Programmable phase shifter

The programmable phase shifter addresses the challenge of frequency-dependent phase shift by using a frequency synchronization section with adaptive filtering and nonlinear signal conversion, enabling it to maintain accurate phase adjustment across changing input signal frequencies.

WO2024247099A1PCT designated stage expired Publication Date: 2024-12-05TDK CORP
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
PCT/JP2023/020079
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-30
Publication Date
2024-12-05

AI Technical Summary

Technical Problem

Existing phase shifters, particularly those using analog elements, face difficulties in generating a desired phase difference for input signals, especially when the frequency of the input signal changes, leading to improper operation.

Method used

A programmable phase shifter is designed with a frequency synchronization section that synchronizes the frequency of the output signal with the fundamental frequency of the input signal, utilizing an adaptive filter and feedback mechanism to adjust waveforms, and can include recurrent neural networks, MEMS microphone arrays, or SAW filters for nonlinear signal conversion, enabling adaptive response to frequency changes.

Benefits of technology

The programmable phase shifter effectively adjusts the phase of the output signal to match the desired phase difference for varying input signal frequencies, ensuring proper operation across different frequencies.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2023020079_05122024_PF_FP_ABST
    Figure JP2023020079_05122024_PF_FP_ABST
Patent Text Reader

Abstract

This programmable phase shifter comprises: a frequency synchronization unit that is configured to receive input of an input signal and to output a frequency synchronization signal; and a phase adjustment unit that is configured to receive input of the frequency synchronization signal and to output an output signal. The frequency synchronization unit is configured such that the frequency of the frequency synchronization signal is synchronized with the fundamental frequency of the input signal. The phase adjustment unit is provided with an adaptive filter that performs adjustment such that the waveforms of the output signal and the reference signal match, and a feedback unit that inputs, into the frequency synchronization unit, a feedback signal based on the output signal.
Need to check novelty before this filing date? Find Prior Art

Description

Programmable Phase Shifter

[0001] The present invention relates to a programmable phase shifter.

[0002] A phase shifter shifts the phase of an input signal and outputs a signal with a fixed phase shift. Phase shifters are widely used in engineering, for example, in high-frequency circuits, directivity control of phased arrays, and control of robot actuators.

[0003] For example, Patent Document 1 describes a variable phase shifter using diode elements, and Patent Document 2 describes a digital phase shifter using transistors.

[0004] JP 2005-142724 A JP 2006-524933 A

[0005] Phase shifters using analog elements have difficulty in generating the desired phase difference for an arbitrary input signal. For example, even if a phase shifter can generate the desired phase difference for an input signal of a specific frequency, the phase shifter may not operate properly if the frequency of the input signal changes.

[0006] The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a programmable phase shifter that can adjust the phase of an output signal programmably relative to an input signal.

[0007] (1) A programmable phase shifter according to a first aspect includes a frequency synchronization unit configured to receive an input signal and output a frequency synchronization signal, and a phase adjustment unit configured to receive the frequency synchronization signal and output an output signal. The frequency synchronization unit is configured to synchronize the frequency of the frequency synchronization signal with a fundamental frequency of the input signal. The phase adjustment unit includes an adaptive filter that adjusts the waveforms of the output signal and a reference signal so that they match, and a feedback unit that inputs a feedback signal based on the output signal to the frequency synchronization unit.

[0008] (2) In the programmable phase shifter according to the above aspect, the frequency synchronization unit may have a plurality of nodes. The input signal is input to at least one of the plurality of nodes. The signal output from the plurality of nodes to the phase adjustment unit is synchronized with the fundamental frequency of the input signal.

[0009] (3) In the programmable phase shifter according to the above aspect, at least one of the plurality of nodes may have a different time constant.

[0010] (4) In the programmable phase shifter according to the above aspect, the frequency synchronization unit may be a recursive neural network.

[0011] (5) In the programmable phase shifter according to the above aspect, the frequency synchronization unit may be an analog device that can perform nonlinear conversion of the input signal.

[0012] (6) In the programmable phase shifter according to the above aspect, the analog device may be a MEMS microphone array.

[0013] (7) In the programmable phase shifter according to the above aspect, the analog device may be a SAW filter.

[0014] (8) In the programmable phase shifter according to the above aspect, the frequency synchronization unit may be a digital circuit that can perform nonlinear conversion of the input signal.

[0015] (9) In the programmable phase shifter according to the above aspect, the phase adjustment unit may be a digital filter.

[0016] The programmable phase shifter according to the above aspect can adaptively respond to changes in the frequency of the input signal.

[0017] FIG. 1 is a block diagram of a programmable phase shifter according to a first embodiment. FIG. 2 is a schematic diagram of characteristic parts of the programmable phase shifter according to the first embodiment. FIG. 3 is a cross-sectional view of an analog device constituting the programmable phase shifter according to the first embodiment. FIG. 4 is a schematic diagram of an analog device constituting the programmable phase shifter according to the first embodiment. FIG. 5 is a schematic diagram of an analog device constituting the programmable phase shifter according to the first embodiment. FIG. 6 is a schematic diagram of a recurrent neural network without a feedback unit. FIG. 7 is a schematic diagram of a recurrent neural network with a feedback unit. FIG. 8 shows results of a first example of the programmable phase shifter according to the present embodiment. FIG. 9 shows results of a second example of the programmable phase shifter according to the present embodiment.

[0018] The present embodiment will be described in detail below with reference to the drawings as appropriate. The drawings used in the following description may show characteristic portions enlarged for the sake of clarity, and the dimensional ratios of each component may differ from the actual ones. The materials, dimensions, etc. exemplified in the following description are merely examples, and the present invention is not limited thereto. Appropriate changes can be made within the scope of the effects of the present invention.

[0019] 1 is a block diagram of a programmable phase shifter 100 according to the first embodiment. The programmable phase shifter 100 includes, for example, a frequency synchronization unit 10, a phase adjustment unit 20, and a memory 30.

[0020] An input signal I(t) is input to the frequency synchronization unit 10. The input signal I(t) is input from an input signal source. There may be one or more input signals I(t). When the input signal I(t) is input, the frequency synchronization unit 10 outputs a frequency synchronization signal. There may be one or more frequency synchronization signals. The frequency synchronization signal is synchronized with the fundamental frequency of the input signal I(t). In other words, the frequencies of the input signal I(t) and the frequency synchronization signal are approximately the same.

[0021] Fig. 2 is a schematic diagram of a characteristic part of the programmable phase shifter 100 according to the first embodiment. The frequency synchronization unit 10 shown in Fig. 2 has a reservoir network, which is one type of recurrent neural network.

[0022] A neural network is a mathematical model that mimics the network of nerve cells in the brain. Recurrent neural networks can handle nonlinear time series data. Nonlinear time series data is data whose values ​​change over time, and examples include stock prices and the number of influenza epidemics. Recurrent neural networks are suitable for processing time series data by feeding the results of processing in neurons in later layers back to neurons in earlier layers.

[0023] The reservoir network includes a plurality of nodes 11. The number of nodes 11 is not particularly limited.

[0024] Each of the nodes 11 is connected to other nodes 11. For example, weights are defined between the respective nodes 11. When a signal propagates from one node 11 to another node 11, the defined weight is applied to the signal, causing the signal to change.

[0025] The number of weights that can be defined is equal to the number of combinations of connections between the nodes 11. Each weight between the nodes 11 is defined in principle and does not change through learning. Each weight between the nodes 11 is arbitrary and may be the same as or different from one another. Some of the connection weights between multiple nodes 11 may change through learning.

[0026] An input signal I(t) is input to one or more nodes 11. The input signal I(t) may be input to all nodes 11. The input signal I(t) interacts with one another as it propagates between multiple nodes 11 within the frequency synchronization unit 10. Signal interaction means that a signal propagated to one node 11 affects signals propagating through other nodes 11. For example, as the input signal I(t) propagates between nodes 11, a weight connecting the nodes 11 is applied to the input signal I(t), causing it to change. The frequency synchronization unit 10 projects the input signal into a multidimensional nonlinear space.

[0027] The input signal I(t) input to the frequency synchronization unit 10 is replaced by another signal as it propagates between the nodes 11 .

[0028] At least one of the multiple nodes 11 is connected to a phase adjustment unit 20. The frequency of the input signal I(t) does not change as it propagates through the frequency synchronization unit 10. Therefore, the signal output from the node 11 connected to the phase adjustment unit 20 is synchronized with the fundamental frequency of the input signal I(t).

[0029] The concept of a reservoir network can be replicated in analog devices that can perform nonlinear transformations on an input signal I(t). Analog devices are composed of a combination of physical elements. For example, analog devices convert input signals into vibrations, electromagnetic fields, magnetic fields, spin waves, etc.

[0030] For example, a MEMS microphone array is an example of an analog device. A MEMS microphone array is an arrangement of multiple MEMS microphones that are electrically connected to one another. MEMS is an abbreviation for Micro Electronics Mechanical System.

[0031] 3 is a cross-sectional view of an example of a MEMS microphone, where MEMS microphone 11A corresponds to each of the nodes 11 in FIG.

[0032] The MEMS microphone 11A includes, for example, a diaphragm 1, a MEMS chip 2, an integrated circuit 3, a substrate 4 having an opening 4A, and a protective plate 5. The diaphragm 1, the MEMS chip 2, and the integrated circuit 3 are formed on the substrate 4 and electrically connected to each other. The diaphragm 1, the MEMS chip 2, and the integrated circuit 3 are protected by the protective plate 5.

[0033] The MEMS microphone 11A converts, for example, sound waves into electrical signals. Sound waves input through the opening 4A vibrate the diaphragm 1. The vibration of the diaphragm 1 changes, for example, the capacitance of a capacitor in the MEMS chip 2, and is converted into an electrical signal. The integrated circuit 3 includes, for example, an analog-to-digital converter, and outputs the electrical signal in analog form.

[0034] An input signal I(t) is input to the MEMS microphone array. The input signal I(t) is, for example, a sound wave. The input signal I(t) vibrates the diaphragm 1 of the MEMS microphone 11A. The vibration of the diaphragm 1 of one MEMS microphone 11A generates a sound wave. This sound wave vibrates the diaphragm 1 of the nearby MEMS microphone 11A. In other words, when the input signal I(t) is input to the MEMS microphone array, the diaphragm 1 of each MEMS microphone 11A vibrates based on the input signal I(t).

[0035] The vibration of the diaphragm 1 of each MEMS microphone 11A affects the diaphragm 1 of the other MEMS microphones 11A, and the vibrations of the diaphragms 1 of the MEMS microphones 11A interact with each other. An input signal I(t) input to the MEMS microphone array is nonlinearly transformed. At least one MEMS microphone 11A in the MEMS microphone array is connected to a phase adjustment unit 20. The signal that has been nonlinearly transformed in the MEMS microphone array is input to the phase adjustment unit 20 as a frequency synchronization signal.

[0036] For example, a SAW filter is an example of an analog device. SAW is an abbreviation for Surface Acoustic Wave.

[0037] 4 is a schematic diagram of a SAW filter 10A. The SAW filter 10A has, for example, an input portion IN, an output portion OUT, and a propagation region IF. The SAW filter 10A has, for example, an input portion IN, an output portion OUT, and a propagation region IF on the same plane of a piezoelectric body Pb. The piezoelectric body Pb is, for example, made of barium titanate (BaTiO 3 ), lithium niobate (LiNbO 3 ), lithium tantalate (LiTaO 3 ), lead zirconate titanate (PZT).

[0038] The input section IN is a section that generates a signal in the propagation region IF. An input signal I(t) is applied to the input section IN. The input section IN converts the signal into an elastic wave and inputs it to the propagation region IF. The input section IN inputs multiple elastic waves of different frequencies to the propagation region IF. The input section IN shown in FIG. 4 has a first input section INa and a second input section INb. The first input section INa and the second input section INb sandwich the propagation region IF. The elastic wave input from the first input section INa and the elastic wave input from the second input section INb collide in the propagation region IF and interfere with each other.

[0039] The propagation region IF is a region in the piezoelectric body where elastic waves propagate. The propagation region IF is also an interference region where elastic waves interfere with each other. The propagation region IF corresponds to the frequency synchronization unit 10 in FIG. 4 .

[0040] The surface acoustic waves converted from the signal at the input IN interfere in the propagation region IF, where the acoustic waves interact and the input signal is subjected to nonlinear transformation and processing.

[0041] The output unit OUT is a part that outputs, as a signal, elastic waves that have interfered in the propagation region IF. The output unit OUT detects elastic waves that are generated when multiple elastic waves input from the input unit IN interfere in the propagation region IF. The output unit OUT, for example, converts the elastic waves into an electric potential. The elastic waves generated in the propagation region IF are detected by the output unit OUT. The elastic waves are converted into an electric potential difference by the output unit OUT and output as a signal to the outside. The output unit OUT is connected to the phase adjustment unit 20. The signal output from the output unit OUT is a frequency synchronization signal.

[0042] The analog device is not limited to these examples, as long as it converts an input signal into vibration, an electromagnetic field, a magnetic field, a spin wave, or the like.

[0043] For example, the analog device may be a spin torque oscillator array in which spin torque oscillators are arranged in an array, each of which corresponds to a node 11 in FIG.

[0044] Furthermore, for example, the analog device may be a combination of RLC circuits. FIG. 5 shows an example of an RC circuit applicable to the node 11. The RC circuit includes an input terminal t1, a resistor R, a capacitor C, a switching element SW, an output terminal t2, and a control terminal t3. The analog device may be, for example, a combination of multiple RC circuits as shown in FIG. 5. At least one of the circuits constituting the analog device has a different time constant. The existence of circuits with different time constants causes a phase difference to occur in the RC circuits corresponding to each node 11.

[0045] Although an example in which an analog device is applied to the frequency synchronization unit 10 has been shown so far, the frequency synchronization unit 10 may also be a digital circuit. The digital circuit performs nonlinear conversion on the input signal I(t).

[0046] A frequency synchronization signal is input to the phase adjustment unit 20. The phase adjustment unit 20 is configured to output an output signal z(t) when the frequency synchronization signal is input. The phase adjustment unit 20 includes, for example, an adaptive filter 21 and a feedback unit 22.

[0047] The adaptive filter 21 includes, for example, a calculator 25 and a comparator 26. The calculator 25 and the comparator 26 are, for example, part of a microcontroller or a processor. The calculator 25 and the comparator 26 include, for example, a CPU, a ROM, a RAM, etc.

[0048] The calculator 25 performs, for example, a multiplication operation to apply a weight w to the frequency synchronization signal output from the frequency synchronization unit 10, and a summation operation to add up the results of the respective product operations. The frequency synchronization signal is output from, for example, at least one of the multiple nodes 11 that make up the frequency synchronization unit 10. The calculator 25 may further perform an operation to substitute the result of the product-sum operation into an activation function.

[0049] The comparator 26 compares the calculation result (corresponding to the output signal z(t)) of the calculator 25 with the reference signal s(t) stored in the memory 30. The comparator 26 compares the degree of match between the waveforms of the output signal z(t) and the reference signal s(t). During learning, the comparator 26 feeds back the result of the degree of match between the waveforms of the output signal z(t) and the reference signal s(t) to the calculator 25. The calculator 25 adjusts the weight w to be applied to the frequency synchronization signal based on the fed-back information. The weight w is adjusted so as to increase the degree of match between the waveforms of the output signal z(t) and the reference signal s(t). If learning to set the weight w is performed separately, the programmable phase shifter 100 does not need to include the comparator 26 and the memory 30.

[0050] The feedback unit 22 inputs a feedback signal u(t) based on the output signal z(t) to the frequency synchronization unit 10. The feedback signal u(t) may be the output signal z(t) as is, or may be the output signal z(t) transformed by a predetermined function. The feedback signal u(t) is input to, for example, each of the nodes 11 of the frequency synchronization unit 10. The feedback signal u(t) may be input to, for example, some of the nodes 11 of the frequency synchronization unit 10.

[0051] The phase adjustment unit 20 may be an analog filter that processes a continuously changing signal, or a digital filter that divides a continuous signal into time domains and processes them. Examples of the digital filter include an FIR filter and an IIR filter.

[0052] The memory 30 stores a reference signal s(t). The reference signal s(t) is, for example, an ideal value of the phase difference between the input signal I(t) and the output signal z(t). The memory 30 stores, for example, an ideal value of the phase difference for each frequency of the input signal I(t). The ideal value of the phase difference is a value indicating how much the phase of the output signal z(t) should be delayed or advanced relative to the phase of the input signal I(t), and varies depending on, for example, the frequency of the input signal I(t). The reference signal s(t) is set in advance, for example, depending on the task that requires processing. The memory 30 stores, for example, the phase difference between the input signal I(t) and the output signal z(t) corresponding to the task, for each frequency.

[0053] Next, the operation of the programmable phase shifter 100 will be described. The programmable phase shifter 100 learns the setting conditions for adding a desired phase difference to the input signal I(t) and outputting it as the output signal z(t). Based on the learning results, the programmable phase shifter 100 outputs the output signal z(t) when the actual input signal I(t) is input.

[0054] First, the learning operation of the programmable phase shifter 100 will be described. First, the ideal state of the signal output from the programmable phase shifter 100 when the input signal I(t) is input to the programmable phase shifter 100 is determined. The ideal state of the signal is determined based on the ideal value of the phase difference between the phase of the input signal I(t) and the phase of the output signal z(t). This ideal signal is the reference signal s(t). The reference signal s(t) is set according to the task and is determined for each frequency of the input signal I(t). The reference signal s(t) has a desired phase difference with respect to the input signal I(t).

[0055] Next, the weights between the nodes 11 of the frequency synchronization unit 10 and the weights w between the frequency synchronization unit 10 and the phase adjustment unit 20 (the weights w between each node 11 and the calculator 25) are arbitrarily set. Hereinafter, this setting will be referred to as the initial setting. For example, the weights may all be constant, or may vary according to a certain probability distribution.

[0056] Next, the input signal I(t) is applied to the programmable phase shifter 100 in the initial setting. The node 11 to which the input signal I(t) is applied is synchronized with the fundamental frequency of the input signal I(t).

[0057] For example, if the node 11 is a MEMS microphone, when an input signal I(t) is input as a sound wave, the diaphragm 1 of each MEMS microphone 11A is forced to vibrate in synchronization with the fundamental frequency of the input signal I(t). The phases of the vibrations of the diaphragms 1 of the MEMS microphones 11A may be the same or different. The phase of the vibrations of the diaphragms 1 of at least one MEMS microphone 11A is different from the phases of the vibrations of the diaphragms 1 of the other MEMS microphones 11A.

[0058] For example, if the frequency synchronization unit 10 is a SAW filter 10A, when the input signal I(t) is input as an acoustic wave, the acoustic waves interfere and interact within the propagation region IF, and as a result, the SAW filter 10A generates an acoustic wave synchronized with the fundamental frequency of the input signal I(t) or a harmonic of the fundamental frequency of the input signal I(t).

[0059] The frequency synchronization unit 10 inputs a frequency synchronization signal to the phase adjustment unit 20. Based on the initial setting, the calculator 25 applies a weight w to each of the frequency synchronization signals input to the phase adjustment unit 20 and adds the respective signals together. The calculator 25 then outputs the result as an output signal z(t).

[0060] The output signal z(t) output based on the initial setting often does not match the reference signal s(t). The comparator 26 determines the degree of match between the output signal z(t) and the reference signal s(t). The comparator 26 adjusts the weight w between the frequency synchronization unit 10 and the phase adjustment unit 20 (the weight w between each node 11 and the calculator 25) based on the degree of match between the output signal z(t) and the reference signal s(t). The weight w is adjusted so as to increase the degree of match between the output signal z(t) and the reference signal s(t).

[0061] The calculator 25 also transmits the output signal z(t) to the feedback unit 22. The feedback unit 22 transmits the feedback signal u(t) to the frequency synchronization unit 10. The feedback signal u(t) may be the output signal z(t) itself, or may be a signal obtained by performing a predetermined functional transformation. The feedback signal u(t) may be applied to each of the nodes 11 of the frequency synchronization unit 10, or may be applied to only some of the nodes 11.

[0062] FIG. 6 is a schematic diagram of a recurrent neural network without a feedback unit 22, and FIG. 7 is a schematic diagram of a recurrent neural network with a feedback unit 22.

[0063] 6 can be expressed as a differential equation as follows: where r represents the internal state of the neural network, J is the connection matrix and represents the adjacency matrix between nodes 11, I(t) represents the input signal, and v represents the weight applied to the input signal.

[0064]

[0065] Similarly, the process of the recurrent neural network shown in Figure 7 can be expressed as a differential equation as follows: where z(t) is the output signal and u is the weight applied to the feedback signal. u z(t) corresponds to u(t) in Figure 2.

[0066]

[0067] Here, z(t) is the product of the signal output from node 11 and the weight, z(t)=w T r, where r is tanh(x). Therefore, by converting equation (2), the following equation (3) is obtained.

[0068]

[0069] Comparing equation (1) and equation (3), when the feedback signal u(t) is fed back, it can be considered that the coupling matrix corresponding to the internal coupling within the frequency synchronization unit 10 changes from J to J'. In other words, when the feedback unit 22 feeds back the feedback signal u(t), the internal coupling of the frequency synchronization unit 10 changes. The internal coupling of the frequency synchronization unit 10 changes, for example, when the weights set between the nodes 11 change.

[0070] In this way, when the programmable phase shifter 100 learns, the weight w between the frequency synchronization unit 10 and the phase adjustment unit 20 (the weight w between each node 11 and the calculator 25) and the internal coupling of the frequency synchronization unit 10 change from the initial settings. As the programmable phase shifter 100 repeats learning, setting conditions that increase the degree of match between the output signal z(t) and the reference signal s(t) are found for each frequency of the input signal I(t).

[0071] Next, the phase shifting operation of the programmable phase shifter 100 will be described. Based on the learning of the programmable phase shifter 100, appropriate setting conditions for achieving a desired phase change are determined. Based on the learning results, the weight w between the frequency synchronization unit 10 and the phase adjustment unit 20 (the weight w between each node 11 and the calculator 25) and the internal coupling of the frequency synchronization unit 10 are determined. These setting conditions are referred to as optimal settings.

[0072] When an actual input signal I(t) is input, the programmable phase shifter 100 outputs an output signal z(t) based on the optimal settings. As a result of learning, the output signal z(t) based on the optimal settings should show a high degree of agreement with the reference signal s(t). As a result, the programmable phase shifter 100 can generate an output signal z(t) that imparts a desired phase difference to the input signal I(t). Because the programmable phase shifter 100 learns the optimal settings for each frequency of the input signal I(t), it can appropriately generate a desired phase difference even if the frequency of the input signal I(t) changes.

[0073] As described above, the programmable phase shifter 100 according to this embodiment not only adjusts the weight w between the frequency synchronization unit 10 and the phase adjustment unit 20 (the weight w between each node 11 and the calculator 25), but also adjusts the internal coupling within the frequency synchronization unit 10 by inputting the feedback signal u(t) to the frequency synchronization unit 10. That is, the programmable phase shifter 100 according to this embodiment can adjust two parameters, the weight w and the internal coupling of the frequency synchronization unit 10, enabling more adaptive learning and automatically adjusting each parameter to achieve an appropriate phase difference.

[0074] FIG. 8 shows the results of using the programmable phase shifter 100 according to this embodiment to advance the phase difference of the output signal z by 0.25π relative to the input signal f. The upper graph in FIG. 8 shows the response during learning when a reference signal is presented, and the lower graph in FIG. 8 shows the response during inference after learning. The dotted line in FIG. 8 represents the input signal f, and the solid line represents the output signal z. As shown in FIG. 8, using the programmable phase shifter 100 according to this embodiment enabled the output signal z to be advanced by 0.25π relative to the input signal f. The results in the upper graph in FIG. 8 indicate that the use of a reference signal during learning, the weight w between the frequency synchronization unit 10 and the phase adjustment unit 20 (the weight w between each node 11 and the calculator 25), and the optimal conditions for the internal coupling of the frequency synchronization unit 10 were appropriately determined. The results in the lower graph in FIG. 8 indicate that the desired phase difference was achieved for the input signal f during inference.

[0075] FIG. 9 also shows the results of delaying the phase difference of the output signal z by 0.25π relative to the input signal f using the programmable phase shifter 100 according to this embodiment. The upper graph in FIG. 9 shows the response during learning when a reference signal is presented, and the lower graph in FIG. 9 shows the response during inference after learning. The dotted line in FIG. 9 represents the input signal f, and the solid line represents the output signal z. As shown in FIG. 9, using the programmable phase shifter 100 according to this embodiment enabled the output signal z to be delayed by 0.25π relative to the input signal f. The results in the upper graph in FIG. 9 indicate that the use of a reference signal during learning, the weight w between the frequency synchronization unit 10 and the phase adjustment unit 20 (the weight w between each node 11 and the calculator 25), and the optimal conditions for the internal coupling of the frequency synchronization unit 10 were appropriately determined. The results in the lower graph in FIG. 9 indicate that the desired phase difference was achieved for the input signal f during inference.

[0076] Up to this point, a preferred embodiment of the present invention has been exemplified based on the first embodiment, but the present invention is not limited to this embodiment.

[0077] The programmable phase shifter according to this embodiment can also be applied to controlling the directionality of a phased array antenna (laser), controlling a robot actuator, and the like.

[0078] For example, when the programmable phase shifter according to the present embodiment is used for controlling the directivity of a phased array antenna, the programmable phase shifter according to the present embodiment is connected to each antenna element of the phased array antenna. The phased array antenna can impart directivity to a signal by changing the phase shift difference imparted to each antenna element. The programmable phase shifter according to the present embodiment can impart an appropriate phase difference to a signal depending on the input signal input to the antenna element, the position of the antenna element within the phased array antenna, etc.

[0079] Furthermore, for example, when the programmable phase shifter according to this embodiment is used to control an actuator of a robot, the programmable phase shifter according to this embodiment is connected to an electric circuit that controls the robot. For example, the movement of a robot arm has a different phase near the fulcrum of the arm and at a position away from the fulcrum. The programmable phase shifter according to this embodiment can provide signals with an appropriate phase difference to each electric circuit that controls each part of the robot arm.

[0080] 10... Frequency synchronization unit, 10A... SAW filter, 11... Node, 11A... MEMS microphone, 20... Phase adjustment unit, 21... Adaptive filter, 22... Feedback unit, 25... Arithmetic unit, 26... Comparator, 30... Memory, 100... Programmable phase shifter

Claims

1. A programmable phase shifter comprising: a frequency synchronization unit configured to receive an input signal and output a frequency synchronization signal; and a phase adjustment unit configured to receive the frequency synchronization signal and output an output signal, wherein the frequency synchronization unit is configured to synchronize the frequency of the frequency synchronization signal with a fundamental frequency of the input signal, and the phase adjustment unit comprises: an adaptive filter that adjusts the waveforms of the output signal and a reference signal so that they match; and a feedback unit that inputs a feedback signal based on the output signal to the frequency synchronization unit.

2. The programmable phase shifter of claim 1, wherein the frequency synchronization unit has a plurality of nodes, the input signal is input to at least one of the plurality of nodes, and the signal output from the plurality of nodes to the phase adjustment unit is synchronized with the fundamental frequency of the input signal.

3. The programmable phase shifter of claim 2, wherein at least one of said plurality of nodes has a different time constant.

4. The programmable phase shifter according to claim 1, wherein the frequency synchronization section is a recurrent neural network.

5. The programmable phase shifter according to claim 1, wherein the frequency synchronization section is an analog device capable of nonlinearly converting the input signal.

6. The programmable phase shifter of claim 5, wherein the analog device is a MEMS microphone array.

7. The programmable phase shifter of claim 5, wherein said analog device is a SAW filter.

8. The programmable phase shifter according to claim 1, wherein the frequency synchronization section is a digital circuit capable of nonlinearly converting the input signal.

9. The programmable phase shifter according to claim 1, wherein the phase adjustment section is a digital filter.

Citation Information

Patent Citations

  • Phase synchronizing circuit

    JP1980037031A

  • Phase-shifting device

    JP2006086857A

  • Phase quadrature signal generator having a variable phase shift network

    US5983082A

  • Information processing device, method for setting hidden node, and method for producing information processing device

    WO2022201425A1