JP7505698B2 - Signal processing device and signal processing method - Google Patents

Description

Application of Article 30, Paragraph 2 of the Patent Act Published at OPIC2019 on April 22, 2019 Published at Unconventional Computation and Natural Computation 2019 on June 3, 2019 Published at the 13th Japan-Finland Joint Symposium on Optical Engineering on August 26, 2019 Published at the 80th Applied Physics Society Academic Lecture Meeting, Fall 2019 on September 18, 2019 Published at the Optical Society of Japan Annual Academic Lecture Meeting on December 2, 2019 Published at Optical Review on March 6, 2020

The present invention relates to a signal processing device and a signal processing method.

In recent years, progress has been made in the development of computational elements that are tiny and can perform calculations with low power consumption by implementing signal processing that utilizes fluorescence resonance energy transfer (FRET) between fluorescent particles (see, for example, non-patent document 1).

T. Nishimura et al., Appl. Phys. Lett. 101, 233703 (2012)

However, current microfabrication technology has the problem that it is difficult to physically implement large-scale FRET pathways as designed.

The present invention aims to provide a signal processing device and a signal processing method in which a large-scale FRET path is physically implemented.

A signal processing device according to one embodiment of the present invention includes an optical energy network including a plurality of fluorescent particles and configured so that the path of energy movement through the plurality of fluorescent particles is determined autonomously and randomly; a signal input unit that excites at least a portion of the plurality of fluorescent particles in response to an input signal to be input to the optical energy network; a signal readout unit that reads out an optical signal from the optical energy network after exciting the portion of the fluorescent particles; and an acquisition unit that acquires a calculation result from the optical energy network based on the optical signal read out by the signal readout unit.

A signal processing method according to one aspect of the present invention excites at least a portion of a plurality of fluorescent particles in response to an input signal to an optical energy network configured to autonomously and randomly determine an energy movement path through the plurality of fluorescent particles, reads an optical signal from the optical energy network after exciting the portion of the fluorescent particles, and obtains a calculation result from the optical energy network based on the read optical signal.

According to the above aspect, signal processing can be performed using an optical energy network with a large-scale FRET path.

1 is a schematic diagram showing a configuration of a signal processing device according to an embodiment of the present invention; FIG. 2 is a schematic diagram showing the configuration of a FRET network portion. 10 is a flowchart illustrating a procedure of a process executed by the signal processing device according to the present embodiment in a learning phase. 10 is a flowchart illustrating a procedure of a process executed by a signal processing device according to the present embodiment in an operation phase. FIG. 11 is a schematic diagram showing the configuration of a FRET network section in embodiment 2. FIG. 13 is a diagram illustrating the results of examining the diversity of fluorescence intensity. FIG. 13 is a diagram illustrating the results of examining the diversity of fluorescence lifetimes. FIG. 13 is a diagram illustrating the results of examining the diversity of fluorescence lifetimes due to differences in excitation intensity.

Hereinafter, the present invention will be described in detail with reference to the drawings showing embodiments thereof.
(Embodiment 1)
Fig. 1 is a schematic diagram showing the configuration of a signal processing device according to this embodiment, and Fig. 2 is a schematic diagram showing the configuration of a FRET network section 12. The signal processing device according to this embodiment is a device that performs calculations by utilizing an optical energy network through which signals are transmitted by FRET, and includes a signal input section 11, a FRET network section 12, a signal readout section 13, a control section 14, a storage section 15, and an input/output section 16.

The FRET network section 12 is composed of a nanostructure 120 in which a plurality of fluorescent particles 121, 121, ..., 121 are randomly arranged in a solid, liquid, or amorphous state. Here, the fluorescent particles 121 may be fluorescent molecules or may be fluorescent materials that function as quantum dots. The nanostructure 120 may have a laminated structure. The number of fluorescent particles 121 arranged in the nanostructure 120 is adjusted so that the distance between adjacent fluorescent particles 121, 121 is 10 nm or less. For example, about 10 ⁶ fluorescent particles 121 are arranged in a nanostructure 120 having a volume of about 1 μm ³ . For the fluorescent particles 121, Cy3 with an excitation peak of 550 nm and a fluorescence peak of 570 nm, Cy5 with an excitation peak of 649 nm and a fluorescence peak of 670 nm, or the like is used.

Since the fluorescent particles 121 are randomly arranged in the nanostructure 120, the distance between any two fluorescent particles 121, 121 varies, but FRET can occur when the distance is approximately 10 nm or less. The conditions for FRET occurrence include not only the distance between the two fluorescent particles 121, 121, but also a large overlap between the emission spectrum on the donor side and the absorption spectrum on the acceptor side, the two fluorescent particles 121, 121 being appropriately oriented, the fluorescence quantum yield on the donor side being large, and the absorption intensity of the acceptor being large. By using such a nanostructure 120, the path along which energy moves between the fluorescent particles 121, 121 in the FRET network section 12 is determined autonomously and randomly. That is, the fluorescent particles 121, 121, ..., 121 in the nanostructure 120 construct an optical energy network through which signals are transmitted by FRET.

The signal input unit 11 excites at least a part of the fluorescent particles 121, 121, ..., 121 in response to an input signal. The input signal is a signal to be calculated in the signal processing device, and may be generated inside the control unit 14 or may be input from the outside through the input/output unit 16. The signal input unit 11 controls the wavelength and time-series light intensity of the light to be irradiated to the nanostructure 120 according to the modulation parameters adjusted in response to the input signal, and irradiates a specific irradiation area of the nanostructure 120 with light whose wavelength and time-series light intensity are controlled (hereinafter referred to as excitation light). The modulation method is arbitrary, and it is sufficient that the modulation is performed according to a preset rule. In addition, the irradiation area to which the excitation light is irradiated is set in advance. When the excitation light is irradiated, the energy state of the fluorescent particles 121, 121, ..., 121 included in the irradiation area transitions, for example, from the ground state to the excited state.

When some of the fluorescent particles 121, 121, ..., 121 transition to an excited state, FRET occurs between two fluorescent particles 121, 121 that satisfy the above-mentioned conditions. That is, excitation energy is transferred from the fluorescent particle 121 on the donor side to the fluorescent particle 121 on the acceptor side. Such FRET between two fluorescent particles 121, 121 occurs one after another in the fluorescent particles 121, 121, ..., 121 that constitute the nanostructure 120. Therefore, after the signal input unit 11 irradiates the nanostructure 120 with excitation light, the energy state of each fluorescent particle 121, 121, ..., 121 that constitutes the nanostructure 120 autonomously changes due to FRET from other fluorescent particles 121 that satisfy the generation condition, and changes to a state that represents the calculation result.

The signal readout unit 13 reads out a signal from the FRET network unit 12 after the excitation light is irradiated from the signal input unit 11 and the energy state of each fluorescent particle 121, 121, ..., 121 changes to a state that represents the calculation result (for example, after about 10 ^-12 seconds have passed since the excitation light was irradiated). The signal readout unit 13 can read out a signal from the FRET network unit 12 by performing wavelength-resolved measurement/time-resolved measurement using a spectrometer. A known method can be used to read out a signal from the FRET network unit 12. In addition, the readout region from which the signal readout unit 13 reads out a signal may be set to the entire region of the nanostructure 120, or may be set to a partial region of the nanostructure 120. The signal readout by the signal readout unit 13 is a signal of a fluorescence spectrum obtained from the fluorescent particles 121, 121, ..., 121 included in the readout region. The signal readout section 13 outputs the signal read out from the FRET network section 12 to the control section 14 .

The control unit 14 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The CPU of the control unit 14 executes various computer programs stored in the ROM or the storage unit 15 to control the operation of each part of the hardware and execute various calculations, causing the entire device to function as the signal processing device of the present application. The RAM of the control unit 14 temporarily stores data, etc., generated during the execution of the various computer programs.

The calculations executed by the control unit 14 include a calculation to learn the relationship between the signal read by the signal readout unit 13 from the FRET network unit 12 and the calculation result by the FRET network unit 12, and a calculation to derive the calculation result from the signal read out by the FRET network unit 12 using the result of the learning calculation. These calculations will be described in detail later.

The control unit 14 is not limited to the above configuration, and may be any processing circuit including one or more CPUs, a multi-core CPU, a GPU (Graphics Processing Unit), a microcomputer, etc. The control unit 14 may also have functions such as a clock that outputs date and time information, a timer that measures the elapsed time from when an instruction to start measurement is given to when an instruction to end measurement is given, and a counter that counts numbers.

The storage unit 15 includes a storage device using a memory, a hard disk drive (HDD), etc. The storage unit 15 stores various computer programs executed by the control unit 14, data necessary for executing the computer programs, calculation results by the control unit 14, etc. The computer programs stored in the storage unit 15 include a learning program for learning the relationship between the signal read by the signal readout unit 13 from the FRET network unit 12 and the calculation result to be obtained as the output for the input signal, and a calculation program for converting the signal read out by the signal readout unit 13 from the FRET network unit 12 into a calculation result based on the learning result, etc.

The computer program stored in the memory unit 15 may be provided by a non-transitory recording medium on which the computer program is recorded in a readable manner. The recording medium is, for example, a portable memory such as a CD-ROM, a USB memory, a Secure Digital (SD) card, a micro SD card, or a CompactFlash (registered trademark). In this case, the control unit 14 can read the computer program from the recording medium using a reading device (not shown) and install the read computer program in the memory unit 15. The computer program stored in the memory unit 15 may also be provided by communication. In this case, the control unit 14 can acquire the computer program by communicating with an external server (not shown) and install the acquired computer program in the memory unit 15.

The input/output unit 16 has an interface for inputting and outputting various information. The input/output unit 16 may have an interface for connecting an input device such as a keyboard or a mouse, and may receive user instructions through the connected input device. The input/output unit 16 may also have an interface for connecting an external computer, and may input a signal output from the connected external computer. Furthermore, the input/output unit 16 may have an interface for connecting an output device such as a liquid crystal display or an organic EL (Electro-Luminescence) display, and may output the results of calculations by the control unit 14 to the output device.

The operation of the signal processing device according to this embodiment will be described below.
3 is a flow chart for explaining the procedure of the process executed by the signal processing device according to this embodiment in the learning phase. In an initial state in which no learning has been performed, the control unit 14 of the signal processing device executes the following process to learn the relationship between the signal read out from the FRET network unit 12 by the signal readout unit 13 and the calculation result by the FRET network unit 12. Furthermore, when a user's instruction is received through the input/output unit 16, the control unit 14 may execute the following process to re-learn the relationship between the signal read out from the FRET network unit 12 by the signal readout unit 13 and the calculation result by the FRET network unit 12.

When an input signal to be calculated is input, the control unit 14 excites at least a portion of the fluorescent particles 121, 121, ..., 121 that constitute the FRET network unit 12 by controlling the signal input unit 11 (step S101). The input signal to be calculated may be input to the control unit 14 through the input/output unit 16, or may be generated inside the control unit 14. The control unit 14 generates modulation parameters according to the input signal to be calculated, and controls the wavelength and time-series light intensity of the excitation light to be irradiated according to the generated modulation parameters. The signal input unit 11 irradiates a predetermined irradiation area of the FRET network unit 12 with excitation light whose wavelength and time-series light intensity are controlled based on the modulation parameters generated by the control unit 14.

By irradiating the excitation light, the energy state of the fluorescent particles contained in the irradiated area transitions, for example, from the ground state to an excited state. When some of the fluorescent particles 121, 121, ..., 121 transition to the excited state, FRET occurs between two fluorescent particles 121, 121 that satisfy the generation condition. FRET between the fluorescent particles 121, 121 occurs one after another in the fluorescent particles 121, 121, ..., 121 that make up the nanostructure 120, and is transmitted to the entire nanostructure 120. In other words, the energy state of the fluorescent particles 121, 121, ..., 121 changes autonomously due to FRET, and changes to a state that represents the calculation result.

Next, the signal readout unit 13 reads out the signal from the FRET network unit 12 (step S102). This signal readout may be performed after the timing when the energy state of the fluorescent particles 121, 121, ..., 121 changes to a state that represents the calculation result (for example, after the timing when about 10 ^-12 seconds have passed since the irradiation of the excitation light). The signal readout unit 13 can read out the signal from the FRET network unit 12 by performing wavelength-resolved measurement/time-resolved measurement using a spectrometer. The signal readout from the FRET network unit 12 is output to the control unit 14.

Next, the control unit 14 calculates a cost function that is set based on the signal read by the signal read unit 13 and a teacher signal that indicates an ideal output for the input signal (step S103). The cost function L({k _i }) is expressed by, for example, the following equation.

Here, x _i (i=1, 2, 3, ...) represents the input signal, and y _i (t, λ) represents the signal read out by the signal readout unit 13 from the FRET network unit 12. f(x _i ) represents a teacher signal that shows an ideal output for the input signal x _i . Furthermore, the calculation result corresponding to the signal y _i (t, λ) can be described as F(y _i (t, λ), {k _i }). Here, F(...) is a function for converting the signal y _i (t, λ) into the calculation result. F(...) may be a linear function or a nonlinear function. {k _i } is a parameter used when converting the signal y _i (t, λ) into the calculation result using the function F(...). The parameter {k _i } is adjusted by learning.

Note that the cost function is not limited to the formula in equation 1, and for example, a cost function expressed by the following formula may be used.

Next, the control unit 14 judges whether the calculated value of the cost function L({k _i }) is equal to or smaller than a threshold ε (step S104). Here, the threshold ε is a threshold for judging whether the cost function L({k _i }) has converged, and is set to an appropriate small value.

When it is determined that the calculated value of the cost function L is not equal to or less than the threshold value ε (S104: NO), the control unit 14 changes the parameter {k _i } (step S105). For example, the control unit 14 may hold the value of the cost function L calculated in step S103, and change the parameter {x _i } using a gradient method so that the value L _n of the currently calculated cost function L({k _i }) is smaller than the value L _n-1 of the previously calculated cost function L({k _i }). After changing the parameter {x _i }, the control unit 14 returns the process to step S101.

On the other hand, if it is determined that the calculated value of the cost function L({k _i }) is equal to or less than the threshold ε (S104: YES), the control unit 14 determines the value of the parameter {k _i } at that time as the parameter {k _i } to be used in the operation phase (step S106). The control unit 14 stores the determined {k _i } in the storage unit 15, and ends the processing according to this flowchart.

Figure 4 is a flowchart explaining the procedure of the processing executed by the signal processing device of this embodiment in the operation phase. In the operation phase after learning is completed, the control unit 14 of the signal processing device executes the following processing to obtain the calculation result of the FRET network unit 12 from the signal read out from the signal readout unit 13.

When an input signal to be calculated is input, the control unit 14 excites at least a part of the fluorescent particles 121, 121, ..., 121 constituting the FRET network unit 12 by controlling the signal input unit 11, as in the learning phase (step S201). The input signal to be calculated may be input to the control unit 14 through the input/output unit 16, or may be generated inside the control unit 14. The control unit 14 generates modulation parameters according to the input signal to be calculated, and controls the wavelength and time-series light intensity of the excitation light to be irradiated according to the generated modulation parameters. The signal input unit 11 irradiates a predetermined irradiation area of the FRET network unit 12 with excitation light whose wavelength and time-series light intensity are controlled based on the modulation parameters generated by the control unit 14.

By irradiating the excitation light, the energy state of the fluorescent particles contained in the irradiated area transitions, for example, from the ground state to an excited state. When some of the fluorescent particles 121, 121, ..., 121 transition to the excited state, FRET occurs between two fluorescent particles 121, 121 that satisfy the generation condition. FRET between the fluorescent particles 121, 121 occurs one after another in the fluorescent particles 121, 121, ..., 121 that make up the nanostructure 120, and is transmitted to the entire nanostructure 120. In other words, the energy state of the fluorescent particles 121, 121, ..., 121 changes autonomously due to FRET, and changes to a state that represents the calculation result.

Next, the signal readout unit 13 reads out the signal from the FRET network unit 12 (step S202). This signal readout may be performed after the timing when the energy state of the fluorescent particles 121, 121, ..., 121 changes to a state that represents the calculation result (for example, after the timing when about 10 ^-12 seconds have passed since the irradiation of the excitation light). The signal readout unit 13 can read out the signal from the FRET network unit 12 by performing wavelength-resolved measurement/time-resolved measurement using a spectrometer. The signal readout from the FRET network unit 12 is output to the control unit 14.

Next, the control unit 14 derives the calculation result by the FRET network unit 12 based on the signal output from the signal readout unit 13 (step S203). At this time, the control unit 14 reads out the parameters {k _i } determined in the learning phase, and inputs the signal y _i (t, λ) to the function F(y, {k _i }) described by the parameters {k _i }, thereby deriving the calculation result by the FRET network unit 12. The control unit 14 outputs the derived calculation result via the input/output unit 16 (step S204).

As described above, in the signal processing device of this embodiment, the physical properties of the fluorescent particles 121 possessed by the FRET network portion 12 are directly used for calculations, thereby making it possible to achieve smaller size and lower power consumption compared to conventional computing devices that use electrical signals.

(Embodiment 2)
In the second embodiment, a signal processing device that utilizes the spatial dynamics of the FRET network section 12 will be described.

Figure 5 is a schematic diagram showing the configuration of the FRET network section 12 in the second embodiment. The nanostructure 120 of the FRET network section 12 in the second embodiment has a laminated structure in which the first layer 120A to the third layer 120C are laminated. The first layer 120A contains fluorescent particles 121A having a fluorescent wavelength of, for example, 540 nm, the second layer 120B contains fluorescent particles 121B having a fluorescent wavelength of, for example, 580 nm, and the third layer 120C contains fluorescent particles 121C having a fluorescent wavelength of, for example, 620 nm. Each of these layers 120A to 120C is formed by spin-coating a polymer solvent containing the fluorescent particles 121A to 121C on a cover glass and then curing it. The fluorescent particles 121A are randomly arranged inside the first layer 120A. Similarly, fluorescent particles 121B are randomly arranged inside second layer 120B, and fluorescent particles 121C are randomly arranged inside third layer 120C. Although the three-layered nanostructure 120 has been described as an example in FIG. 5, nanostructure 120 may be a laminate of two layers or four or more layers. Nanostructure 120 may not have a layered structure, and may have a structure in which fluorescent particles 121A, 121B, and 121C are randomly arranged. The types of fluorescent particles randomly arranged are not limited to three types, and may be two types or four or more types.

In the nanostructure 120 shown in FIG. 5, energy transfer from fluorescent particle 121A with a fluorescent wavelength of 540 nm to fluorescent particle 121B with a fluorescent wavelength of 580 nm is permitted, and transfer in the reverse direction is prohibited. Similarly, energy transfer from fluorescent particle 121B with a fluorescent wavelength of 580 nm to fluorescent particle 121C with a fluorescent wavelength of 620 nm is permitted, and transfer in the reverse direction is prohibited. Therefore, when excitation light is irradiated onto the nanostructure 120 from the first layer 120A side to excite fluorescent particle 121A in the first layer 120A, the energy state is transmitted in the order of fluorescent particle 121A in the first layer 120A, fluorescent particle 121B in the second layer 120B, and fluorescent particle 121C in the third layer 120C. The fluorescence emitted by fluorescent particle 121C can be observed from the third layer 120C side of the nanostructure 120.

Since the fluorescent particles 121A-121C are randomly arranged within each layer, even if the irradiation area of the excitation light is the same, the number of fluorescent particles 121A-121C contributing to the energy transfer will differ depending on the location where the excitation light is irradiated. For example, if excitation light is irradiated to each of the irradiation areas 111, 112, and 113 shown in Figure 5 and fluorescence is observed in each of the observation areas 131, 132, and 133, the number of fluorescent particles 121A-121C contributing to the energy transfer will differ in each area, and therefore the fluorescence intensity and fluorescence lifetime obtained as the observation result will be expected to differ in each area.

Figure 6 is a diagram explaining the results of verifying the diversity of fluorescence intensity. Figure 6 shows the results of irradiating different areas of the nanostructure 120 shown in Figure 5 with excitation light and observing the fluorescence intensity in each area. The irradiation areas 111, 112, and 113 were circular areas with a diameter of 16 μm. The observation areas 131, 132, and 133 corresponding to each irradiation area 111, 112, and 113 were square areas with one side measuring 1 μm. The spacing between each area was 100 μm. Laser light having a wavelength of 515 nm was used as the excitation light irradiated to the irradiation areas 111, 112, and 113. In this embodiment, the diversity of fluorescence intensity was verified by performing wavelength-resolved measurement using a spectroscope.

The wavelength dependence of the fluorescence intensity observed in each observation region 131, 132, 133 is shown in the graph in the right column of FIG. 6. The horizontal axis of each graph indicates wavelength (nm), and the vertical axis indicates fluorescence intensity (arbitrary scale). The fluorescence intensity observed in each observation region 131, 132, 133 has peaks near 540 nm, 580 nm, and 620 nm, indicating that energy transfer occurs between fluorescent particles. In addition, the graph in FIG. 6 indicates that the wavelength dependence of the fluorescence intensity differs depending on the observation region 131, 132, 133. It is presumed that the difference in fluorescence intensity is due to the difference in density of the fluorescent particles 121A, 121B, and 121C in each region.

In this way, various fluorescence spectra are obtained depending on the observation regions 131, 132, and 133, suggesting that the above-mentioned nanostructure 120 can be applied to a reservoir. That is, the signal processing device can provide a spatial modulation pattern to the input signal, selectively excite the fluorescent particles 121A contained in the first layer 120A, and then read out a signal determined according to the spatial dynamics of the FRET network portion 12.

(Embodiment 3)
In the third embodiment, a signal processing device that utilizes the time dynamics of the FRET network portion 12 will be described.

Figure 7 is a diagram explaining the results of verifying the diversity of the fluorescence lifetime. Figure 7 shows the results of irradiating different areas of the nanostructure 120 shown in Figure 5 with excitation light and observing the fluorescence lifetime in each area. The irradiation areas 111, 112, and 113 were circular areas with a diameter of 16 μm, as in embodiment 2. The observation areas 131, 132, and 133 corresponding to each irradiation area 111, 112, and 113 were rectangular areas with a short side of 1 μm and a long side of 400 μm. The interval between each area was 100 μm. Laser light having a wavelength of 515 nm was used as the excitation light irradiated to the irradiation areas 111, 112, and 113. In this embodiment, the diversity of the fluorescence lifetime was verified by performing time-resolved measurement using a spectrometer.

The change over time in fluorescence intensity observed in each observation region 131, 132, 133 is shown in the graph in the right column of Figure 7. The horizontal axis of each graph indicates time (ns), and the vertical axis indicates fluorescence intensity (arbitrary scale). A curve fitting with a fitting function (Aln(-t/τ1) + Bln(-t/τ2)) was used to show the fluorescence lifetime. Here, τ1 and τ2 represent time constants, and their respective values are as shown in Figure 7. From the graph in Figure 7, it can be seen that the fluorescence lifetime observed in each observation region 131, 132, 133 differs depending on the region. It is presumed that the difference in fluorescence lifetime is due to the length of the energy transmission path caused by the difference in the distribution of fluorescent particles 121A, 121B, 121C in each region.

In this way, various fluorescence spectra are obtained depending on the observation regions 131, 132, and 133, suggesting that the above-mentioned nanostructure 120 can be applied to a reservoir. That is, the signal processing device can provide a spatial modulation pattern to the input signal, selectively excite the fluorescent particles 121A contained in the first layer 120A, and then read out a signal determined according to the time dynamics of the FRET network portion 12.

(Embodiment 4)
In the fourth embodiment, another example of a signal processing device utilizing the time dynamics of the FRET network portion 12 will be described.

Figure 8 is a diagram illustrating the results of a verification of the diversity of fluorescence lifetimes due to differences in excitation intensity. Figure 8 shows the results of irradiating the nanostructure 120 shown in Figure 5 with excitation light of different excitation intensities and observing the fluorescence lifetime for each. The excitation light irradiated to the nanostructure 120 was laser light with a wavelength of 510 nm and an intensity of 135 mW for strong excitation and a wavelength of 510 nm and an intensity of 20 mW for weak excitation. In this embodiment, the diversity of fluorescence lifetimes was verified by performing time-resolved measurements using a spectrometer.

The change in fluorescence intensity over time when fluorescent particle 121A was excited under each excitation condition was as shown in the graph in Figure 8. The horizontal axis of each graph shows time (ns), and the vertical axis shows fluorescence intensity (arbitrary scale). A curve fitted with a fitting function (Aln(-t/τ1) + Bln(-t/τ2)) was used to show the fluorescence lifetime. The graph on the left side of Figure 8 shows the fluorescence lifetime when strong excitation was performed, and the graph on the right side shows the fluorescence lifetime when weak excitation was performed. Note that the five curves shown in each graph show the results of observations performed by changing the concentrations of fluorescent particles 121A to 121C.

From these graphs, it can be seen that when excited under weak excitation conditions, the time constants τ1 and τ2 are longer than when excited under strong excitation conditions. As a result, it is presumed that during strong excitation, excitons are generated in almost all fluorescent particles 121A-121C, which then emit light, whereas during weak excitation, there are fluorescent particles 121A-121C which do not have excitons, and the excitation energy transferred via the network structure of FRET network section 12 emits light with a delay.

In this way, a variety of fluorescence spectra can be obtained depending on the excitation intensity, suggesting that the above-mentioned nanostructure 120 can be applied to a reservoir. That is, the signal processing device selectively excites the fluorescent particles 121A-121C with an excitation intensity defined by the excitation conditions, and then reads out a signal determined according to the time dynamics of the FRET network portion 12.

The embodiments disclosed herein are illustrative in all respects and should not be considered limiting. The scope of the present invention is indicated by the claims, not by the above meaning, and is intended to include all modifications within the meaning and scope of the claims.

REFERENCE SIGNS LIST 11 Signal input section 12 FRET network section 13 Signal read-out section 14 Control section 15 Storage section 16 Input/output section

Claims (7)

an optical energy network including a plurality of fluorescent particles, the optical energy network being configured such that a transfer path of the fluorescence resonance energy via the plurality of fluorescent particles is determined autonomously and randomly;
a signal input section for exciting at least a portion of the plurality of fluorescent particles in response to an input signal to be input to the optical energy network;
a signal readout unit for reading out an optical signal from the optical energy network after exciting the part of the fluorescent particles;
and an acquisition unit that acquires a result of a calculation performed by the optical energy network based on the optical signal read out by the signal readout unit. the input signal is a signal having a spatial modulation pattern;
The signal input unit selectively excites fluorescent particles in the optical energy network based on a spatial modulation pattern of the input signal;
The signal processing device according to claim 1 , wherein the signal readout unit reads out an optical signal determined according to spatial dynamics of the optical energy network. the input signal is a signal having a spatial modulation pattern;
The signal input unit selectively excites fluorescent particles in the optical energy network based on a spatial modulation pattern of the input signal;
The signal processing device according to claim 1 , wherein the signal readout unit reads out an optical signal determined according to time dynamics of the optical energy network. the signal input excites fluorescent particles in the optical energy network at an intensity defined by an excitation condition;
The signal processing device according to claim 1 , wherein the signal readout unit reads out an optical signal determined according to time dynamics of the optical energy network. 5. The signal processing device according to claim 1, further comprising: a learning unit that learns a relationship between the optical signal read out by the signal readout unit and the calculation result acquired by the acquisition unit so that an output signal for the input signal reproduces a teacher signal indicating an ideal output. 6. The signal processing device according to claim 1, wherein the transfer of the fluorescence resonance energy is fluorescence resonance energy transfer in which excitation energy is transferred by resonance between an electron of a fluorescent particle serving as a donor and an electron of a fluorescent particle serving as an acceptor. Exciting at least a portion of the plurality of fluorescent particles in response to an input signal to an optical energy network configured to autonomously and randomly determine a transfer path of the fluorescence resonance energy of the plurality of fluorescent particles;
reading out an optical signal from the optical energy network after exciting the portion of the fluorescent particles;
A signal processing method comprising: acquiring a result of a calculation performed by the optical energy network based on the read optical signal.

Images