RU2575005C2 - Modulation for encoded light transmission - Google Patents

Abstract

FIELD: electricity.

SUBSTANCE: invention relates to the lighting control system, in particular containing multiple light sources based on use of the visible light (VL) and infrared (IR) communication for the light sources selection and improved control. The technical result is achieved in that sequences of the channel symbols for the light source control are determined from the sequences of the source symbols such, that there is no visible flicker effect in the encoded light emitted by the light sources. Each source symbol is converted in composite channel symbol containing at least one first channel symbol, that can be identical to the current source symbol, and at least one second channel symbol, that can be function of the current source symbol, and at least one next and/or previous source symbol.

EFFECT: assurance of the light coding, ie encoded information inclusion in the light output of the light source, that does not result in visible flicker effect.

18 cl, 21 dwg

Description

FIELD OF THE INVENTION

The present invention relates to a coded light system. In particular, it relates to methods and devices for controlling at least one light source configured to emit coded light in a coded light system, and for decoding such coded light into an information sequence.

BACKGROUND

Light sources are currently used in lighting systems consisting of a large number of light sources. Since the introduction of LED lighting, several parameters of these light sources can be changed and controlled in the system of light sources. Such parameters include light intensity, color of light, color temperature of light, and even the direction of light. By changing and controlling these parameters of various light sources, the light designer or system user is able to generate light scenes. This process is often referred to as scene setup, and it is usually a rather complicated process due to the large number of light sources and parameters that need to be controlled. Typically, each light source requires one controller or control channel. This makes it difficult to control a system of more than ten light sources.

To make it possible to more intuitively and easily control light sources and to create scenes, it has previously been proposed to embed information sequences, such as invisible identifiers, in the light output of lighting devices. Such embedding of identifiers can be based on the unique modulation of visible light (VL) of the lighting device or on the placement of an additional infrared (IR) light source in the lighting device and the unique modulation of this IR light. Manchester code is commonly used as a means to modulate coded light.

Embedding identifiers in the light will be called coded light (CL). A communication using CL will be called visible light communication (VLC). For CL transmission, light emitting diodes (LEDs) are generally considered that allow an acceptable high modulation frequency and bandwidth. This, in turn, may result in a quick response from the control system. Identifiers, however, can also be integrated into other light sources, such as incandescent, halogen, fluorescent, and high intensity discharge (HID) lamps. These identifiers of light sources, also called codes, provide for the identification and assessment of the intensity of individual local contributions to lighting. This can be applied in light management applications such as commissioning, light source selection, and interactive scene setup. These applications are used, for example, in homes, offices, shops and hospitals. Consequently, these light source identifiers enable a simple and intuitive operation to control the lighting system, which would otherwise be very complex.

SUMMARY OF THE INVENTION

In connection with the foregoing, it is desirable that the embedding of coded information (eg, packets) in the light output of the light source is invisible to users who are interested in the lighting function of the light source. In other words, it may be desirable that the coding of the light does not cause apparent flicker. This requirement means that the spectrum of the modulating code should not contain low frequencies. Since light sources connected to a conventional electrical network tend to produce strong interference, for example, at a constant current of 50 Hz or 100 Hz, it may be advantageous to suppress this interference by appropriate filtering in the receiver. In order not to reduce the desired signal too much by this filtering, it is advantageous that the desired signal is absent in those parts of the spectrum where interference exists. This consideration also implies that the spectrum of the modulating code should not contain low frequencies. It may also be desirable that the spectrum of the modulating code does not contain high frequencies for the reason that it may be difficult to implement circuits in low cost LED control devices for such high frequencies.

It is also desirable on the receiver side to be able to detect digital information contained in the VLC, even in the presence of interference from other light sources, such as the sun, a fluorescent lamp or tube, an incandescent lamp, etc. Therefore, it may be desirable that the spectrum of the modulated signal does not contain any contribution to the positions of the frequencies of the spectrum where such interference exists.

The purpose of the present invention is to overcome these problems and provide methods, devices and systemic approaches that solve or at least reduce the problems mentioned above; for example, by providing methods, devices, and system approaches that allow a better spectral limitation, in particular, for low frequencies than modern devices in the art, while having comparable complexity.

In General, the above objectives are achieved using methods and devices in accordance with the attached independent claims.

According to a first aspect, the above objectives are achieved with a light control device for controlling at least one light source configured to emit coded light based on a control signal, the light control device comprising: a receiver configured to receive a sequence u = [u ₁ , ..., u _k , ..., u _K ] characters u _{k of the} source, representing the information sequence of the source of information; a processing unit configured to determine from a sequence of source symbols a sequence z = [z ₁ , ..., z _k , ..., z _K ] channel symbols z _k forming a control signal; a transmitter configured to provide at least one light source with a control signal, thereby controlling at least one light source; the processing unit is configured to determine the sequence of symbols z of the channel by converting each symbol u _{k of the} source at time k into a composite symbol z _k = (z _{k, 1} , z _{k, 2} ) of the channel containing at least one first symbol z _{k, 1} channels, one of which is identical to u _k , and at least one second symbol z _{k, 2} channels selected from the set M, where z _{k, 2} is a function of the source symbol u _k and at least one subsequent and / or of the previous symbol u _{i of the} source, i ≠ k, while the transformation leads to the fact that the spectrum The total power density at the zero frequency of the channel symbol sequence z is zero, and the control signal thereby makes it possible that visible flicker is not present in the coded light emitted by at least one light source.

The proposed transformations provide an opportunity for a better spectral limitation, in particular, for low frequencies than, for example, the Manchester pulse, while having comparable complexity.

The symbol z _{k, 2} can be a signed weighted average of L elements u _{k-l + 1} + u _{k + 1} for l = 1, ... L, leading to the fact that the power spectral density of the sequence of channel symbols z decays as (fT ) ^(4L) , where f means the frequency in Hz, and T means the time in seconds to transmit u _k . The Manchester coding spectrum fades as f ² for f close to zero. Thus, the proposed coding scheme decays faster than Manchester coding.

In accordance with the second aspect, the above objectives are achieved using a lighting device comprising at least one light source, modulation means and a light control device in accordance with the foregoing, while the modulation means is configured to, in accordance with a control signal provided by the control device light, modulating coded light that must be emitted by at least one light source.

In accordance with a third aspect, the above objectives are achieved by a method for controlling at least one light source configured to emit coded light based on a control signal, the method comprising: receiving a sequence u = [u ₁ , .. ., u _k , ..., u _K ] characters u _{k of the} source, representing the information sequence of the source of information; determining a sequence z = [z ₁ , ..., z _k , ..., z _K ] of symbols z _{k of a} channel forming a control signal from a sequence of source symbols; providing at least one light source with a control signal, thereby controlling at least one light source; wherein the sequence of symbols z of the channel is determined by converting each symbol u _{k of the} source at time k into a composite symbol z _k = (z _{k, 1} , z _{k, 2} ) of the channel containing at least one first symbol z _{k, 1 of the} channel, one of which is identical to u _k , and at least one second channel symbol z _{k, 2} selected from the set M, where z _{k, 2} is a function of the source symbol u _k and at least one subsequent and / or previous source symbol u _i , i ≠ k, while the conversion leads to the fact that the power spectral density at zero hour the frequency of the z-channel symbol sequence is zero, and the control signal thereby makes it possible that visible flicker is not present in the coded light emitted by at least one light source.

According to a fourth aspect, the above objectives are achieved by an information decoder for decoding an information sequence from a signal received from a light detector, the signal indicating that the encoded light is emitted from at least one light source controlled by the light control device in accordance with with the foregoing, comprising: a receiver configured to receive a signal from a light detector, wherein the signal characterizes a sequence z = [z ₁ , ..., z _k , ..., z _K ] symbol in z _k = (z _{k, 1} , z _{k, 2} ) channels; a processing unit configured to determine from the signal the sequence û = [û ₁ , ..., û _k , ..., û _K ] decoded symbols û _{k of the} source forming the decoded information sequence; wherein the control unit is configured to determine the sequence of decoded source symbols û, whereby the decoded source symbol û _k is determined from the difference between z _{k, 1} and the weighted average of at least one symbol z _{k, 2} channels and one subsequent and / or previous symbol z _{i of the} channel, i ≠ k.

The processing unit may comprise a hard decision decoder. A hard decision decoder may be configured to determine a sequence of decoded source symbols û as a sign of the difference. Such a decoder can allow fast and accurate decoding.

According to a fifth aspect, the above objectives are achieved by an information decoder for decoding an information sequence from a signal received from a light detector, the signal indicating that the encoded light is emitted from at least one light source controlled by the light control device in accordance with the foregoing comprising: a receiver configured to receive a signal from a light detector, wherein the signal characterizes a sequence z = [z ₁ , ..., z _k , ..., z _K ] of characters z _k = (z _{k, 1} , z _{k, 2} ) channel; a processing unit configured to determine from the signal the sequence û = [û ₁ , ..., û _k , ..., û _K ] decoded symbols û _{k of the} source forming the decoded information sequence; the control unit is configured to determine the sequence of decoded source symbols символов, whereby the decoded source symbol û _k is determined by taking linear combinations of z _{l, 1} and half of the symbols z _{l, 2, L} and z _{l, 2, R} for values l in neighborhood k, where z _{k, 2, L} is the first half of z _{k, 2} , and z _{k, 2, R} is the second half of z _{k, 2} .

The processing unit may be configured to determine a sequence of decoded source symbols посредством by using a matched filter having a set of coefficients containing a sequence of [-1, 0, 1, 1] in terms of half the symbols. According to an embodiment, the matched filter has 2N + 4 coefficients in terms of half the characters, starting with the sequence [-1, 0], followed by N pairs of coefficients ranging between [1, 1] and [-1, -1] , to the Nth pair, which is [(-1) ^N-1 , (-1) ^N-1 ], and ending with the sequence [0, (-1) ^N ].

It should be noted that the invention relates to all possible combinations of features set forth in the claims. Similarly, the advantages of the first aspect apply to the second aspect, the third aspect, the fourth aspect and the fifth aspect, and vice versa.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other aspects of the present invention will be described in more detail below with reference to the accompanying drawings, showing an embodiment (s) of the invention.

Figure 1 illustrates a lighting system in accordance with an embodiment;

2 illustrates a light source in accordance with an embodiment;

3 illustrates an information decoder in accordance with an embodiment;

4a-4c illustrate shift register encoders in accordance with embodiments;

5 illustrates a shift register decoder in accordance with an embodiment;

6a-e illustrate pulse sequences for source symbols and channel symbols in accordance with embodiments;

7a-7g illustrate spectral representations of channel symbol sequences in accordance with embodiments;

8a-8b are block diagrams in accordance with embodiments.

DETAILED DESCRIPTION OF THE INVENTION

The embodiments described below are exemplary, so that this disclosure will be comprehensive and complete and will fully convey the scope of the invention to those skilled in the art. Identical numbers refer to identical elements throughout.

Figure 1 illustrates a lighting system 1, which contains at least one light source, schematically indicated by the number 2 position. At least one light source 2 may be a lighting device and / or be part of a lighting control system, thus the lighting system 1 may be designated as a coded lighting system. The lighting device may comprise at least one light source 2. The term "light source" means a device that is used to provide light in a room to illuminate objects in a room. A room in this context is typically a room in an apartment or a room in an office, a gym, a room in a public place, or part of an outdoor environment, such as part of a street. Each light source 2 is capable of emitting coded light, as schematically illustrated by arrow 6. The emitted light thus contains a modulated portion associated with the coded light containing information sequences. The emitted light may also contain an unmodulated portion associated with a contribution to the lighting. Each light source 2 may be associated with a number of lighting settings, inter alia, related to the contribution to the illumination of the light source, such as color, color temperature, and intensity of the emitted light. In general, the contribution to the illumination of the light source can be defined as the time-averaged output of the light emitted by the light source 2. The light source 2 will be further described with reference to FIG. 2.

As noted above, at least one light source 2 emits information sequences through visible light 6. Before information sequences are emitted through visible light 6, they are converted into a channel symbol sequence to form a modulated signal. This modulation signal may then act as a control signal to control at least one light source. The control signal can thereby determine a pulse train that switches at least one light source 2 between the emission of light (in the "ON" state) and the non-emission of light (in the "OFF" state).

The lighting system 1 further comprises a device called an information decoder 4. The information decoder 4 is configured to decode the information sequence from encoded light emitted by at least one light source 2. The information decoder 4 will be further described with reference to FIG. 3.

The lighting system 1 may further comprise other devices 10 configured to control and / or provide information to at least one light source 2.

2 schematically illustrates a light source 2 in terms of a number of functional blocks. The light source 2 comprises an emitter 14 for emitting coded light. The emitter 14 may contain one or more LEDs, but could also contain one or more sources of FL or HID, etc. In general, coding schemes may use multiple light sources. For example, a 3-level coding scheme may have two LEDs that use transforms (OFF, OFF) for level “-A”, (ON, OFF) for level “0” and (ON, ON) for level “+ A”. How levels are determined is described below. The emitter is controlled by the light control device 18. The light control device 18 may comprise or be part of a processing unit 16, such as a central processing unit (CPU). Essentially, the light control device 18 comprises a receiver 20 and a transmitter 24. The receiver 20 may be configured to receive settings, control information, code parameters, and the like. The receiver 20 may be a receiver configured to receive coded light. The receiver 20 may include an infrared interface for receiving infrared light. Alternatively, receiver 20 may be a radio receiver for receiving wirelessly transmitted information. In addition, alternatively, the receiver 20 may include a connector for receiving information transmitted over a wired connection. The wire may be a power line cable. The wire may be a computer cable. Information related to the settings, control information, code parameters and the like can be stored in the memory 22.

The light control device 18 may receive information through a receiver 20 associated with an information sequence to be transmitted by the light-encoded light source 2. For example, using the processing unit 16, the light control device 18 can change the encoding of the encoded light so that the encoded light emitted by the emitter 14 will contain an information sequence (an encoded version of the information sequence). To achieve such a transmission, the light control device 18 may be configured to perform a number of functions. For example, the receiver 20 is configured to obtain a sequence u = [u ₁ , ..., u _k , ..., u _K ] of source symbols u _k representing the information sequence of the information source. The processing unit 16 is configured to determine from a sequence of symbols a source of a sequence z = [z ₁ , ..., z _k , ..., z _K ] channel symbols z _k forming a control signal. The transmitter 24 is configured to provide the light source 2 with a control signal, thereby controlling the light source 2. These functions will be described in more detail below with reference to the flowchart of FIG. 8a. Alternatively, the light source 2 does not comprise a light control device. The light control device 18 may thus be part of the lighting system 1.

The information decoder 4 may be configured to detect and receive light, such as coded light, containing information sequences emitted by at least one light source 2, as well as light emitted by light sources outside of the lighting system 1 (not shown). The receiver 4 is configured to determine from the detected and received light information sequences transmitted by at least one light source 2. A functional block diagram for the information decoder 4 in accordance with an embodiment of the present invention is shown in FIG. The information decoder 4 comprises a receiver 34 configured to receive a signal from the light detector 32, the signal serving as a sign of the sequence z = [z ₁ , ..., z _k , ..., z _K ] characters z _k = (z _{k , 1} , z _{k, 2} ) channels. The information decoder 4 further comprises a processing unit 26 adapted to determine from the signal a sequence û = [û ₁ , ..., û _k , ..., û _K ] of decoded source symbols û _k forming a decoded information sequence. To achieve such a determination, information decoder 4 may be configured to perform a number of functions. These functions will be described below with reference to the flowchart of FIG. 8b. The information decoder 4 may further comprise a memory 28 and a transmitter 30. The memory 28 may store instructions related to functions for evaluating the information sequence. The transmitter 30 may be used to transmit information to at least one light source 2 in the lighting system 1.

As noted above, before information sequences are emitted through visible light 6, they are converted into a sequence of channel symbols to form a modulated signal. The conversion includes determining a modulation scheme by which the channel symbol sequence is to be modulated. The Manchester code (or biphasic code) is usually used as a means to modulate coded light. In Manchester coding, each input bit (also called a user bit) is converted to a pair of channel characters, where the channel characters in each pair have different characters. Namely, “1” is converted to the pair {+1, -1}, while “0” is converted to the pair {-1, +1} (or vice versa). Obviously, the Manchester code is free of DC because the average transmitted amplitude is zero. Fig. 7a illustrates at 52 position number a frequency representation (in the form of a power spectral density, PSD) of a transmitted waveform for an example with a data rate of 500 user bits per second using the Manchester code. However, due to the relatively slow decay of the PSD, visible flicker may be introduced into the emitted light.

This problem can be overcome using the methods, devices and ideas of the system, as described below, which provide transformations from the source symbols u to the channel z symbols, leading to the PSD at zero frequency of the channel z sequence of symbols being zero, and where the PSD has faster attenuation than the Manchester code. A control signal (i.e., a signal based on a sequence of channel symbols generated by the light control device 18 and used to control at least one light source 2) thereby makes it possible that visible flicker is not present in the coded light emitted by at least one light source 2. Means for providing such a transform as well as decoding an information sequence that has been modulated by such a transform will be disclosed below. In short, a set of transformations is provided that can be considered as a modification of the Manchester Code Modulation Rule, which translates bits into oscillatory signals. The proposed transformations are especially useful for transmitting information embedded in the light. Good low-frequency properties ensure no visible flicker and low susceptibility to low-frequency interference. In particular, the proposed transformations make it possible to improve the "DC-free" property of the Manchester code, thus leading to much less visible flicker (or even complete absence of flicker) of the emitted coded light, while improving the suppression of 50 Hz and 100 Hz interference in the information decoder for a given (low) bit rate due to faster decay of the PSD. Such various aspects of a modulation system for communication through visible light will be discussed below.

In accordance with the first proposed conversion, each user bit u _k , u _k ∈ {-1, +1} from the string of user bits u is converted to a group of two symbols (z _{k, 1} , z _{k, 2} ) of the channel, where z _{k, 1} , z _{k, 2} ∈ {-1, 0, +1}, where z _{k, 2} depends on u _{k of the} current group and u _{k + 1 of the} next group. Thus, the transformation for z _{k, 1 is} defined as z _{k, 1} = u _k , and the transformation for z _{k, 2 is} defined as: u _{k, 2} = - (u _k + u _{k + 1} ) / 2. The sequence of symbols z of the channel is thus formed by pairs of two symbols (z _{k, 1} , z _{k, 2} ) of the channel. One of the advantages of coding using the first proposed transform over Manchester coding is the power spectrum, such as the spectral power density, which corresponds to the Fourier transform of the autocorrelation function of the channel sequence. For amplitude modulation with amplitude A near the average level A _{0, the} sequence of channel symbols z is transmitted with amplitude A0 + A · z. The power spectrum for Manchester coding is thus equal to

,

,

where T is the user bit transmission time. For the first proposed conversion, the power spectrum is

.

.

Thus, the spectrum of the first proposed transform is more strongly suppressed around frequencies that are multiples of 2 / T than in Manchester coding. In particular, for frequencies close to zero, the Manchester power spectrum is O (f ² ), while the power spectrum for the first proposed conversion is O (f ⁴ ). In other words, the Manchester coding spectrum fades out as f ² for f close to zero, while the power spectrum using the first proposed transform fades out as f ⁴ for f close to zero. At position number 54 in FIG. 7b, the PSD for the first proposed conversion is compared with the Manchester (biphasic) conversion at position number 56 for an equal user bit rate, respectively. As can be seen in this figure, the first proposed transformation has a much wider zone of signal absence around DC (zero frequency) than the Manchester (biphasic) transformation.

If the spectral “hole” around DC is to be expanded, the first proposed transformation can be extended to the second class of proposed transformations. In accordance with the modulation using the second proposed transformation, one bit u _k ∈ {-1, +1} of the source is converted to the word [z _{k, 1} ¹ ... z _{k, 1} ^j ... z _{k, 1} ^N , z _{k, 2} ] containing N + 1 characters, where the first N characters are defined as: z _{k, 1} ^j = (- 1) ^j-1 u _k , for j = 1, ..., N, and where z _{k, 2 is} defined by the transformation z _{k, 2} = - (z _{k, 1} ^N + z _{k + 1,1} ¹ ) / 2. The effect of the second proposed conversion is that the signal band is shifted up in frequency, thereby freeing up low frequencies. The second proposed transformation can be considered as a type of phase modulation of a square carrier, where the unifying symbol z _{k, 2} serves to “smooth” the transitions to keep the required bandwidth small. Fig. 7c shows by PSD position number 58 for the second proposed conversion for N = 2 in comparison with the Manchester pulse by 60 position number and a transformation defined as [+1, -1, -1, +1], by 62 position number, all at the same bit rate of 500 bps. Fig. 7d shows at PSD position number 64 for the second proposed conversion for N = 4 all for the same bit rate of 500 bps and again compared to the Manchester pulse at position number 66 and the conversion defined as [+1 , -1, -1, +1], by the number 68 positions.

However, even stronger suppression than achieved by the first and second proposed transformations may be necessary in some applications. The first and second proposed transformations can be further extended to the class of third proposed transformations. The class of the third proposed transformations is denoted by an integer L and includes the conversion of the source symbols u _k to the channel symbols z _{k of} value 3L. The resulting transmitted signal (when amplitude modulation is used) has low power at low frequencies and attenuates as (fT) ^4L , where f is the frequency and T is the bit transmission time. The corresponding threshold matched filter receiver has a similar low-frequency interference suppression. For L = 1, the first proposed transformation is covered. For L = 2, the circuit provides a good balance between practicality and the desired low-frequency properties.

In particular, in accordance with the class of third proposed transformations, each channel symbol z _k represents a pair (z _{k, 1} , z _{k, 2} ), where z _{k, 1} = u _k , and where intermediate symbols z _{k, 2} will be chosen as

,

,

where for a given value of L, the coefficients a _{L, l are} determined by the condition that

,

,

so the power spectrum

.represents

.

For each value of L, the search for the coefficients a _{L, l} is a matter of linear algebra. The results for the first few L values are

Intermediate symbols z _{k, 2} , being the weighted sum of L ternary symbols, can take 3 ^L different values. As noted above, for L = 1, the first proposed transformation is covered. Setting L = 2 gives a significant low-frequency improvement over L = 1, while maintaining the number of levels limited to nine, namely the values in the set:

.

.

Since the maximum amplitude is 5/4 in this case compared to 1 for L = 1, the maximum modulation depth A for L = 2 is 4/5 times the maximum amplitude for L = 1.

The user bit sequence is thus converted to a channel symbol sequence, as described above. These channel symbols are transmitted over the channel (i.e., from at least one light source 2 to the information decoder 4) using communication through visible light, for example, using amplitude modulation. In the information decoder 4, the received soft symbols can be sent through a matched filter, and the calculated sequence of the source, or user bits, can be detected by a threshold circuit (using, for example, a sign function), see below. Fig. 7e shows the PSD as a function of fT divided by A ² T for Manchester coding at position number 76, as well as for the third class of the proposed conversion for L = 1 at position number 72, L = 2 at position 74 number and L = 3 by number 78 positions, respectively. As noted in the figure, the frequency gap near the zero frequency increases with increasing L.

As described above, in accordance with the first proposed transformation, each user bit u _k , u _k ∈ {-1, +1}, from the string of user bits u is converted to a group of two symbols (z _{k, 1} , z _{k, 2} ) of the channel , where z _{k, 1} , z _{k, 2} ∈ {-1, 0, +1}. The first and second proposed transformations, for which the output oscillating signal can thus take values in the ternary alphabet {A _low , 0, A _high }, can be further extended to the class of the fourth proposed transformations for which the output oscillating signal takes binary values Alphabet {A _low , A _high }. Thus, for the class of fourth proposed transformations, each user bit u _k , u _k ∈ {-1, +1} from the string of user bits u is transformed into a group of two symbols (z _{k, 1} , z _{k, 2} ) of the channel, where z _{k , 1} , z _{k, 2} ∈ {-1, +1}. Therefore, while the first and second proposed transforms usually require three light output levels, the fourth proposed transform usually requires only two light output levels; for the first and second proposed transformations, the vibrational signal corresponding to the +1 symbol can, as described above, be a pulse with a high amplitude A _high , the vibrational signal corresponding to the symbol -1 can be a pulse with a low amplitude A _low , and the vibrational signal corresponding to symbol '0', may be a pulse with an amplitude (A _high + A _low ) / 2.

For the class of fourth proposed transformations, the symbol '0', which is used only for z _{k, 2} , and therefore has nonzero support in [T / 2, T), can be formed using a combination of _high -level amplitude A _high and amplitude A _low low level while maintaining a fourth-order low frequency power spectrum. For the oscillatory signal '0' W ₀ (t) with support in [T / 2, T), the requirements can thus be formulated as:

.

.

If the number of level changes within a symbol is limited, the number of vibrational signals satisfying these restrictions is, of course. The minimum number of level changes for such an oscillatory signal W ₀ (t) is two, and in this case there are two possibilities: the first possibility has W ₀ (t) = A _low for t∈ [T / 2, 5T / 8) and [7T / 8, T) and W ₀ (t) = A _high for t∈ [5T / 8, 7T / 8), the second possibility is the same as the first, but with high and low levels exchanged.

Let W _{+ be the} oscillation signal for transmitting the symbol '+1' and W _{is the} oscillatory signal for transmitting the symbol '-1'. The symbol '0' can thus be transmitted as an oscillating signal denoted by W _{+ -} if it is preceded by a symbol '+1', followed by a symbol '-1', and as another oscillatory signal denoted by W _{- +} , if it is preceded by a '-' character and is followed by a '+' character. For example, if the vibrational signals W _{+ -} and W _{- +} satisfy the following conditions (which, therefore, correspond to conditions a), b) and c) above):

,

,

The power spectral density is free of delta peaks at nonzero frequencies and has a fourth-order low-frequency behavior.

With one level change inside each oscillatory signal W _{+ -} and W _{- + there is} no level change during transitions between the '0' symbol and previous symbols and no level change during transitions between the '0' symbol and the following symbol, W _{+ -} and W _{- +} cannot be selected as

W _{+ -} (t) = A _high for all t∈ [T / 2, 5T / 4), and W _{+ -} (t) = A _low for all t∈ [5T / 4, T)

W _{- +} (t) = A _low for all t∈ [T / 2, 5T / 4), and W _{- +} (t) = A _high for all t∈ [5T / 4, T).

Thus, z _{k, 2} is represented by either W _{+ -} or W _{- +} depending on the values of u _k and u _{k + 1} . If u _k = + 1 and u _{k + 1} = -1, then z _{k, 2} is represented by W _{+ -} and vice versa. In addition, since W _{+ -} and W _{- +} change levels, each z _{k, 2} will also change levels and, in accordance with the above, will consist of two sub-parts; for W _{+ - the} first sub-part will be set to A _high , and the second sub-part will be set to A _low , and for W _{- + the} first sub-part will be set to A _low , and the second sub-part will be set to A _high . At position number 92 in FIGS. 7f-7g, where FIG. 7g gives an enlarged version around the zero frequency of the graph in FIG. 7f, the PSD for the fourth proposed conversion is compared with the Manchester (biphasic) conversion at position number 94 and the first proposed conversion at number 96 Positions for equal user bit rates. As can be seen in FIGS. 7f-7g, the fourth proposed conversion has a narrower zone of no signal around DC (zero frequency) than the first proposed conversion, but still wider than the Manchester (biphasic) conversion.

The use of two different vibrational signals for intermediate symbols (i.e., the symbols '0') can also be applied to the class of second proposed transformations. In particular, if the ternary symbol z _{k, 2} , which must be inserted between z _{k, 1} ^N-1 and z _{k + 1,1} ¹ , is '0' (i.e., if u _{k + 1} = (- 1) ^N u _k ), the vibrational signal W _{+ -} can be used if z _{k + 1,1} ¹ = -1, and the vibrational signal W _{- +} can be used if z _{k + 1,1} ¹ = 1.

For the third class of proposed transformations, z _{k, 2} can, in accordance with the above, take 3 ^L different values, some of which are more than +1, or less than -1. The values of z _{k, 1} and z _{k, 2} characters of the third class of the proposed transformations can be scaled so that they all lie in the interval [-1, 1]. As an example, for L = 2, the possible values become -1, -9/10, -4/5, -1/10, 0, 1/10, 4/5, 9/10, 1 for z _{k, 2} and -4/5, +4/5 for z _{k, 1} . A vibration signal representing the z symbol can be generated using a high level for a fraction of (1 + z) / 2 symbol transmission time and a low level for a fraction (1-z) / 2 symbol transmission time. In addition, for characters z _{k, 2} three 'high' values of 4/5, 9/10 and 1 character can occur only between a pair of characters (z _{k, 1} , z _{k + 1,1} ) equal to (-4/5 , -4/5), therefore, for high symbol values, the oscillatory signal can preferably start and end with 'high' and have a 'low' part in the middle. Similarly, vibrational signals for three low values of -1, -9/10 and -4/5 characters can start and end with 'low' and have a 'high' part in the middle. Three 'average' values of -1/10, 0, 1/10 characters can only occur between a pair of (-4/5, 4/5) or (4/5, -4/5), so for these characters it may be preferable have two different vibrational signals; the first waveform starting with 'low' and ending with 'high', the other waveform starting with 'high' and ending with 'low', and where the waveform is used, which is consistent with adjacent signal levels.

In general, both Manchester coding and coding based on the first proposed transformation allow a simple threshold receiver based on matched filters:

and

,

,

respectively, where z _k on the right side shows the received parts in the information decoder 4. The low-frequency suppression of these filters directly relates to the power spectra of the corresponding signals. A simple decoder implementation for the first proposed transform may thus comprise an integrating filter with a reset for each received channel symbol z. The hard decision decoder can thus evaluate the expression

and compare the result with a threshold of 0 to choose between transmitted +1 or -1 (i.e. either u _k = + 1 or u _k = -1). From this and the foregoing, one skilled in the art understands how a matched threshold receiver can be implemented for the second and third proposed transforms, respectively. In particular, the matched filter threshold receiver for the third proposed transform may be expressed as

,

,

where the z-values on the right side are the measured values of these symbols in the receiver.

The above expression (z _{k, 1} - (z _k-1,2 + z _{k, 2} ) / 2) for the first proposed conversion corresponds to the output of a receiver having a matched filter with coefficients [-1/2, 1, -1/2] filter. Similarly, matched filters corresponding to the proposed second conversion class may have the following sets of coefficients:

N = 1: [-1/2, 1, -1/2]

N = 2: [-1/2, 1, -1, 1/2]

N = 3: [-1/2, 1, -1, 1, -1/2]

N = 4: [-1/2, 1, -1, 1, -1, 1/2]

and so on (as noted above, for N = 1, the second transformation class corresponds to the first transformation).

In general, a receiver based on a matched filter convolves the filter coefficients with the received vibrational signal and then outputs the values at sample times. If the user bit u _{k is} preceded and followed by user bits with the same value, the characters preceding and following the character corresponding to the bit have the opposite meaning, so convolution with a matched filter gives u _k · (A _high -A _low ) · T / 2. If one of the neighboring bits has the opposite sign, one of the neighboring symbols is '0', and the output of the matched filter is (3/4) · u _k · (A _high -A _low ) · T / 2. In addition, if both adjacent bits have the opposite sign, the output of the matched filter is (1/2) · u _k · (A _high -A _low ) · T / 2.

The receiver for the fourth proposed conversion may also be based on a matched filter. In particular, the aforementioned matched filter threshold receiver for the first proposed transform may also be used for the fourth proposed transform with good results. In general terms, for optimal decoding of the fourth transformation, the receiver must take into account the signal levels in both halves of the symbol of the oscillatory signal representing the symbol z _{k, 2} (for all k), i.e. characters that can represent '0'. Denote the first half of z _{k, 2} as z _{k, 2, L} , and the second half of z _{k, 2} as z _{k, 2, R.} In other words, z _{k, 2, L} corresponds to z _{k, 2} for t∈ [T / 2, 5T / 4), and z _{k, 2, R} corresponds to z _{k, 2} for t∈ [5T / 4, T) inside of each interval [0, T). Thus, decoding u _k involves taking linear combinations of z _{l, 1} , z _{l, 2, L} and z _{l, 2, R} for values of l in the neighborhood of k. In particular, the matched filter for the fourth proposed conversion may preferably have an impulse response characteristic corresponding to the [-1, 0, +1, +1, 0, -1] filter coefficients, and therefore be divided into six blocks, each having a duration T / 4. Thus, for these filter coefficients, the output signal is independent of the value of the adjacent bits.

The matched filter filter coefficients for the fourth proposed transform may also be used in the matched filter based receiver for the first proposed transform. Using the matched filter filter coefficients for the fourth proposed transform in the matched filter receiver for the first proposed transform can for user bits that are detected by taking the sign of the output of the matched filter can reduce the probability of incorrect detection due to additional noise if one or both of the adjacent user bits have the opposite sign. In particular, for the class of second proposed transformations, matched filters can have the following sets of coefficients in terms of 'half-characters':

N = 1: [-1, 0, 1, 1, 0, -1]

N = 2: [-1, 0, 1, 1, -1, -1, 0, 1]

N = 3: [-1, 0, 1, 1, -1, -1, 1, 1, 0, -1]

N = 4: [-1, 0, 1, 1, -1, -1, 1, 1, -1, -1, 0, 1]

and so on (as noted above, for n = 1, the second class of transformations corresponds to the first transform). Thus, a generally consistent filter can have 2N + 4 coefficients in terms of 'half characters'. Odds can start with the sequence [-1, 0]. This starting sequence can then be followed by N pairs of coefficients ranging between [1, 1] and [-1, -1] to the Nth pair, which is [(-1) ^N-1 , (-1) ^N-1 ]. The coefficients may then end with the sequence [0, (-1) ^N ]. The receiver can use either a filter for decoding, or a transform with the same value of N, of course. Filters for the fourth transform are equally good or better than filters for the first or second transform. Detection of a scaled third class of proposed transforms may require higher temporal accuracy (such as oversampling in a detector) than for the originally proposed third class of transforms. However, a matched filter for the amplitude-based L = 2 system can be used in the detector for the above-described scaled symbols, thereby reducing the time accuracy requirements.

Both the encoder (i.e., the conversion process) and the decoder (the inverse conversion process) can be implemented using shift registers. Fig. 4a illustrates, at position number 36, an implementation on shift registers of an encoder based on the first proposed transform. The source sequence u is supplied to a shift register having two memory cells, one for storing u _{k + 1} and one for storing u _k . z _{k, 1 is} directly obtained from u _k . z _{k, 2 is} obtained by first adding u _k and u _{k + 1} , and then multiplying this sum by a factor of -1/2. z _{k, 1} and z _{k, 2 are} then serialized in a multiplexer (mux) to form the final sequence of z channel symbols.

Fig. 4b illustrates, at position number 38, an implementation on shift registers of an encoder based on the second proposed transform. The source sequence u is supplied to a shift register having two memory cells, one for storing u _{k + 1} and one for storing u _k . z _{k, 1} is formed as a sequence z _{k, 1} = [z _{k, 1} ¹ , ..., z _{k, 1} ⁿ , ..., z _{k, 1} ^N ], where each z _{k, 1} ^{n is} obtained from the sequence alternating u _k with a sign; z _{k, 1} ⁿ = u _k if n is odd, and z _{k, 1} ⁿ = -u _k if n is even. In other words, to get z _{k, 1} ⁿ , u _{k is} either multiplied by +1 (i.e. no operation is required), or by -1. The resulting sequence z _{k, 1} = [z _{k, 1} ¹ , ..., z _{k, 1} ⁿ , ..., z _{k, 1} ^N ] is then fed to the multiplexer (mux). z _{k, 2 is} obtained by first adding z _{k, 1} ^N and u _{k + 1} , and then multiplying this sum by a factor of -1/2. z _{k, 1} and z _{k, 2 are} then serialized in the multiplexer to form the final sequence of z channel symbols.

Fig. 4c illustrates, at reference numeral 40, an implementation on shift registers of an encoder based on the third proposed conversion. For L = 2, the source sequence u is fed into a shift register having four memory cells for storing u _{k + 2} , u _{k + 1} , u _k and u _k-1 , respectively. z _{k, 1 is} directly obtained from u _k . z _{k, 2 is} obtained from the weighted sum of u _{k + 2} , u _{k + 1} , u _k and u _k-1 , respectively. As noted above, the weighting factors for L = 2 are 1/16, -9/16, -9/16 and 1/16, respectively. Thus, according to the first implementation, u _{k + 2} , u _{k + 1} , u _k and u _{k-1 are} first multiplied by 1, -9, -9 and 1, respectively, then they are added and then multiplied by a factor of 1/16, to get z _{k, 2} . Alternatively, z _{k, 2} can be found by directly multiplying u _{k + 2} , u _{k + 1} , u _k and u _k-1 by 1/16, -9/16, -9/16 and 1/16, respectively, thus eliminating the need for an additional multiplier. z _{k, 1} and z _{k, 2 are} then serialized in a multiplexer (mux) to form the final sequence of z channel symbols.

Figure 5 illustrates at position number 42 the implementation of the shift registers of the decoder based on the first proposed conversion. The resulting channel z sequence is fed to a demultiplexer (dmux) to obtain z _{k, 1} and z _{k, 2} . z _{k, 2} is supplied to the shift register, which has two memory cells, one for storing z _k-1,2 and one for storing z _{k, 2} . z _k-1,2 and z _{k, 2 are} added and multiplied by a factor of -1/2. The calculated sequence û of the source is then obtained by adding the result of this multiplication with z _{k, 1} and then taking the sign of the sum.

Figa-d give an example of the application of the first proposed conversion. 6a illustrates at position number 44 the sequence u = [- 1, -1, +1, +1, +1, -1, ...] of the source, which, among other things, can represent the information sequence [0, 0, 1, 1, 0 , ...]. The representation of the waveform for the source sequence u is illustrated by the position number 46 in FIG. 6b, where each source bit has a time duration of T seconds. For illustrative purposes, only the amplitude of the oscillatory signal is limited by the interval from -1 to +1, but in general it can be from -A to + A if amplitude modulation with amplitude A is used. As one skilled in the art understands, other methods are equally possible modulation, in addition to amplitude modulation. FIG. 6c illustrates at position number 48 the corresponding sequence of channel symbols z, i.e. z = [- 1, +1, -1, 0, +1, -1, +1, 0, -1,?, ...], where "?" is either "0" or "+1", depending on the next source character. The representation of the vibrational signal for the channel z sequence is illustrated by the position number 50 in FIG. 6d. Thus, a conversion is similar to biphasic if the user bits following each other are the same. However, if the two consecutive bits are not the same, the first proposed conversion outputs “0” as the merge bit z _{k, 2} , which leads to a smooth transition of the transmitted vibrational signal, thus preventing the generation of frequencies that are “very far” from the quasi-carrier frequency (500 Hz in the example above).

Fig.6e gives an example of the application of the fourth proposed transformation used for the same example as for figa-d, i.e. for the sequence u = [- 1, -1, +1, +1, -1, ...] of the source, which, among other things, can represent the information sequence [0, 0, 1, 1, 0, ...] . The representation of the waveform for the channel z sequence for the fourth proposed transform is illustrated at 90 in FIG. 6e. The representation of the vibrational signal is similar to that of the first proposed transformation illustrated in FIG. 6d. However, if two consecutive user bits u _k , u _{k + 1 are} not identical, the fourth proposed conversion does not produce '0' as a merge bit z _{k, 2} but instead returns either W _{+ -} (if it is preceded by a ' +1 'and followed by the character' -1 '), or W _{- +} (if preceded by the character' -1 'and followed by the character' +1 '), where W _{+ -} and W _{- + are} determined in accordance with and where, for the present example, A _high = '+ 1' and A _low = '- 1'.

The application of the proposed conversions is suitable for communication through visible light (VLC) between lamps or between lamps and a control panel. In terms of the OSI stack (Open Systems Interconnection) defining such a connection, this would form part of the physical layer definition. One skilled in the art will recognize that the present invention is in no way limited to the preferred embodiments described above. On the contrary, many modifications and variations are possible within the scope of the attached claims.

For example, although the disclosed embodiments have been illustrated in the context of a coded light environment, the proposed transformations are applicable to general data transmission and information transmission, such as radio communication. The disclosed light control device may then be replaced by a control device configured to control a radio transmitter. Similarly, the information decoder may be configured to decode the information sequence from the received radio signal. In addition, the various levels included in the individual channel symbols were converted to the amplitude of the transmitted light level. These levels can be converted to a pulse width modulation circuit (which may also be suitable for an LED based VLC system). The proposed schemes can even be implemented in such a way that the information decoder does not have to determine which system (amplitude modulation or pulse width modulation) is used in the transmitter.

Claims (18)

1. A light control device (18) for controlling at least one light source (2) configured to emit coded light based on a control signal, wherein the light control device comprises:
a receiver (20) configured to receive a sequence u = [u ₁ , ..., u _k , ..., u _K ] of source symbols u _k representing the information sequence of the information source;
a processing unit (16) configured to determine from a sequence of symbols the source of the sequence z = [z ₁ , ..., z _k , ..., z _K ] symbols z _{k of the} channel forming the control signal;
a transmitter (24), configured to provide at least one light source with a control signal, thereby controlling at least one light source;
the processing unit is configured to determine the sequence of symbols z of the channel by converting each symbol u _{k of the} source at time k into a composite symbol z _k = (z _{k, 1} , z _{k, 2} ) of the channel containing at least one first symbol z _{k, 1} channels, one of which is identical to u _k , and at least one second symbol z _{k, 2} channels having a value selected from the set M, and the value z _{k, 2} is determined according to the weight function having the source symbol u _k and at least one subsequent and / or previous source symbol u _i , and y The word i ≠ k is specified as input parameters, while the weight function contains at least two nonzero weight coefficients, and the transformation leads to the fact that the spectral power density at zero frequency of the sequence of channel symbols z is zero, and
wherein the control signal thereby makes it possible that visible flicker is not present in the coded light emitted by at least one light source. 2. The light control device according to claim 1, wherein the number of elements in the set M is odd. 3. The light control device according to claim 1 or 2, in which the set M is symmetric with respect to zero. 4. The light control device according to claim 1 or 2, in which the zero element is included in the set M. 5. The light control device according to claim 1 or 2, in which z _{k, 2} is a weighted average of L elements u _{k-l + 1} + u _{k + 1} for l = 1, ... L, leading to that the power spectral density of the sequence of channel symbols z decays as (fT) ^(4L) , where f means the frequency in Hz and T means the time in seconds for transmitting u _k . 6. The light control device according to claim 5, in which the weighting factor a _{L, 1} for L elements is determined so that the equality holds:

. 7. The light control device according to claim 6, in which the weighting factor a _{L, 1 is} further determined so that the equality holds:

for amplitude modulation with amplitude A of the channel symbol sequence. 8. The light control device according to claim 1 or 2, in which z _{k, 2} = - (u _k + u _{k + 1} ) / 2. 9. The light control device according to claim 1 or 2, where z _{k, 1} is the sequence of the first characters z _{k, 1} = [z _{k, 1} ¹ ... z _{k, 1} ^j ... z _{k, 1} ^N ] channel, where the jth symbol is given as z _{k, 1} ^j = (- 1) ^j-1 u _k , for j = 1, ..., N, and where z _{k, 2} = - (z _{k, 1} ^N + z _{k + 1,1} ¹ ) / 2. 10. A lighting device comprising at least one light source (2), modulation means and a light control device (18) according to any one of claims. 1-9, wherein the modulation means is configured to, in accordance with the control signal provided by the light control device, modulate the encoded light that should be emitted by at least one light source. 11. A control method for at least one light source (2) configured to emit coded light based on a control signal, the method comprising the steps of:
accept (80) the sequence u = [u ₁ , ..., u _k , ..., u _K ] characters u _{k of the} source, representing the information sequence of the source of information;
determine (82) the sequence z = [z ₁ , ..., z _k , ..., z _K ] of the symbols z _{k of the} channel forming the control signal from the sequence of source symbols;
provide (84) at least one light source with a control signal, thereby controlling at least one light source;
wherein the sequence of channel symbols z is determined by converting each source symbol u _k at time k into a composite channel symbol z _k = (z _{k, 1} , z _{k, 2} ) containing at least one first channel symbol z _{k, 1} , one of which is identical to u _k , and at least one second channel symbol z _{k, 2} having a value selected from the set M, the value z _{k, 2 being} determined according to a weight function having a source symbol u _k and at least one subsequent and / or the previous symbol u _{i of the} source, and the condition i ≠ k is specified as input pairs meters, while the weight function contains at least two non-zero weight coefficients, and the conversion leads to the fact that the spectral power density at zero frequency of the sequence of z channel symbols is zero, and
wherein the control signal thereby makes it possible that visible flicker is not present in the coded light emitted by at least one light source. 12. An information decoder (4) for decoding an information sequence from a signal received from the light detector (32), the signal indicating that the encoded light is emitted from at least one light source (2) controlled by the light control device (18) according to any one of paragraphs. 1-9, containing:
a receiver (34) configured to receive a signal from a light detector, the signal characterizing a sequence z = [z ₁ , ..., z _k , ..., z _K ] of characters z _k = (z _{k, 1} , z _{k , 2} ) a channel;
a processing unit (26) configured to determine from a signal a sequence û = [û ₁ , ..., û _k , ..., û _K ] of decoded source symbols û _k forming a decoded information sequence;
the processing unit is configured to determine the sequence of decoded source symbols û, whereby the decoded source symbol û _k is determined from the difference between z _{k, 1} and the weighted average of at least z z _{, 2} channels and one subsequent and / or previous z symbol _i channel, i ≠ k. 13. The information decoder according to claim 12, wherein the processing unit comprises a hard decision decoder and in which the hard decision decoder is configured to determine a sequence of decoded source symbols û as a sign of said difference. 14. The information decoder according to claim 12 or 13, in which û _k is determined in accordance with û _k = sign (z _{k, 1} - (z _k-1,2 + z _{k, 2} ) / 2). 15. The information decoder according to claim 12 or 13, depending on claim 6, in which û _k is determined in accordance with

. 16. An information decoder (4) for decoding an information sequence from a signal received from the light detector (32), the signal being a sign that the encoded light is emitted from at least one light source (2) controlled by the control device (18) light according to any one of paragraphs. 1-9, containing:
the receiver (34), configured to receive a signal from the light detector, the signal characterizes the sequence z = [z ₁ , ..., z _k , ..., z _K ] characters z _k = (z _{k, 1} , z _{k, 2} ) a channel;
a processing unit (26) configured to determine from a signal a sequence û = [û ₁ , ..., û _k , ..., û _K ] of decoded source symbols û _k forming a decoded information sequence;
the processing unit is configured to determine the sequence of decoded source symbols символов, whereby the decoded source symbol û _k is determined by taking linear combinations z _{l, 1} and half the symbols z _{l, 2, L} and z _{l, 2, R} for values l in neighborhood k, where z _{k, 2, L} is the first half of z _{k, 2} , and z _{k, 2, R} is the second half of z _{k, 2} . 17. The information decoder according to claim 16, wherein the processing unit is configured to determine a sequence of decoded source symbols посредством by using a matched filter having a set of coefficients containing a sequence of [-1, 0, 1, 1] in terms of half the characters. 18. The information decoder according to claim 16 or 17, wherein the matched filter has 2N + 4 coefficients in terms of half the characters, starting with the sequence [-1, 0], followed by N pairs of coefficients ranging between [1, 1] and [-1, -1], up to the Nth pair, which is [(-1) ^N-1 , (-1) ^N-1 ], and ending with the sequence [0, (-1) ^N ].

Images