9-11, configured to change from the first configuration mode to the second configuration mode in case it is determined that the channel has a quality worse than a predetermined threshold; and / or configured to change from the second configuration mode to the first configuration mode in case it is determined that the channel has a quality better than a predetermined threshold.

13. The optical transmitter of any of the preceding claims, configured, in case it is determined that the channel has a quality worse than a predetermined threshold, to reduce the number of the groups for each group and / or in case it is determined that the channel has a quality better than a predetermined threshold, to increase the number of the groups for each group.

14. The optical transmitter of any of the preceding claims, configured to select between at least one full-power mode and at least one second, power-saving mode, wherein, in the power-saving configuration mode, each group has at least one emitter which is deactivated, and, in the full-power mode, the at least one emitter is reactivated.

15. The optical transmitter of any of the preceding claims, wherein the optical emitters are VCSEL emitters or sub-arrays of VCSEL emitters.

16. The optical transmitter of any of the preceding claims, wherein the emitters are collected in subarrays, all the emitters of one subarrays being selectable to have the same state.

17. The optical transmitter of claim 16, wherein each group is defined using a discrete number of subarrays.

18. The optical transmitter of any of the preceding claims, wherein the emitter selector is configured to define each group of emitters in such a way that light emitted by different groups overlap, so as to reach the power level of the PAM symbol by cumulative power level from all thegroups.

19. A method for performing an optical transmission, through at least a plurality of emitters which are preferably optical emitters, each emitter being configured to emit light according to a state selected between at least a first state, in which the light is emitted at a first power level, and a second state, in which the light is emitted at a second power level different from the first power level, wherein the plurality of emitters are partitioned into a plurality of groups, so that all the emitters of the same group have the same state; wherein the method includes: receiving a sequence of input values to be transmitted by the plurality of emitters; mapping the sequence of input values onto one or more pulse amplitude modulation, PAM, symbols, each PAM symbol having a power level; and performing a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol.

20. A non-transitory storage unit storing instruction which, when executed by a processor, cause the processor to control an optical transmission by mapping a sequence of input values onto one or more pulse amplitude modulation, PAM, symbols, each PAM symbol having a power level; and performing a selection, for each group of emitters, between the first state and the second state so as to reach the power level of each PAM symbol.

Images