METHOD, TRANSMITTER, STRUCTURE, TRANSCEIVER AND ACCESS POINT FOR PROVISION OF MULTI-CARRIER ON-OFF KEYING SIGNAL

Abstract

A method of transmitting an On-Off Keying, OOK, signal which includes an ON waveform and an OFF waveform forming a pattern representing transmitted information. The method includes obtaining a basic baseband waveform; scrambling the basic baseband waveform by applying a first binary randomised sequence where one of the binary values cause transformation to a complex conjugate; modulating the information to be transmitted by applying the scrambled basic baseband waveform for the ON waveform and applying no waveform for the OFF waveform; and transmitting the modulated information.

Claims

1. A method of transmitting an On-Off Keying, OOK, signal which comprises an ON waveform and an OFF waveform forming a pattern representing transmitted information, the method comprising: obtaining a basic baseband waveform; scrambling the basic baseband waveform by applying a first binary randomised sequence where at least one of the binary values of the first binary randomised sequence causes transformation of the basic baseband waveform to a complex conjugate of the basic baseband waveform; modulating the information to be transmitted by applying the scrambled basic baseband waveform for the ON waveform and applying no waveform for the OFF waveform; and transmitting the modulated information.

2. The method of claim 1, wherein the obtaining of the basic baseband waveform comprises generating an Orthogonal Frequency Division Multiplexing signal mimicking a desired baseband waveform.

3. The method of claim 2, wherein the desired baseband waveform corresponds to a multicarrier on-off keying, MC-OOK, symbol.

4. The method of any one of items 1 to 3 claim 1, wherein the scrambling of the basic baseband waveform further comprises applying a second binary randomised sequence where binary values apply phase rotations which are mutually separated by π.

5. The method of claim 4, wherein the first randomised sequence is generated in a shift register mechanism representing a first polynomial and the second randomised sequence is generated in a shift register mechanism representing a second polynomial different from the first polynomial.

6. The method of claim 5, wherein the shift register mechanism uses a single shift register for the generation of both the first and the second binary randomised sequences, where the first binary randomised sequence is tapped at a first position of the single shift register and the second binary randomised sequence is tapped at a second position of the single shift register, and the first and second positions of the single shift register are different.

7. A transmitter for transmitting an On-Off Keying, OOK, signal which comprises an ON waveform and an OFF waveform forming a pattern representing transmitted information, the transmitter comprising: a basic waveform input configured to obtain a basic baseband waveform; a scrambler configured to scramble the basic baseband waveform by applying a first binary randomised sequence where at least one of the binary values of the first binary randomised sequence causes transformation of the basic baseband waveform to a complex conjugate of the basis baseband waveform; a modulator configured to modulate the information to be transmitted by applying the scrambled basic baseband waveform for the ON waveform and applying no waveform for the OFF waveform; and a transmitter circuit configured to transmit the modulated information.

8. The transmitter of claim 7, comprising a basic baseband waveform generator, wherein the basic baseband waveform generator is arranged to generate the basic baseband waveform as an Orthogonal Frequency Division Multiplex signal mimicking a desired baseband waveform, and is arranged to provide the basic baseband waveform to the basic waveform input.

9. The transmitter of claim 8, wherein the desired baseband waveform corresponds to a multicarrier on-off keying, MC-OOK, symbol.

10. The transmitter of claim 7, wherein the scrambler is arranged to apply a second binary randomised sequence where binary values apply phase rotations which are mutually separated by π.

11. The transmitter of claim 10, wherein the first randomised sequence is generated in a shift register mechanism representing a first polynomial and the second randomised sequence is generated in a shift register mechanism representing a second polynomial different from the first polynomial.

12. The transmitter of claim 11, further comprising a shift register, wherein the shift register mechanism uses the shift register for the generation of both the first and the second binary randomised sequences, where the first binary randomised sequence is tapped at a first position of the shift register and the second binary randomised sequence is tapped at a second position of the shift register, and the first and second positions of the shift register are different.

13. (canceled)

14. A structure for generating sequences, the structure comprising: a binary shift register; a feedback structure connected to the shift register arranged to define a linear feedback shift register according to a polynomial; a first output configured to collect one or more state values from a first group of elements of the shift register, the one or more state values from the first group forming a value of a first sequence; a second output configured to collect one or more state values from a second group of elements of the shift register, the one or more state values from the second group forming a value of a second sequence, no element of the second group belonging to the first group; and a third output configured to apply a binary randomised sequence to the second sequence to cause transformation of the second sequence to a complex conjugate of the second sequence.

15. The structure of claim 14, wherein the second output is configured to collect state values from the second group of elements, the second group comprising a plurality of elements of the shift register such that the second sequence comprises symbols having more than two possible values.

16. The structure of claim 14, wherein the second sequence is a binary sequence.

17. The structure of claim 16, wherein the second output is configured to collect state values from the second group of elements, where the second group comprises a single element of the shift register.

18. The structure of claim 14, wherein the first output is configured to collect state values from the first group of elements comprising a plurality of elements of the shift register such that the first sequence comprises symbols having more than two possible values.

19. The structure of claim 14, wherein the first sequence is a binary sequence.

20. The structure of claim 19, wherein the first output is configured to collect state values from the first group of elements, where the first group comprises a single element of the shift register.

21. A transceiver comprising: a transmitter, the transmitted configured to transmit an On-Off Keying, OOK, signal which comprises an ON waveform and an OFF waveform forming a pattern representing transmitted information, the transmitter comprising: a basic waveform input configured to obtain a basic baseband waveform; a scrambler configured to scramble the basic baseband waveform by applying a first binary randomised sequence where at least one of the binary values of the first binary randomised sequence causes transformation of the basic baseband waveform to a complex conjugate of the basis baseband waveform; a modulator configured to modulate the information to be transmitted by applying the scrambled basic baseband waveform for the ON waveform and applying no waveform for the OFF waveform; and a transmitter circuit configured to transmit the modulated information; and a structure, the structure being configured to provide the first and second binary randomized sequences for the transmitter, the structure comprising: a binary shift register; a feedback structure connected to the shift register arranged to define a linear feedback shift register according to a polynomial; a first output configured to collect one or more state values from a first group of elements of the shift register, the one or more state values from the first group forming a value of a first sequence; a second output configured to collect one or more state values from a second group of elements of the shift register, the one or more state values from the second group forming a value of a second sequence, no element of the second group belonging to the first group; and a third output configured to apply a binary randomised sequence to the second sequence to cause transformation of the second sequence to a complex conjugate of the second sequence.

22. The transmitter according to claim 7, wherein the transmitter is comprised in an access point and causes the access point to transmit a wake-up packet using multicarrier on-off keying.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The above, as well as additional objects, features and advantages of the present disclosure, will be better understood through the following illustrative and non-limiting detailed description of preferred embodiments of the present disclosure, with reference to the appended drawings.

[0035] FIG. 1 schematically illustrates a receiver having a traditional WUR and PCR structure.

[0036] FIG. 2 schematically illustrates a traditional OOK structure.

[0037] FIG. 3 schematically illustrates a structure for generating a basic baseband waveform using IFFT.

[0038] FIG. 4 is a signal diagram illustrating power spectral density of a basic baseband waveform generated by a structure according to FIG. 3.

[0039] FIG. 5 schematically illustrates a structure for a phase randomisation technique to smooth power spectral density spikes of a signal as of FIG. 4.

[0040] FIG. 6 is a signal diagram illustrating power spectral density of a smoothened waveform by the structure of FIG. 5.

[0041] FIG. 7 schematically illustrate a structure for flattening power spectral density of a waveform according to an embodiment.

[0042] FIG. 8 illustrates signal diagrams of power spectral densities of a signal (Cf. FIG. 6) and a complex conjugate of the signal.

[0043] FIG. 9 illustrates an alternative structure for flattening power spectral density of a waveform according to an embodiment.

[0044] FIG. 10 is a signal diagram illustrating a flattened power spectral density using an embodiment.

[0045] FIG. 11 schematically illustrates a transmitter according to an embodiment.

[0046] FIG. 12 schematically illustrates a transmitter according to an embodiment.

[0047] FIG. 13 schematically illustrates a transmitter according to an embodiment.

[0048] FIG. 14 schematically illustrates a linear feedback shift register for generating a first sequence and an addon tap to extract a second sequence according to an embodiment.

[0049] FIG. 15 schematically illustrates a linear feedback shift register for generating a first sequence and an addon tap mechanism to extract a second sequence according to an example.

[0050] FIG. 16 schematically illustrates a linear feedback shift register for generating a first sequence and an addon tap mechanism to extract a second sequence according to an embodiment.

[0051] FIG. 17 schematically illustrates a transmitter according to an embodiment.

[0052] FIG. 18 is a signal diagram illustrating power spectral density of a basic baseband waveform generated when applying a structure as of FIG. 15.

[0053] FIG. 19 is a flow chart illustrating a method according to an embodiment.

[0054] FIG. 20 is a block diagram schematically illustrating a network node according to an embodiment.

[0055] FIG. 21 schematically illustrates a computer-readable medium and a processing device.

DETAILED DESCRIPTION

[0056] FIG. 7 schematically illustrate a structure for flattening power spectral density of a waveform according to an embodiment. A binary bit sequence with proper randomisation is provided, and for each bit value b1 a baseband waveform is substituted by its complex conjugate for one of the states of the bit value and kept unchanged for the other state of the bit value. The randomised complex conjugation will flatten the PSD of the output of the structure.

[0057] FIG. 8 illustrates signal diagrams of power spectral densities of a signal (Cf. FIG. 6) and a complex conjugate of the signal. By the randomised substitution with the complex conjugate version of the basic baseband waveform, an alternative OFDM signal is randomly provided, with spectral content as illustrated to the right in FIG. 8. The alternative OFDM signal provides the same envelope as the basic baseband waveform, but with a different spectral content. This is due to that the waveform obtained by complex conjugating a time domain OFDM waveform can also be generated by transforming to the time domain a frequency domain signal comprising frequency domain symbols that are the complex conjugates of the frequency domain symbols of the original signal, and reversing the order of the subcarriers. For example, if an OFDM signal is generated from complex-valued frequency domain symbols X.sub.k, where k=−M, . . . , M, by means of an IFFT, then the complex conjugate of said OFDM signal can be generated by applying an inverse discrete Fourier transform to the frequency domain symbols X*.sub.−k, where the star * represents complex conjugation and the minus sign in the index k indicates reversal of the order of the subcarriers. The randomised provision of the variants, i.e. sometimes with spectral content as illustrated to the left in FIG. 8 and sometimes with spectral content as illustrated to the right in FIG. 8, provides for the flatter waveform in average, which will be demonstrated with reference to FIG. 10 below illustrating a simulated result of one embodiment of the approach.

[0058] FIG. 9 illustrates an alternative structure for flattening power spectral density of a waveform according to an embodiment. Here, the spectral line suppression feature demonstrated with FIG. 5 is applied together with an approach similar to the one demonstrated with reference to FIG. 7. This structure will thus do spectral line suppression and PSD flattening.

[0059] FIG. 10 is a signal diagram illustrating a flattened power spectral density using an embodiment. This is for example a result achieved when applying the structure demonstrated with reference to FIG. 9 on a basic baseband signal generated by a generator structure as demonstrated with reference to FIG. 3. The PSD is fairly flat and free from spectral lines, and thus provides good performance for use in an OOK provision structure.

[0060] Referring back to the discussion in the background section about the limitations in output power, a discussion about the benefits of the flattened PSD illustrated by the diagram of FIG. 10 will now be given. Subject to the PSD limits in for example Europe, a WUP having a PSD as in the diagram of FIG. 10 would have a total output power of 35 mW, given the same other features as of the example in the background section. Note that although the same basic baseband On waveform is used to generate FIG. 6 and FIG. 10, the output power in PSD limited regulatory domains is 1 dB larger if the transmitter is implemented according to the approach providing the PSD as of FIG. 10 than according to traditional techniques providing the PSD as of FIG. 6.

[0061] FIG. 11 schematically illustrates a transmitter according to an embodiment. In brief, the transmitter is arranged for OOK similar to the structure which has been demonstrated with reference to FIG. 2 but with a PSD flattening structure similar to the structure which has been demonstrated with reference to FIG. 7. The ON waveform generator (WG) providing the waveform to the PSD flattening structure may be similar to the generator demonstrated with reference to FIG. 3.

[0062] FIG. 12 schematically illustrates a transmitter according to an embodiment. In brief, the transmitter is arranged for OOK similar to the structure which has been demonstrated with reference to FIG. 2 but with a PSD flattening structure similar to the structure which has been demonstrated with reference to FIG. 7 and a spectral line suppression structure similar to the one which has been demonstrated with reference to FIG. 5. The ON waveform generator (WG) providing the waveform to the PSD flattening structure may be similar to the generator demonstrated with reference to FIG. 3.

[0063] FIG. 13 schematically illustrates a transmitter according to an embodiment. In brief, the transmitter has a similar structure as the one demonstrated with reference to FIG. 12, but with the spectral line suppression structure similar to the one which has been demonstrated with reference to FIG. 5 connected to the waveform generator and then the PSD flattening structure similar to the structure which has been demonstrated with reference to FIG. 7 provided between the spectral line suppression structure and the OOK structure.

[0064] The bit sequences provided to the PSD flattening structure for providing a randomised application of the complex conjugate may be provided in a variety of ways. One way is to use a pseudorandom sequence generator based on a linear feedback shift register. Another way is to collect a sequence from a look-up table. Below, with reference to FIGS. 14 to 16, there are demonstrated approaches for achieving multiple sequences from a single shift register structure. The multiple sequences may be desired for example for the structures demonstrated with reference to FIGS. 12 and 13 where the PSD flattening structure demands one sequence and the spectral line suppression structure demands one sequence. For the sake of not risk causing new kinds of spurs in the signal for the OOK structure, it is desired to have separate sequences in those cases, which sequences have limited mutual correlations. The approaches demonstrated with reference to FIGS. 14 to 16 have the advantage of keeping implementation complexity low.

[0065] An alternative way of flattening a signal as discussed above is taught in international application PCT/EP2018/066984, which is here incorporated by reference in its entirety. That approach comprises transmitting a first on-off keyed signal corresponding to the data symbols, the first signal comprising a plurality of on periods and a plurality of off periods. Each on period comprises a first signal portion cyclically shifted within the on period by a respective random or pseudorandom factor. The cyclic shifting of the first signal portion may be performed within the on period. For example, the first signal portion may be shifted in the on period by a factor such as a delay or percentage, and any part of the first signal that is shifted outside of the on period may be reintroduced into the on period at the opposite end of the on period. In this way, for example, the on period may in some examples remain filled with a signal formed from the first signal portion. In some examples, therefore, the first signal may have a flatter frequency response than other signals. In an example, Manchester coding may be applied to the data part of a wake up packet (WUP). For example, a logical “0” is encoded as “10” and a logical “1” as “01”. Therefore, every data symbol comprises an “ON” part (where there is energy) and an “OFF” part, where there is no energy, wherein the order of these parts is dependent on the data symbol. In addition, the WUP may be generated in some examples by means of an inverse fast Fourier transform (IFFT), as this block may already be available in some transmitters such as for example Wi-Fi transmitters supporting e.g. IEEE 802.11a/g/n/ac. An example approach for generating the OOK signal representing a WUP is to use the 13 sub-carriers in the centre of an OFDM multi-carrier signal, and populating these 13 sub-carriers with a signal to represent ON and to not transmit anything at all to represent OFF, similar as demonstrated with reference to FIG. 3. This may be referred to as multicarrier OOK (MC-OOK). In one example, the IFFT has 64 points and is operating at a sampling rate of 20 MHz, and just as for ordinary orthogonal frequency division multiplexing (OFDM) a cyclic prefix (CP) is added after the IFFT operation in order to have the OFDM symbol duration as being used in 802.11a/g/n/ac. In some examples of MC-OOK for a WUP, the same OFDM symbol is used. In other words, the same frequency domain symbols are used to populate the non-zero subcarriers for all data symbols. Using the same OFDM symbol to generate the “ON” part of every Manchester coded data symbol may result in strong periodic time correlations in the data part of the WUP. These correlations give rise to spectral lines, which are spikes in the Power Spectral Density (PSD) of the WUP. These spectral lines may in some examples be undesirable because there may be local geographic regulations that limit the power that can be transmitted in narrow portions of the spectrum.

[0066] In a first example embodiment, a signal is transmitted from a single antenna. Suppose that the data part of the WUP consists of a number N of OFDM symbols. This example embodiment consists of the following steps:

[0067] 1. Determine a set of K delays, K≥2. These are {T.sub.1.sup.CS, . . . , T.sub.k.sup.CS}.

[0068] 2. Generate a random or pseudorandom sequence consisting of N integers taking values between 1 and K. These are {m.sub.1, . . . , m.sub.N}.

[0069] 3. Apply a random or pseudorandom cyclic shift to each of the OFDM symbols corresponding to the “ON” parts of the data symbols, wherein the cyclic shift corresponds to one of the N integers in the sequence. For example, apply the delay T.sub.m.sub.n.sup.CS (a negative value) to the OFDM symbol corresponding to the “ON” part of the n-th data symbol. That is, if s(t), 0≤t<T.sub.s is the time domain signal corresponding to the “ON” part, having a duration T.sub.S, then the cyclic shift s.sub.CS(t;T.sub.m.sub.n.sup.CS) of s(t) by the delay T.sub.m.sub.n.sup.CS≤0 is generated by setting:

[00001]$s_{CS}\left(t;T_{m_n}^{\mathrm{CS}}\right)=\{\begin{array}{ccc}s\left(t-T_{m_n}^{\mathrm{CS}}\right)& \mathrm{if}& 0\le t<T_s+T_{m_n}^{\mathrm{CS}}\\ s\left(t-T_{m_n}^{CS}-T_s\right)& \mathrm{if}& T_{m_n}^{CS}+T_s\le t<T_s\end{array}$

[0070] 4. Transmit the MC-OOK signal, comprising the cyclically shifted OFDM symbol s.sub.CS(t;T.sub.m.sub.n.sup.CS) in in the “ON” part of the n-th data symbol.

[0071] In one particular example, T.sub.s=.sup.4 μs. A set of K=8 cyclic shifts {T.sub.1.sup.CS, . . . , T.sub.8.sup.CS} is defined as shown in the table below.

TABLE-US-00001 T.sub.1.sup.CS −0 ns T.sub.2.sup.CS −400 ns T.sub.3.sup.CS −800 ns T.sub.4.sup.CS −1200 ns T.sub.5.sup.CS −1600 ns T.sub.6.sup.CS −2000 ns T.sub.7.sup.CS −2400 ns T.sub.8.sup.CS −2800 ns

[0072] In another particular example, T.sub.s=2 μs. A set of K=8 cyclic shifts {T.sub.1.sup.CS, . . . , T.sub.8.sup.CS} is defined as shown in the table below.

TABLE-US-00002 T.sub.1.sup.CS −0 ns T.sub.2.sup.CS −400 ns T.sub.3.sup.CS −600 ns T.sub.4.sup.CS −800 ns T.sub.5.sup.CS −1000 ns T.sub.6.sup.CS −1200 ns T.sub.7.sup.CS −1400 ns T.sub.8.sup.CS −1800 ns

[0073] A sequence of random or pseudorandom integers having values between 1 and 8 is generated for each data symbol, and a cyclic shift by the corresponding delay is applied to the “ON” part of the signal for each data symbol. For example, if T.sub.s=2 μs and the integer m generated for the n-th data symbol is 6, then a cyclic shift of T.sub.6.sup.CS=1200 ns is applied to the “ON” part of the n-th transmitted data symbol.

[0074] A suitable approach for generating pseudorandom sequence generation is desired for this solution as well for the approach demonstrated with reference to FIGS. 1 to 13. As an example, consider the case where K is a power of 2, i.e. K=2.sup.P. The 802.11 standard utilizes the linear feedback shift register with generator polynomial z.sup.−7+z.sup.−4+1 to generate pseudorandom bit sequences. Any of these sequences can be used, by grouping the output in groups of p bits. Any such group can be mapped to an integer between 1 and K.

[0075] Another example embodiment involves transmission from multiple antennas (e.g. transmit diversity or spatial diversity). For each of the antennas, an MC-OOK signal is generated from data symbols according to any given multi-antenna transmit (TX) diversity technique. Then, the embodiment given for a single transmit antenna can be applied to a signal to be transmitted from each antenna. The TX diversity technique applied to the signals from the antennas may comprise delay diversity (e.g. as used in the GSM cellular system) or cyclic delay diversity (e.g. as used in the LTE cellular system).

[0076] In an example, suppose that there are L transmit antennas, MC-OOK is used, and CSD is the TX diversity technique employed by the transmitter. In this case, cyclic delays Δ.sub.l, l=1, . . . , L are applied to the OFDM symbol s(t). Thus, the signal transmitted through the 1-th antenna is s.sup.l(t)=s.sub.CS(t;Δ.sub.l), where s.sub.CS(t;Δ.sub.l) denotes the cyclic shift of s(t) by Δ.sub.l and is defined as given above for the single-antenna example. This example embodiment consists of the following steps:

[0077] 1. Determine a set of K delays, K≥2. These are {T.sub.1.sup.CS, . . . , T.sub.k.sup.CS}.

[0078] 2. Generate a random or pseudorandom sequence consisting of N integers taking values between 1 and K These are {m.sub.1, . . . , m.sub.N}.

[0079] 3. For each of the L antennas, apply the delay T.sub.m.sub.n.sup.CS (a negative value) to the OFDM symbol corresponding to the “ON” part of the n-th data symbol. That is, if s.sup.l(t), 0≤T.sub.s is the time domain signal corresponding to the “ON” part, then for the l_th antenna, the cyclic shift s.sub.CS.sup.l(t;T.sub.m.sub.n.sup.CS) of s.sup.l(t) is generated by applying a cyclic delay by T.sub.m.sub.n.sup.CS. Note the delay T.sub.m.sub.n.sup.CS may change from one data symbol to the next.

[0080] 4. Transmit the MC-OOK signal, comprising the cyclically shifted OFDM symbol s.sub.CS.sup.l(t;T.sub.m.sub.n.sup.CS) in the “ON” part of the n-th data symbol in the signal transmitted through the l-th antenna.

[0081] As an example, if CSD is used, then:

[00002]$s_{CS}^l\left(t;T_{m_n}^{\mathrm{CS}}\right)=s\left(t;{\Delta }_l+T_{m_n}^{CS}\right)=\{\begin{array}{ccc}s\left(t-{\Delta }_l-T_{m_n}^{CS}\right)& \mathrm{if}& 0\le t<{\Delta }_l+T_s+T_{m_n}^{CS}\\ s\left(t-{\Delta }_l-T_{m_n}^{\mathrm{CS}}-T_s\right)& \mathrm{if}& {\Delta }_l+T_{m_n}^{\mathrm{CS}}+T_s\le t<T_s\end{array}$

[0082] Cyclic shift symbol randomization suppresses spectral lines and flattens the spectrum. In an example where T.sub.sym=4 μs and there are 8 possible syclic shifts, by 0 ns, 400 ns, 800 ns, 1200 ns, 1600 ns, 2000 ns, 2400 ns and 2800 ns.

[0083] A slight drawback of the cyclic shift symbol randomization technique is that it can't eliminate spectral lines arising from the DC component in the On waveform. A cyclic shift applied to an OFDM signal can be implemented by a rotation of the frequency domain symbols. Thus, when applied to OFDM waveforms, cyclic shift randomization can be thought of as randomization of the phases of the subcarriers. However, the rotation applied to the DC subcarrier by any cyclic shift is zero, and hence the phase of the DC subcarrier can't be randomized by means of cyclic shift randomization. A practical solution to this drawback may be to use waveforms without a DC component as On waveforms. This can be achieved by nulling or blanking the DC subcarrier of an OFDM waveform. However, there might be circumstances where having a non-null DC subcarrier is desirable, for example to have more degrees of freedom to optimize the On waveform for performance or for other metric.

[0084] Symbol randomization techniques that suppress spectral lines as demonstrated with reference to FIG. 5 combined with cyclic shift randomization as demonstrated above provides for a low complexity technique to suppress spectral lines and flatten the spectrum. A structure for achieving this is illustrated in FIG. 17, where an example of a low-complexity sequence generation as will be demonstrated below is applied.

[0085] A well-known approach for generating pseudorandom sequences are the above-mentioned linear feedback shift register using a proper polynomial. Considering the approaches demonstrated herein for removal of spectral lines and flattening of spectral properties of a signal, there is a desire for an efficient and low-resource consuming solution for producing two or more sequences. Here, the two or more sequences are preferably having limited correlation not to risk introducing new undesired spurs in the signal. A straightforward solution is to have one generation mechanism for each sequence to generate, and to carefully select e.g. structure and polynomials of the respective generation mechanism to provide limited correlation. However, in this disclosure it is suggested an approach for generating two or more sequences from a single shift register structure where register elements and their states are reused for the different sequences. A basic sequence generated by the structure will have the same properties as of a linear feedback shift register. The additional generated sequences will not have the same characteristics but will have low enough correlation for the purposes of the signal shaping approaches of this disclosure and will also have sufficient performance for other applications where multiple sequences with low correlation is desired.

[0086] An approach according to this disclosure is implemented in a transmitting network node, such as an access point, AP. An embodiment is illustrated in FIG. 17. The LFSR is updated every T.sub.sym.

[0087] FIG. 14 schematically illustrates a linear feedback shift register for generating a first sequence and an addon tap to extract a second sequence according to an embodiment. The LFSR in this figure has generator polynomial x.sup.−7+x.sup.−4+1, but other generator polynomials can be used. Referring back to the Figs illustrating OOK, The LFSR is updated every symbol time, T.sub.sym, and the bits b0 and b1 are read from different states of the LFSR. For example b0 may be extracted from the first position in the register, labelled X.sup.1 in FIG. 14, while b1 may be extracted from the seventh position in the register, labelled X.sup.7 in FIG. 14.

[0088] FIG. 15 schematically illustrates a structure comprising a Linear Feedback Shift Register (LFSR) with generator polynomial x.sup.−7+x.sup.−4+1 where the structure is used to generate a pseudo-random bit sequence, but other polynomials could be used. The register contains seven elements labelled X.sup.1 to X.sup.7. The bits b5 to b7 are extracted from the elements 5 to 7 of the register. Moreover, the LFSR is updated every T.sub.sym. Note that both phase randomization and cyclic shift randomization require a source of randomness, in order to generate random phase shifts and random cyclic shifts. Ideally, independent sources of randomness should be used. However, for ease of implementation the same LFSR is used to generate both the sequences.

[0089] A problem with the combined solution is that there is a strong correlation between the source of randomness used for the phase randomizer (i.e. b7) and the source of randomness for the cyclic shift randomizer (i.e. b5, b6, b7). In a combined flattening and spectral line suppression structure including e.g. a complex conjugation structure and a phase shifter, or a cyclic shifter and a phase shifter, this may cause remaining spectral lines as illustrated in the diagram of FIG. 18. The reason is that in this example cyclic shift randomization imparts a random phase shift by zero or 180 degrees to two of the subcarriers, but due to the perfect correlation, the phase randomizer reverses the 180 degrees phase shift, so that as a result these two subcarriers don't have their phase randomized, resulting in the two spectral lines shown in FIG. 18. In order to eliminate spectral lines, the phases applied to each subcarrier in the On waveform preferably have zero mean. But because of the strong correlations, they may fail to do so. As an illustration, suppose that there are 8 possible cyclic shifts, by 0 ns, 400 ns, 800 ns, 1200 ns, 1600 ns, 2000 ns, 2400 ns and 2800 ns.

[0090] Hence, since a symbol randomization technique based on a combination of phase randomization and cyclic shift randomization is desirable, and since due to ease of implementation it is also desirable to use only one LFSR as source of randomness for both randomization techniques, it is sought a method to achieve symbol randomization by means of a combination of phase randomization, cyclic shift randomization and using only one LFSR. The basic idea in the present disclosure is to create two sources of entropy or randomness from the same LFSR in such a way that the two randomization techniques are sufficiently decorrelated.

[0091] FIG. 16 schematically illustrates a structure comprising a Linear Feedback Shift Register (LFSR) with generator polynomial x.sup.−7+x.sup.−4+1 where the structure is used to generate a pseudo-random bit sequence, but other polynomials could be used. The register contains seven elements labelled X.sup.1 to X.sup.7. In FIG. 17 the source of randomness for the phase randomizer, enclosed by dotted lines, is labelled b7 and is a bit stream drawn from the seventh element X.sup.7 in the register. The sources of randomness for the cyclic shift randomizer, enclosed by dashed lines, are labelled b1, b2, b3, and are three-bit streams drawn from the first, second and third elements of the register. This breaks the strong correlations between the randomness sources for the phase and cyclic shift randomizers.

[0092] The decreased correlation between the tapped sequences is achieved by choosing the source of randomness for a first sequence to depend on a first set of elements in the LFSR register, and to choose the sources of randomness for a second sequence to depend on a second set of elements of the register, such that the first and second sets are non-overlapping. The respective set may comprise one element, producing a binary sequence, or a plurality of sets, producing a higher order sequence, in any combination.

[0093] Although binary phase randomization is the simplest phase randomization technique, it is possible to use quaternary or higher order phase randomization techniques. As an illustration, in the case of quaternary phase randomization, for each occurrence of an On waveform, a randomly chosen phase of either 0, 90, 180 or 270 degrees is applied to said On waveform. Thus, it is necessary to choose randomly among 4 phases. This can be achieved by feeding bitstreams b1 and b2, drawn from elements 1 and 2 of the register, to the phase randomizer, and feeding bitstreams b5, b6, b7 drawn from elements 5, 6 and 7 of the register, to the cyclic shift randomizer. Once again, the key is that the two sets of elements of the register, namely {1,2} (used for phase randomization) and {5,6,7} (used for cyclic shift randomization) are non-overlapping.

[0094] FIG. 19 is a flow chart schematically illustrating methods of this disclosure. The method is for transmitting an On-Off Keying, OOK, signal which comprises an ON waveform and an OFF waveform forming a pattern representing transmitted information. A basic baseband waveform is obtained 1900. The obtaining 1900 of the basic baseband waveform may comprise generating an Orthogonal Frequency Division Multiplex signal mimicking a desired baseband waveform. The basic baseband waveform is scrambled 1902 by applying a first binary randomised sequence where one of the binary values cause transformation to a complex conjugate. The scrambling 1902 of the basic baseband waveform may further comprise applying a second binary randomised sequence where binary values apply phase rotations which are mutually separated by π. The first randomised sequence may be generated in a shift register mechanism representing a first polynomial and the second randomised sequence may be generated in a shift mechanism representing a second polynomial different from the first polynomial. The shift register mechanism may use a single shift register for the generation of both the first and the second binary randomised sequences, where the first binary randomised sequence is tapped at a first position of the single shift register and the second binary randomised sequence is tapped at a second position of the single shift register, and the first and second positions of the single shift register are different.

[0095] The information to be transmitted is modulated 1904 by applying the scrambled basic baseband waveform for the ON waveform and applying no waveform for the OFF waveform. The modulated information is then transmitted 1906.

[0096] FIG. 20 is a block diagram schematically illustrating a network node 2000, e.g. an access point, according to an embodiment. The network node comprises an antenna arrangement 2002, a receiver 2004 connected to the antenna arrangement 2002, a transmitter 2006 connected to the antenna arrangement 2002, a processing element 2008 which may comprise one or more circuits, one or more input interfaces 2010 and one or more output interfaces 2012. The interfaces 2010, 2012 can be operator interfaces and/or signal interfaces, e.g. electrical or optical. The network node 2000 is arranged to operate in a cellular communication network. In particular, by the processing element 2008 being arranged to perform the features demonstrated with reference to FIG. 19, the network node 2000 is capable of efficiently providing WUPs and be implemented with low complexity. The processing element 2008 can also fulfill a multitude of tasks, ranging from signal processing to enable reception and transmission since it is connected to the receiver 2004 and transmitter 2006, executing applications, controlling the interfaces 2010, 2012, etc.

[0097] The methods according to the present disclosure is suitable for implementation with aid of processing means, such as computers and/or processors, especially for the case where the processing element 2008 demonstrated above comprises a processor handling WUP provision. Therefore, there is provided computer programs, comprising instructions arranged to cause the processing means, processor, or computer to perform the steps of any of the methods according to any of the features described with reference to FIG. 19. The computer programs preferably comprise program code which is stored on a computer readable medium 2100, as illustrated in FIG. 21, which can be loaded and executed by a processing means, processor, or computer 2102 to cause it to perform the methods, respectively, according to embodiments of the present disclosure, preferably as any of the features described with reference to FIG. 19. The computer 2102 and computer program product 2100 can be arranged to execute the program code sequentially where actions of the any of the methods are performed stepwise or perform the methods on a real-time basis. The processing means, processor, or computer 2102 is preferably what normally is referred to as an embedded system. Thus, the depicted computer readable medium 2100 and computer 2102 in FIG. 21 should be construed to be for illustrative purposes only to provide understanding of the principle, and not to be construed as any direct illustration of the elements.