An electric injection coding device and method based on two-segment semiconductor laser

Abstract

The present invention discloses an electric injection encoding device and method based on a two-segment semiconductor laser. The device comprises: a laser comprising a gain region and an absorption region, and performing frequency encoding or spatiotemporal encoding according to a gain region signal and / or an absorption region signal; a controller for controlling the temperature of the laser; a current source for outputting a current signal; a field programmable logic gate array for providing a signal to be encoded to the gain region; a bias device for loading the current signal and the electric signal to be encoded to the gain region; a voltage source for providing a voltage signal; a field programmable logic gate array for providing a signal to be encoded to the absorption region; a bias device for loading the voltage signal and the signal to be encoded to the absorber region; an optical isolator connected to the laser and an adjustable optical attenuator; the adjustable optical attenuator connected to a spectrometer; an optical coupler for dividing the encoded optical signal into two light paths; a photodetector for converting one light path into an electrical signal; an oscilloscope for measuring the electrical signal; and a spectrometer for measuring the other light path.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical computing, and in particular relates to an electric injection coding device and method based on a two-segment semiconductor laser. Background Art

[0002] In spiking neural networks, information is represented in the form of spikes, which are continuous in time but discrete in intensity, making spiking neural networks more energy-efficient. In recent years, spike-based neuromorphic units have flourished in the fields of electronics and optics. Spike-based photonic neuromorphic processing aims to support information processing and communication on a unified device, and has advantages over the von Neumann architecture in terms of efficiency, correctness, and adaptability in solving specific tasks such as pattern recognition, decision-making, optimization, and learning. In addition, spike coding is recognized by the neuroscience community as a sparse coding strategy that is widely present in the nervous system. Similarly, in photonic pulse neural networks, pulse coding plays an important role as a hub connecting external networks with photonic pulse neural networks.

[0003] Among the existing coding schemes, there are optical injection coding based on graphene fiber lasers, optical injection and electrical injection coding based on vertical cavity surface emitting lasers, optical injection coding based on microdisk lasers, and optical injection coding based on distributed feedback lasers.

[0004] However, existing optical injection photon pulse neuron encoding schemes require the network to provide an additional optical path to modulate information onto the optical path and inject it into the laser for encoding. This makes the pulse neural network more complex and more dependent on external light. Furthermore, existing electrical injection encoding schemes based on vertical cavity surface emitting lasers have a single electrical injection interface, and the optical pulse neuron (laser) has a single function. Summary of the Invention

[0005] In order to solve the above-mentioned problems existing in the related art, the present invention provides an electric injection encoding device and method based on a two-segment semiconductor laser. The technical problem to be solved by the present invention is achieved through the following technical solutions:

[0006] The present invention provides an electric injection encoding device based on a two-segment semiconductor laser, comprising: a two-segment semiconductor laser, a temperature controller, a current source, a voltage source, a first field programmable logic gate array and a second field programmable logic gate array, a first bias device and a second bias device, an optical isolator, an adjustable optical attenuator, an optical coupler, a photodetector, an oscilloscope, and a spectrometer;

[0007] The two-segment semiconductor laser comprises a gain region and a saturation absorption region; is used to perform frequency encoding or spatiotemporal encoding based on a signal in the gain region or a signal in the saturation absorption region to obtain a coded optical signal; and performs calculation based on a signal in the gain region and a signal in the saturation absorption region, and performs frequency encoding or spatiotemporal encoding to obtain a coded optical signal;

[0008] The temperature controller is connected to the two-segment semiconductor laser and is used to control the operating temperature of the two-segment semiconductor laser;

[0009] The current source is connected to the first bias device and is used to output a current signal;

[0010] The first field programmable logic gate array is connected to the first bias device and is used to provide the electrical signal to be encoded to the gain region;

[0011] The first biaser is connected to the gain region and is used to load the current signal and the electrical signal to be encoded into the gain region;

[0012] The voltage source is connected to the second bias device and is used to output a voltage signal;

[0013] The second field programmable logic gate array is connected to the second bias device and is used to provide the electrical signal to be encoded to the saturated absorption region;

[0014] The second biaser is connected to the saturated absorption region and is used to load the voltage signal and the coded electrical signal to the saturated absorber region;

[0015] The optical isolator is connected to the two-segment semiconductor laser and the adjustable optical attenuator respectively, and is used to prevent reflected light from entering the two-segment semiconductor laser;

[0016] The adjustable optical attenuator is connected to the photodetector and the spectrometer respectively through the optical coupler; the adjustable optical attenuator is used to control the power of the coded optical signal from the optical isolator; the optical coupler is used to divide the coded optical signal after power control into a first optical signal and a second optical signal;

[0017] The photodetector is configured to convert the first optical signal into an electrical signal;

[0018] The oscilloscope is connected to the photodetector and is used to measure the converted electrical signal;

[0019] The spectrometer is used to measure the second optical signal.

[0020] The present invention also provides an electric injection coding method based on a two-segment semiconductor laser, which is implemented based on the above-mentioned electric injection coding device based on a two-segment semiconductor laser, and includes:

[0021] The temperature controller is used to control the operating temperature of the two-segment semiconductor laser to a preset temperature;

[0022] Using the voltage source and the current source, a voltage value and a current value are inputted into the electrical injection encoding device each time. Simultaneously, using the first field programmable logic gate array and / or the second field programmable logic gate array, a test electrical signal is inputted into the gain region and / or the saturated absorption region each time. The coded signal is measured each time by the oscilloscope and the spectrometer to obtain a measurement result each time. The voltage value and current value inputted each time are a set of driving values.

[0023] When multiple sets of driving values ​​are input into the electrical injection encoding device, each set of driving values ​​and corresponding measurement results are obtained; at least one current value or voltage value is different between any two sets of driving values;

[0024] Based on each of the plurality of drive value groups and the corresponding measurement results, analyzing and obtaining an interaction relationship between the voltage value and the range length of the current value, and obtaining an influence relationship between the voltage value and the current value on the measurement results; the range length of the current value is the size of the range of the current value that supports the electrical injection encoding device to perform signal encoding;

[0025] According to the action relationship and the influence relationship between the two, the driving voltage value and the driving current value are selected, and the driving voltage value, the driving current value, and the electrical signal to be encoded are input into the electrical injection encoding device for frequency encoding or space-time encoding to obtain the encoding result.

[0026] The present invention has the following beneficial technical effects:

[0027] The electric injection encoding device based on a two-segment semiconductor laser proposed in the present invention can provide two electric injection interfaces, and signal encoding can be performed through either of the two electric injection interfaces. Moreover, when both electric injection interfaces are used simultaneously, the two-segment semiconductor laser can calculate and encode the signal in the gain region and the signal in the saturated absorption region. Compared with encoding methods based on light injection, this eliminates the dependence of information encoding in photon pulse neural networks on external light injection and does not require additional photoelectric conversion. Moreover, compared with electric injection encoding based on vertical cavity surface emitting lasers, it not only increases the number of electric injection interfaces but also provides certain computing functions on the two-segment semiconductor laser as a light pulse neuron. In addition, due to the high power of the light output by the two-segment semiconductor laser, the encoded signal encoded by the electric injection encoding device proposed in the present invention is more convenient for information transmission after entering the pulse neural network.

[0028] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic structural diagram of an electric injection encoding device based on a two-segment semiconductor laser provided in an embodiment of the present invention;

[0030] Figure 2 A schematic flow chart of an exemplary electric injection encoding method based on a two-segment semiconductor laser provided by an embodiment of the present invention;

[0031] Figure 3a A schematic diagram of a self-pulse signal corresponding to an exemplary set of drive values ​​provided in an embodiment of the present invention;

[0032] Figure 3b A schematic diagram of an exemplary self-pulse signal corresponding to another set of driving values ​​provided by an embodiment of the present invention;

[0033] Figure 3c A schematic diagram of a self-pulse signal corresponding to another exemplary set of driving values ​​provided in an embodiment of the present invention;

[0034] Figure 4 A schematic diagram of an exemplary FP-SA frequency curve provided in an embodiment of the present invention;

[0035] Figure 5a A schematic diagram of an exemplary signal to be encoded provided by an embodiment of the present invention;

[0036] Figure 5b The exemplary electric injection device provided by the embodiment of the present invention is Figure 5a Schematic diagram of the self-pulse signal obtained after encoding the signal to be encoded shown;

[0037] Figure 5c An exemplary embodiment of the present invention provides statistics Figure 5b Schematic diagram of a signal period obtained from the time interval between every two adjacent spike pulses in the pulse signal shown;

[0038] Figure 5d A schematic diagram of the self-pulse frequency of the self-pulse signal shown in 5b provided in an embodiment of the present invention;

[0039] Figure 5e A schematic diagram showing the peak pulses of an output signal corresponding to an input signal of 24 cycles provided by an embodiment of the present invention;

[0040] Figure 6a A schematic diagram of another exemplary signal to be encoded provided by an embodiment of the present invention;

[0041] Figure 6b The exemplary electric injection device provided by the embodiment of the present invention is Figure 6a Schematic diagram of the self-pulse signal obtained after encoding the signal to be encoded shown;

[0042] Figure 7a A schematic diagram of another exemplary signal to be encoded provided by an embodiment of the present invention;

[0043] Figure 7b The exemplary electric injection device provided by the embodiment of the present invention is Figure 7a Schematic diagram of the self-pulse signal obtained after the signal to be encoded is encoded. DETAILED DESCRIPTION

[0044] The present invention will be further described in detail below with reference to specific examples, but the embodiments of the present invention are not limited thereto.

[0045] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0046] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification.

[0047] Although the present invention is described herein in conjunction with various embodiments, in the process of implementing the claimed invention, those skilled in the art can understand and implement other variations of the disclosed embodiments by reviewing the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude multiple situations. A single processor or other unit can implement several functions listed in the claims. Certain measures are recorded in different dependent claims, but this does not mean that these measures cannot be combined to produce good results.

[0048] Figure 1 FIG. 1 is a structural diagram of an electric injection encoding device based on a two-segment semiconductor laser according to an embodiment of the present invention. Figure 1 As shown, the device includes: a two-segment semiconductor laser ( Figure 1 The figure exemplarily shows a Fabry-Perot laser (FP-SA), a temperature controller (LDT), a current source (LDX), a voltage source (VS), a first field programmable gate array (FPGA) and a second field programmable gate array (FPGA), a first bias tee (Bias Tee) and a second bias tee (Bias Tee), an optical isolator (OI), a variable optical attenuator (VOA), an optical coupler (OC), a photodetector (PD), an oscilloscope (OSC) and a spectrometer (OSA).

[0049] like Figure 1As shown, the FP-SA includes a gain region (Gian) and a saturated absorption region (SA); the FP-SA is used to perform frequency encoding or spatiotemporal encoding based on the signal in the Gian region or the signal in the SA region to obtain a coded optical signal; or, it is used to perform calculations based on the signal in the Gian region and the signal in the SA region, and perform frequency encoding or spatiotemporal encoding to obtain a coded optical signal. The LDT is connected to the FP-SA through a thermoelectric cooler and is used to control the operating temperature of the FP-SA, for example, to control the operating temperature of the FP-SA at a certain constant temperature. The LDX is connected to the first bias tee for outputting a current signal; the FPGA is connected to the first bias tee for providing the electrical signal to be encoded to the Gian region; the first bias tee is connected to the Gain region of the FP-SA and is used to load the current signal and the electrical signal to be encoded to the Gian region of the FP-SA together. The VS is connected to the second bias tee to output a voltage signal. The FPGA is connected to the second bias tee to provide the encoded electrical signal to the SA region of the FP-SA. The second bias tee, connected to the SA region of the FP-SA, applies both the voltage signal and the encoded electrical signal to the SA region of the FP-SA. The OI is connected to the FP-SA and VOA, respectively, to prevent reflected light from entering the FP-SA, thereby reducing damage to the FP-SA caused by external light injection. The VOA is connected to the PD and OSC, respectively, via the OC. The VOA controls the power of the encoded optical signal from the OI to prevent excessive power from entering the PD, potentially damaging it. The OC splits the power-controlled encoded optical signal into a first optical signal and a second optical signal. The PD converts the first optical signal into an electrical signal. The OSC is connected to the PD to measure the converted electrical signal. The OSA measures the second optical signal.

[0050] like Figure 1When the LDX provides a direct current (DC) current signal to the Bias Tee and the FPGA provides an alternating current (AC) electrical signal to be encoded to the Bias Tee, the Bias Tee can superimpose the DC current signal and the AC electrical signal to be encoded, and then load the signal into the Gain region of the FP-SA. The FP-SA performs frequency encoding or spatiotemporal encoding based on the signal in the Gain region to obtain a coded optical signal. When the VS provides a direct current voltage signal to the Bias Tee and the FPGA provides an AC electrical signal to be encoded to the Bias Tee, the Bias Tee can superimpose the DC voltage signal and the AC electrical signal to be encoded, and then load the signal into the SA region of the FP-SA. The FP-SA performs frequency encoding or spatiotemporal encoding based on the signal in the SA region to obtain a coded optical signal. When the LDX provides a DC current signal to the Bias Tee and the FPGA provides an AC electrical signal to be encoded to the Bias Tee, the Bias Tee can superimpose the DC current signal and the AC electrical signal to be encoded and load it into the Gain area of ​​the FP-SA. At the same time, when the VS provides a DC voltage signal to the Bias Tee and the FPGA provides an AC electrical signal to be encoded to the Bias Tee, the Bias Tee can superimpose the DC voltage signal and the AC electrical signal to be encoded and load it into the SA area of ​​the FP-SA. The FP-SA simultaneously calculates and performs frequency encoding or space-time encoding based on the signals in the Gain area and the SA area to obtain a coded optical signal. After the FP-SA generates a coded signal, it transmits the coded optical signal to the OI. The coded optical signal is then transmitted to the VOA via the OI. The VOA attenuates the power of the received coded optical signal and transmits the attenuated coded optical signal to the OC. The OC divides the attenuated coded optical signal into a first optical signal and a second optical signal. The first optical signal is transmitted to the PD and converted into an electrical signal by the PD. The OSC measures the self-pulse signal of the converted electrical signal to obtain the self-pulse signal of the converted electrical signal, thereby measuring the self-pulse of the coded signal. The second optical signal is transmitted to the OSA, which performs spectral measurement on the second optical signal, thereby measuring the spectrum of the coded signal.

[0051] In some embodiments, the OC may equally divide the power-attenuated coded optical signal into the first optical signal and the second optical signal; in some embodiments, the OC may also divide the power-attenuated coded optical signal into the first optical signal and the second optical signal using a 28 ratio.

[0052] In some embodiments, the two FPGAs correspondingly connected to the two Bias Tees may provide the same electrical signal or different electrical signals.

[0053] The embodiment of the present invention further provides an electric injection coding method based on a two-segment semiconductor laser, and the method is implemented based on the above-mentioned electric injection coding device, such as Figure 2 As shown, the method includes:

[0054] S101, using a temperature controller to control the operating temperature of the two-segment semiconductor laser to a preset temperature.

[0055] S102. Using a voltage source and a current source, input a voltage value and a current value into the electric injection encoding device each time. At the same time, using a first field programmable logic gate array and / or a second field programmable logic gate array, input a test electrical signal into the gain region and / or the saturated absorption region each time, and measuring the encoding signal each time through an oscilloscope and a spectrometer to obtain each measurement result; the voltage value and current value input each time are a set of driving values.

[0056] Here, the test electrical signal is any electrical signal output by the FPGA.

[0057] Here, the voltage value input each time may be a preset voltage value, and the current value input each time may be a preset current value.

[0058] In the embodiment of the present invention, the measurement result obtained each time may include: the self-pulse frequency of the coded optical signal and the spectrum of the coded optical signal.

[0059] Here, when the oscilloscope measures the self-pulsation signal of the coded optical signal, the self-pulsation frequency of the coded optical signal can be obtained by counting the time interval between every two adjacent peak pulses in the self-pulsation signal and determining the frequency of the self-pulsation signal based on the time interval.

[0060] S103 , when multiple sets of driving values ​​are input to the electrical injection encoding device, each set of driving values ​​and corresponding measurement results are obtained; at least one current value or voltage value is different between any two sets of driving values.

[0061] Here, the number of the multiple sets of driving values ​​can be set according to actual needs, for example, 30 sets or 40 sets, etc., and this is not limited in the embodiment of the present invention. Among the multiple sets of driving values, some driving values ​​have different voltage values ​​but the same current values, and some driving values ​​have different current values ​​but the same voltage values.

[0062] In some embodiments, a voltage value and a current value can be input into the electrical injection encoding device each time, and an FPGA is used to input a test electrical signal into the Gain area each time. Thus, while multiple sets of driving values ​​are input into the electrical injection encoding device, multiple electrical signals are also input into the Gain area to perform frequency encoding or space-time encoding each time.

[0063] In some embodiments, a voltage value and a current value can be input into the electrical injection encoding device each time, and an FPGA is used to input a test electrical signal into the SA area each time. Thus, while multiple sets of driving values ​​are input into the electrical injection encoding device, multiple electrical signals are also input into the SA area to perform frequency encoding or space-time encoding each time.

[0064] In some embodiments, a voltage value and a current value can be input into the electrical injection encoding device each time, and two FPGAs are used to input electrical signals into the Gain area and the SA area respectively each time. Therefore, while inputting multiple sets of driving values ​​into the electrical injection encoding device, multiple electrical signals are also input into the Gain area and the SA area to perform frequency encoding or space-time encoding each time.

[0065] Here, when multiple sets of driving values ​​are input to the electrical injection encoding device and multiple test electrical signals are input to the gain region, each set of driving values ​​and corresponding measurement results can be obtained when the test electrical signal is provided to the gain region.

[0066] Here, when multiple sets of driving values ​​are input into the electrical injection encoding device and multiple test electrical signals are input into the saturated absorption region at the same time, each set of driving values ​​in the multiple sets of driving values ​​and the corresponding measurement results can be obtained when the test electrical signal is provided to the saturated absorption region.

[0067] Here, when multiple sets of driving values ​​are input into the electrical injection encoding device and multiple test electrical signals are input into the gain region and the saturated absorption region at the same time, each set of driving values ​​in the multiple sets of driving values ​​and the corresponding measurement results can be obtained when the test electrical signals are provided to both the gain region and the saturated absorption region.

[0068] S104. Analyze the interaction between the voltage value and the range length of the current value based on each of the multiple drive value groups and the corresponding measurement results, and obtain the influence of the voltage value and the current value on the measurement results; the range length of the current value is the range of the current value that supports the electrical injection encoding device to perform signal encoding.

[0069] In the embodiment of the present invention, the interaction relationship between the two may be a first interaction relationship between the two, a second interaction relationship between the two, or a third interaction relationship between the two; correspondingly, the measurement result may be a first measurement result, a second measurement result, or a third measurement result.

[0070] In some embodiments, a first interaction relationship between the voltage value and the range length of the current value can be obtained by analyzing each group of drive values ​​in the multiple groups of drive values ​​and the corresponding first measurement result when a test electrical signal is provided to the gain region, as well as a first influence relationship between the voltage value and the current value on the measurement result can be obtained.

[0071] In some embodiments, based on each group of driving values ​​in multiple groups of driving values ​​and the corresponding second measurement results when a test electrical signal is provided to the saturated absorption region, a second interaction relationship between the voltage value and the range length of the current value is analyzed to obtain a second influence relationship between the voltage value and the current value on the measurement result.

[0072] In some embodiments, a third interaction relationship between the voltage value and the range length of the current value can be obtained by analyzing each group of driving values ​​in the multiple groups of driving values ​​and the corresponding third measurement results when a test electrical signal is provided to both the gain region and the saturated absorption region, as well as a third influence relationship between the voltage value and the current value on the measurement result can be obtained.

[0073] In an embodiment of the present invention, a coordinate point can be formed in an XY coordinate system with the current value in each group of driving values ​​as the X coordinate and the measurement result corresponding to each group of driving values ​​as the Y coordinate, and the voltage value in each group of driving values ​​is used as the voltage value corresponding to the coordinate point; by connecting the coordinate points with the same corresponding voltage values, a relationship curve between the current value, the voltage value and the measurement result is obtained; the relationship curve is used to represent the interaction relationship between the magnitude of the voltage value and the range length of the current value, and is also used to represent the influence of the magnitude of the voltage value and the magnitude of the current value on the measurement result.

[0074] For example, a coordinate point can be formed in an XY coordinate system with the current value in each group of drive values ​​as the X coordinate and the self-pulse frequency in the measurement result corresponding to each group of drive values ​​as the Y coordinate, and the voltage value in each group of drive values ​​is used as the voltage value corresponding to the coordinate point; by connecting the coordinate points with the same corresponding voltage value, a relationship curve between the current value, voltage value and self-pulse frequency is obtained; and, in an XY coordinate system, a coordinate point can be formed with the current value in each group of drive values ​​as the X coordinate and the spectrum in the measurement result corresponding to each group of drive values ​​as the Y coordinate, and the voltage value in each group of drive values ​​is used as the voltage value corresponding to the coordinate point; by connecting the coordinate points with the same corresponding voltage value, a relationship curve between the current value, voltage value and spectrum is obtained.

[0075] S105. Select a driving voltage value and a driving current value according to the interaction and influence relationship between the two, input the driving voltage value, the driving current value, and the electrical signal to be encoded into an electrical injection encoding device for frequency encoding or spatiotemporal encoding to obtain an encoding result.

[0076] In an embodiment of the present invention, when the interaction relationship between the voltage value and the range length of the current value, as well as the influence relationship between the voltage value and the current value on the measurement result are obtained through the above steps S101 to S104, appropriate driving voltage and driving current values ​​can be selected based on these obtained relationships and the characteristics of the desired coded optical signal, and the driving voltage and driving current values ​​are input into the electrical injection coding device. At the same time, the electrical signal to be encoded is input into the electrical injection coding device through the FPGA, and spatiotemporal encoding or frequency encoding is performed by the laser in the device to obtain the encoded signal.

[0077] In some embodiments, when the first action relationship and the first influence relationship are obtained, the driving voltage value and the driving current value can be selected according to the first action relationship and the first influence relationship, and the driving voltage value is input into the second bias device through the voltage source, and the driving current value is input into the first bias device through the current source, and the electrical signal to be encoded is input into the first bias device through the FPGA to perform frequency encoding or space-time encoding to obtain the encoding result.

[0078] In some embodiments, when the second action relationship and the second influence relationship are obtained, the driving voltage value and the driving current value can be selected according to the second action relationship and the second influence relationship, and the driving voltage value is input into the second bias device through the voltage source, and the driving current value is input into the first bias device through the current source, and the electrical signal to be encoded is input into the second bias device through the FPGA to perform frequency encoding or space-time encoding to obtain the encoding result.

[0079] In some embodiments, when the third action relationship and the third influence relationship are obtained, the driving voltage value and the driving current value can be selected according to the third action relationship and the third influence relationship, and the driving voltage value is input into the second bias device through the voltage source, and the driving current value is input into the first bias device through the current source, and the electrical signal to be encoded is input into the first bias device through an FPGA, and at the same time, the electrical signal to be encoded is input into the second bias device through another FPGA to perform frequency encoding or space-time encoding to obtain the encoding result.

[0080] The electric injection encoding device based on a two-segment semiconductor laser proposed in the present invention can provide two electric injection interfaces, and signal encoding can be performed through either of the two electric injection interfaces. Moreover, when both electric injection interfaces are used simultaneously, the two-segment semiconductor laser can calculate and encode the signal in the gain region and the signal in the saturated absorption region. Compared with encoding methods based on light injection, this eliminates the dependence of information encoding in photon pulse neural networks on external light injection and does not require additional photoelectric conversion. Moreover, compared with electric injection encoding based on vertical cavity surface emitting lasers, it not only increases the number of electric injection interfaces but also provides certain computing functions on the two-segment semiconductor laser as a light pulse neuron. In addition, due to the high power of the light output by the two-segment semiconductor laser, the encoded signal encoded by the electric injection encoding device proposed in the present invention is more convenient for information transmission after entering the pulse neural network.

[0081] The following takes an example in which an FPGA is used to provide an electrical signal to the Gain area of ​​an FP-SA, and the FP-SA performs frequency encoding and space-time encoding on the provided electrical signal, to further illustrate the method and effect of the present invention.

[0082] 1. FP-SA frequency encodes the provided electrical signal:

[0083] 1) Determine a preset set of driving values, each set of driving values ​​including a voltage value and a current value;

[0084] 2) While inputting a set of drive values ​​into the electrical injection encoding device, a test electrical signal is provided to the Gain region of the FP-SA via the FPGA, and the self-pulse signal corresponding to each set of drive values ​​is measured using an oscilloscope. By counting the time intervals between each two adjacent spike pulses in the self-pulse signal corresponding to each set of drive values, the frequency (self-pulse frequency) of the self-pulse signal corresponding to each set of drive values ​​can be obtained; Figure 3a 、 Figure 3b and Figure 3c are the self-pulse signals corresponding to any three sets of driving values ​​in the preset set of driving values, and, Figure 3a 、 Figure 3b and Figure 3c In the figure, the horizontal axis represents time (Time), the unit can be ns, and the vertical axis represents voltage (Voltage), the unit is V.

[0085] 3) By using the current value in each set of driving values ​​as the X coordinate and the self-pulse frequency value in the measurement result corresponding to each set of driving values ​​as the Y coordinate in the XY coordinate system, a coordinate point corresponding to each set of driving values ​​is obtained, and the voltage value in each set of driving values ​​is used as the voltage value corresponding to the coordinate point; by connecting the coordinate points with the same corresponding voltage value, a relationship curve between the current value, the voltage value and the self-pulse frequency (hereinafter referred to as the FP-SA frequency curve) can be obtained. The frequency curve represents the interaction between the voltage value and the range length of the current value, as well as the influence of the voltage value and the current value on the self-pulse frequency; for example, Figure 4 The frequency curve of FP-SA is shown, wherein the unit of the horizontal axis is mA and the unit of the vertical axis is GHz; Figure 4 It can be seen that when the absolute value of the voltage is small, for example, the voltage is -3.4V, the current range of the Gain region that can cause the FP-SA to self-pulse is small, only 83.5mA to 88.5mA; as the absolute value of the voltage increases, the current range of the Gain region that can cause the FP-SA to self-pulse gradually increases. For example, when the voltage is -3.6V, the current range of the Gain region that can cause the FP-SA to self-pulse is 70.5mA to 87.5mA; and as Figure 4 As shown in FIG. 1 , within a certain range, as the voltage decreases and the absolute value of the current increases, the self-pulse frequency of the output signal of the FP-SA increases.

[0086] 4) According to Figure 4 The frequency curve of FP-SA shown in FIG1 is used to select a set of driving values ​​that can cause FP-SA to have self-pulsation, and the set of driving values ​​is input into the above-mentioned electric injection encoding device, and at the same time, the Gain area of ​​FP-SA is provided with the following signal through FPGA: Figure 5a The periodic signal shown can be obtained through an oscilloscope. Figure 5b The self-pulse signal of the encoded signal shown in FIG; Figure 5b As shown, the output signal is a spike pulse of different frequencies; by measuring Figure 5b The time interval between each two adjacent spike pulses can be plotted Figure 5c , and according to Figure 5c Can be drawn Figure 5d , Figure 5d represents the frequency of different signals; among them, Figure 5c The vertical axis is the time interval (TI), in ns, and the horizontal axis is the number of spike pulses; Figure 5d The vertical axis is the frequency, and the horizontal axis is the number of spike pulses. Figure 5dAs shown, the frequency of the input signal with lower intensity is lower and is below the horizontal line; the frequency of the input signal with higher intensity is higher and is above the horizontal line. Therefore, the intensity of the input signal (the electrical signal provided by the FPGA) can be judged by the frequency of the output signal, that is, the greater the frequency of the output signal, the higher the intensity of the input signal. In addition, by counting the number of spike pulses within a specific time, the frequency of the output signal within the specific time can be determined, and based on the frequency of the output signal within the specific time, the intensity of the electrical signal to be encoded by the FP-SA can be finally determined. For example, by counting the spike pulses of the output signal corresponding to the input signal of 24 cycles, it can be obtained Figure 5e , combined with Figure 5b and Figure 5e It can be seen that a low-intensity input signal will generate 21 pulses within 10ns, while a high-intensity input signal will generate 22 or 23 pulses within 10ns. Figure 5e It can be seen that the statistics of multiple periods show that FP-SA is relatively stable when performing frequency encoding.

[0087] 2. FP-SA performs spatiotemporal encoding on the provided electrical signal:

[0088] Here, the principle of space-time coding is the same as that of frequency coding, but the periodic signal provided to the Gain area of ​​FP-SA by FPGA can be Figure 6a The input signals of different strengths and widths are shown in FIG. The output signal obtained by FP-SA after spatiotemporal encoding of the input signal is as follows: Figure 6b As shown, according to Figure 6a and Figure 6b As shown, for signals with lower intensity, the FP-SA as a light pulse neuron has no response (no spatiotemporal encoding is performed). For signals with higher intensity, the light pulse neuron will generate pulses at the corresponding time. In this way, the threshold characteristics of the photon pulse neuron are realized, so that according to the different input signals, spike pulses at different times can be encoded to complete the spatiotemporal encoding of the photon pulse neural network.

[0089] In addition, according to Figure 6a and Figure 6b , or, according to Figure 7a The input signal shown and Figure 7b From the corresponding output signal, it can be seen that the longer the input signal to be encoded lasts, the longer the duration of the photon pulse neuron encoding, the more spike pulses are obtained by encoding, and the moment when the spike pulse appears corresponds to the moment when the signal to be encoded appears, thereby realizing corresponding encoding of the signal to be encoded input at each moment.

[0090] The embodiments of the present invention can eliminate the dependence of information encoding in photon pulse neural networks on external light injection, and cooperate with FPGA during encoding without the need for additional photoelectric conversion; electrical signals can be injected into the gain area or absorption area of ​​FP-SA separately, or can be injected into the gain area and saturation absorption area at the same time. When electrical signals are injected into the gain area and saturation absorption area at the same time, FP-SA can perform signal calculation and encoding simultaneously, realizing a multifunctional neuron.

[0091] The above is a further detailed description of the present invention in conjunction with specific preferred embodiments, and the specific implementation of the present invention should not be considered to be limited to these descriptions. For those skilled in the art of the present invention, without departing from the concept of the present invention, several simple deductions or substitutions can be made, which should be considered to fall within the scope of protection of the present invention.

Claims

1. An electric injection encoding device based on a two-segment semiconductor laser, characterized in that: include: A two-segment semiconductor laser, a temperature controller, a current source, a voltage source, a first field programmable logic gate array and a second field programmable logic gate array, a first bias device and a second bias device, an optical isolator, an adjustable optical attenuator, an optical coupler, a photodetector, an oscilloscope, and a spectrometer; The two-segment semiconductor laser comprises a gain region and a saturation absorption region; is used to perform frequency encoding or spatiotemporal encoding based on a signal in the gain region or a signal in the saturation absorption region to obtain a coded optical signal; and performs calculation based on a signal in the gain region and a signal in the saturation absorption region, and performs frequency encoding or spatiotemporal encoding to obtain a coded optical signal; The temperature controller is connected to the two-segment semiconductor laser and is used to control the operating temperature of the two-segment semiconductor laser; The current source is connected to the first bias device and is used to output a current signal; The first field programmable logic gate array is connected to the first bias device and is used to provide the electrical signal to be encoded to the gain region; The first biaser is connected to the gain region and is used to load the current signal and the electrical signal to be encoded into the gain region; The voltage source is connected to the second bias device and is used to output a voltage signal; The second field programmable logic gate array is connected to the second bias device and is used to provide the electrical signal to be encoded to the saturated absorption region; The second biaser is connected to the saturated absorption region and is used to load the voltage signal and the coded electrical signal to the saturated absorber region; The optical isolator is connected to the two-segment semiconductor laser and the adjustable optical attenuator respectively, and is used to prevent reflected light from entering the two-segment semiconductor laser; The adjustable optical attenuator is connected to the photodetector and the spectrometer respectively through the optical coupler; the adjustable optical attenuator is used to control the power of the coded optical signal from the optical isolator; the optical coupler is used to divide the coded optical signal after power control into a first optical signal and a second optical signal; The photoelectric detector is used to convert the first optical signal into an electrical signal; The oscilloscope is connected to the photodetector and is used to measure the converted electrical signal; The spectrometer is used to measure the second optical signal.

2. The electric injection encoding device based on a two-segment semiconductor laser according to claim 1, characterized in that: The two-segment semiconductor laser is a Fabry-Perot laser.

3. The electric injection encoding device based on a two-segment semiconductor laser according to claim 1, characterized in that: The first field programmable logic gate array and the second field programmable logic gate array may provide the same electrical signal or different electrical signals.

4. An electric injection coding method based on a two-segment semiconductor laser, characterized in that: The method is implemented based on the electric injection encoding device based on a two-segment semiconductor laser according to any one of claims 1 to 3, and the method comprises: The temperature controller is used to control the operating temperature of the two-segment semiconductor laser to a preset temperature; Using the voltage source and the current source, a voltage value and a current value are inputted into the electrical injection encoding device each time. Simultaneously, using the first field programmable logic gate array and / or the second field programmable logic gate array, a test electrical signal is inputted into the gain region and / or the saturated absorption region each time. The coded signal is measured each time by the oscilloscope and the spectrometer to obtain a measurement result each time. The voltage value and current value inputted each time are a set of driving values. When multiple sets of driving values ​​are input into the electrical injection encoding device, each set of driving values ​​and corresponding measurement results are obtained; at least one current value or voltage value is different between any two sets of driving values; Based on each of the plurality of drive value groups and the corresponding measurement results, analyzing and obtaining an interaction relationship between the voltage value and the range length of the current value, and obtaining an influence relationship between the voltage value and the current value on the measurement results; the range length of the current value is the size of the range of the current value that supports the electrical injection encoding device to perform signal encoding; According to the action relationship and the influence relationship between the two, the driving voltage value and the driving current value are selected, and the driving voltage value, the driving current value, and the electrical signal to be encoded are input into the electrical injection encoding device for frequency encoding or space-time encoding to obtain the encoding result.

5. The electric injection coding method based on a two-segment semiconductor laser according to claim 4, characterized in that: When a plurality of sets of driving values ​​are input into the electrical injection encoding device, obtaining each set of driving values ​​and a corresponding measurement result in the plurality of sets of driving values ​​includes any one of the following: When multiple sets of driving values ​​are input to the electrical injection encoding device and multiple test electrical signals are input to the gain region, a corresponding measurement result of each set of driving values ​​in the multiple sets of driving values ​​is obtained when the test electrical signal is provided to the gain region; When multiple sets of driving values ​​are input to the electrical injection encoding device and multiple test electrical signals are input to the saturated absorption region, each set of driving values ​​in the multiple sets of driving values ​​and a corresponding measurement result are obtained when the test electrical signal is provided to the saturated absorption region; When multiple sets of driving values ​​are input into the electrical injection encoding device and multiple test electrical signals are input into both the gain region and the saturated absorption region, each set of driving values ​​in the multiple sets of driving values ​​and the corresponding measurement results are obtained when the test electrical signals are provided to both the gain region and the saturated absorption region.

6. The electric injection coding method based on a two-segment semiconductor laser according to claim 4, characterized in that: The analyzing, based on each set of driving values ​​in the plurality of sets of driving values ​​and the corresponding measurement results, to obtain an action relationship between the magnitude of the voltage value and the range length of the current value, and to obtain an influence relationship between the magnitude of the voltage value and the magnitude of the current value on the measurement results, includes: In the XY coordinate system, a coordinate point is formed with the current value in each set of driving values ​​as the X coordinate and the measurement result corresponding to each set of driving values ​​as the Y coordinate, and the voltage value in each set of driving values ​​is used as the voltage value corresponding to the coordinate point; By connecting the coordinate points with the same corresponding voltage values, a relationship curve between the current value, the voltage value and the measurement result is obtained; the relationship curve is used to represent the interaction between the magnitude of the voltage value and the range length of the current value, and is also used to represent the influence of the magnitude of the voltage value and the magnitude of the current value on the measurement result.

7. The electric injection coding method based on a two-segment semiconductor laser according to claim 5, characterized in that: The action relationship between the two is a first action relationship between the two, a second action relationship between the two, or a third action relationship between the two; correspondingly, the measurement result is a first measurement result, a second measurement result, or a third measurement result; The analyzing, based on each set of driving values ​​in the plurality of sets of driving values ​​and the corresponding measurement results, to obtain an interaction relationship between the voltage value and the range length of the current value, and to obtain an influence relationship between the voltage value and the current value on the measurement results, includes any of the following: analyzing, based on each of the plurality of drive value groups and the corresponding first measurement result when a test electrical signal is provided to the gain region, the first interaction relationship between the voltage value and the current value range length, and the first influence relationship between the voltage value and the current value on the measurement result; analyzing, based on each of the plurality of drive value groups and the corresponding second measurement result when a test electrical signal is provided to the saturated absorption region, the second interaction relationship between the voltage value and the current value range length, and the second influence relationship between the voltage value and the current value on the measurement result; Based on the analysis of each group of driving values ​​in the multiple groups of driving values ​​and the corresponding third measurement results when the test electrical signals are provided to the gain region and the saturated absorption region, the third interaction relationship between the voltage value and the range length of the current value is obtained, as well as the third influence relationship between the voltage value and the current value on the measurement result is obtained.

8. The electric injection coding method based on a two-segment semiconductor laser according to claim 7, characterized in that: The driving voltage value and the driving current value are selected according to the action relationship and the influence relationship between the two, and the driving voltage value, the driving current value, and the electrical signal to be encoded are input into the electrical injection encoding device for frequency encoding or spatiotemporal encoding to obtain an encoding result, including any one of the following: selecting a driving voltage value and a driving current value based on the first action relationship between the two and the first influence relationship, inputting the driving voltage value into the second bias device via the voltage source, inputting the driving current value into the first bias device via the current source, and inputting the electrical signal to be encoded into the first bias device via the first field programmable logic gate array to perform frequency encoding or spatiotemporal encoding to obtain an encoding result; selecting a driving voltage value and a driving current value based on the second action relationship between the two and the second influence relationship, inputting the driving voltage value into the second bias device via the voltage source, inputting the driving current value into the first bias device via the current source, and inputting the electrical signal to be encoded into the second bias device via the second field programmable logic gate array to perform frequency encoding or spatiotemporal encoding to obtain an encoding result; A driving voltage value and a driving current value are selected based on the third action relationship between the two and the third influence relationship. The driving voltage value is input into the second bias switch via the voltage source, and the driving current value is input into the first bias switch via the current source. The electrical signal to be encoded is input into the first bias switch via the first field programmable logic gate array, and the electrical signal to be encoded is input into the second bias switch via the second field programmable logic gate array. Frequency encoding or spatiotemporal encoding is performed to obtain an encoding result.

Images