CN113376445A - Deep learning enhanced Reedberg atom multi-frequency microwave receiver and detection method - Google Patents
Deep learning enhanced Reedberg atom multi-frequency microwave receiver and detection method Download PDFInfo
- Publication number
- CN113376445A CN113376445A CN202110636376.1A CN202110636376A CN113376445A CN 113376445 A CN113376445 A CN 113376445A CN 202110636376 A CN202110636376 A CN 202110636376A CN 113376445 A CN113376445 A CN 113376445A
- Authority
- CN
- China
- Prior art keywords
- light
- laser
- dichroic mirror
- deep learning
- rubidium bubble
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R29/00—Arrangements for measuring or indicating electric quantities not covered by groups G01R19/00 - G01R27/00
- G01R29/08—Measuring electromagnetic field characteristics
- G01R29/0864—Measuring electromagnetic field characteristics characterised by constructional or functional features
- G01R29/0871—Complete apparatus or systems; circuits, e.g. receivers or amplifiers
Landscapes
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention provides a deep learning enhanced multi-frequency microwave receiver for a Reedberg atom and a detection method, wherein the deep learning enhanced multi-frequency microwave receiver comprises a laser, a polarization beam splitter divides incident light of the laser into two parallel beams, one beam is detection light, and the other beam is reference light; the polarization directions of the two beams of light are mutually orthogonal; two beams of light are transmitted into the rubidium bubble through a dichroic mirror; the detection light and the reference light are finally received by the differential photoelectric detector; the coupling light sequentially passes through a second dichroic mirror, a rubidium bubble and a first dichroic mirror on a light path, enters the rubidium bubble after being reflected by the second dichroic mirror, is reversely coincided with the detection light, and is reflected by the second dichroic mirror; the outer side of the rubidium bubble is provided with a loudspeaker for loading a multi-frequency microwave signal to atoms of the rubidium bubble. The number of the signals of the frequency division multiplexing can reach 20, thereby improving the information transmission rate, quickly responding to the input signals and detecting the frequency spectrums of multiple frequencies at one time.
Description
Technical Field
The invention belongs to the technical field of microwave electric field sensors, and particularly relates to a deep learning enhanced Reedberg atom multi-frequency microwave receiver and a detection method.
Background
With the development of microwave technology (radar, 5G, WiFi, etc.), the requirements for microwave receivers are becoming higher and higher, and the size of conventional receiver antennas increases with increasing microwave wavelengths. The size of the recently proposed rydberg atom receiver is not limited by the wavelength of the microwave, and the rydberg atoms are very sensitive to the microwave signal due to their large dipole moment, and can be used as a high-precision and high-sensitive microwave receiver, and in addition, the atoms have many different states of rydberg, and can be used for receiving microwave signals of different frequencies.
The detection of the relative phase signal of two microwave signals based on rydberg atoms is reported in the literature [ appl.phys.lett.114,114101(2019) ]. However, the document can only detect the relative phase between two microwave signals, cannot detect the phase between a plurality of microwave signals, and has no expandability. The utilization rate of the signal frequency band is not high.
The document IEEE Antennas and Wireless Transmission Letters, vol.18, No.9, pp.1853-1857, Sept.2019 proposes to use the phase between two microwave signals to transmit information for digital communication. The literature describes experiments in which phase shift keying signals and quadrature amplitude modulation signals are received using rydberg atoms. However, the receiving apparatus is complicated, and when different frequency division multiplexing signals are separated, a band pass filter is used. The document shows only two frequency division multiplexed signals for demonstration purposes, if multiple frequency division multiplexed signals are used, multiple band pass filters are required.
Disclosure of Invention
The present invention proposes to decode frequency division multiplexing phase shift keying signals (FDM-2PSK) using the rydberg atoms and a deep learning model. The invention consists of a Reidberg atom and a deep learning model. The multi-frequency microwave signals are loaded on the rydberg atoms through a loudspeaker to influence the transmission spectrum of the detection light, then the transmission spectrum is input into a trained deep learning model, the model outputs the relative phase between the microwave signals, and the original signals are demodulated.
The specific technical scheme is as follows:
the deep learning enhanced multi-frequency microwave receiver for the rydberg atoms comprises a laser, wherein a light path of the laser sequentially comprises a half-wave plate, a polarization beam splitter, a first dichroic mirror, a rubidium bubble, a second dichroic mirror and a differential photoelectric detector;
the polarization beam splitter divides the incident light of the laser into two parallel beams, one beam is probe light, and the other beam is reference light; the polarization directions of the two beams of light are mutually orthogonal; two beams of light are transmitted into the rubidium bubble through a dichroic mirror; the detection light and the reference light are finally received by the differential photoelectric detector;
the coupling light sequentially passes through a second dichroic mirror, a rubidium bubble and a first dichroic mirror on a light path, enters the rubidium bubble after being reflected by the second dichroic mirror, is reversely coincided with the detection light, and is reflected by the second dichroic mirror;
the outer side of the rubidium bubble is provided with a loudspeaker for loading a multi-frequency microwave signal to atoms of the rubidium bubble.
The half-wave plate is used for changing the polarization direction of light and is combined with the polarization beam splitter to adjust the relative light intensity of the two beams of light split by the polarization beam splitter.
The propagation direction of the multi-frequency microwave signal loaded by the horn is perpendicular to the detection light.
The laser emitted by the laser is 795 nm; the coupled light is 474nm laser.
Laser emitted by the laser is divided into two parallel beams by a polarization beam splitter and penetrates through the rubidium bubble, one beam is probe light, and the other beam is reference light; the coupling light and the detection light are reversely coincided, and atoms are excited into a Reedberg state through a two-photon process;
the detection light and the reference light are finally received by a differential photoelectric detector, and an output signal of the differential photoelectric detector is a transmission spectrum of the detection light;
the microwave is loaded on atoms through a loudspeaker, and the propagation direction of the microwave is vertical to the probe light, the reference light and the coupling light; and inputting the output signal of the differential photoelectric detector into the trained deep learning model to obtain a prediction result.
The deep learning model builds a neural network by using a bidirectional long-short term memory layer and a one-dimensional convolutional layer;
sequentially adding a batch normalization layer, an activation function layer and a one-dimensional pooling layer in front of a two-way long-short term memory layer behind the one-dimensional convolution layer;
the final fully-connected layer of the model is used to scale the output to the appropriate dimension.
The invention has the technical effects that:
the invention has expandability, and the number of the signals of the frequency division multiplexing can reach 20, thereby improving the information transmission rate.
The device is simple, and because a deep learning model is used, the invention does not need a plurality of band-pass filters to separate frequency division multiplexing signals.
Robustness, due to the fact that the rydberg atoms are very sensitive to an electromagnetic field, measured data often have large noise, and robustness of the rydberg atom receiver to the noise can be improved by means of deep learning.
And fourthly, the method can be used for multi-target detection. Due to the Doppler effect, different speeds correspond to different frequency spectrums, and the frequency spectrums of multiple frequencies can be detected completely by the model provided by the invention at one time.
And the speed is high, and compared with the method of fitting by using a principal equation, the trained model can quickly respond to the input signal.
Drawings
FIG. 1a is an energy level diagram of an embodiment;
FIG. 1b is a schematic structural view of the present invention;
FIG. 1c is a probe light transmission spectrum of the example upon incidence of microwaves of 4 frequencies;
FIG. 2 is an example deep learning model and input-output data dimensions for each layer;
FIG. 3 is a loss curve for the training set and validation set of an embodiment;
FIG. 4 shows the test results of the model of the example;
fig. 5 shows the result of transmitting the two-dimensional code according to the embodiment.
Detailed Description
The specific technical scheme of the invention is described by combining the embodiment.
As shown in fig. 1b, the deep learning enhanced rydberg atom multi-frequency microwave receiver comprises a laser 1, wherein a light path of the laser 1 sequentially comprises a half-wave plate 2, a polarization beam splitter 3, a first dichroic mirror 4, a rubidium bubble 5, a second dichroic mirror 7 and a differential photoelectric detector 8;
the polarization beam splitter 3 splits the incident light of the laser 1 into two parallel beams, one beam is probe light 12, and the other beam is reference light 11; the polarization directions of the two beams of light are mutually orthogonal; the two beams of light are transmitted into a rubidium bubble 5 through a first dichroic mirror 4, and atoms in the rubidium bubble are rubidium 85 atoms; the detection light 12 and the reference light 11 are finally received by the differential photodetector 8;
the coupling light 10 is emitted by a second laser 9, the light path sequentially passes through a second dichroic mirror 7, a rubidium bubble 5 and a first dichroic mirror 4, the coupling light 10 enters the rubidium bubble 5 after being reflected by the second dichroic mirror 7, is parallel to and reversely coincided with the detection light 12, and is then reflected by the second dichroic mirror 7;
the rubidium bubble 5 is provided with a loudspeaker 6 at the outer side for loading multi-frequency microwave signals on atoms of the rubidium bubble 5.
The half-wave plate 2 is used for changing the polarization direction of light and is combined with the polarization beam splitter 3 to adjust the relative light intensity of the two beams of light split by the polarization beam splitter 3.
The propagation direction of the multi-frequency microwave signal loaded by the horn 6 is perpendicular to the probe light 12.
The laser 1 emits 795nm laser; the coupled light 10 is a 474nm laser.
Atoms undergo a two-photon process |5S1/2,F=2>→|5P1/2,F'=3>→|51D3/2>Excited into a rydberg state |51D3/2>. The energy level diagram is shown in FIG. 1a, and the probe light 12 excites rubidium 85 atom transition |5S1/2,F=2>→|5P1/2,F'=3>Coupled light 10 excites atomic transition |5P1/2,F'=3>→|51D3/2>Multifrequency microwave driven atomic transition |51D3/2>→|50F5/2>. The detuning of the probe light 12, the coupling light 10 and the microwave is respectively deltap,Δc,ΔsAnd its ratio frequency omegap,Ωc,Ωs. The detection light 12 and the reference light 11 are finally received by the differential photodetector 8, and the output signal of the differential photodetector 8 is the transmission spectrum of the atom.
FIG. 1c shows the transmission spectrum of the detected light obtained from the differential photodetector 8, in this case Δp=0,Δc=0,ΔsI.e. the laser resonates with an atom. The microwave field is the superposition of 4 microwaves with different frequencies, and the frequencies of the microwaves are respectively f1=17.62GHz-3kHz f2=17.62GHz-1kHz f317.62GHz +1kHz, and f417.62GHz +3kHz, the frequency spacing Δ f is 2 kHz.
In the experiment, multi-frequency microwave signalsWhereinThe phase of the single-frequency microwave signal takes 0 or pi, omega1,2,3Is the angular frequency of a single frequency microwave signal. The single-frequency signals are loaded on atoms through a horn 6 after being superposed, and the propagation direction of the microwave is vertical to the detection light 12, the reference light 11 and the coupling light 10. The output signal of the differential photoelectric detector 8 is input into a trained deep learning model to obtain a prediction resultOne of the multi-frequency microwave signals is a reference signal whose amplitude is greater than the other signals in order to eliminate non-linearities, i.e., the phenomenon of corresponding identical waveforms with different phases. The modulation signal of the multi-frequency microwave signal is the relative phase of the other microwave signals relative to the reference signal.
Since data is time series and requires a model to have long-term memory capability, the present embodiment uses a long short-term memory layer (LSTM layer) in Keras to build a neural network. In order to further improve the memory of the network, the present embodiment uses a Bi-directional LSTM layer and a one-dimensional convolution layer (1-d convolution layer). A batch normalization layer (batch normalization layer), an activation function layer (ReLU layer), and a one-dimensional pooling layer (1-d max-firing layer) are sequentially added after the one-dimensional convolution layer in order to make the model converge faster during training [ series Ioffe, Christian Szegedy Proceedings of the 32nd International Conference on Machine Learning, PMLR 37: 448-. The last fully connected layer (dense layer) of the model is used to adjust the output to the appropriate dimension.
Fig. 2 shows the neural network structure used, and the input-output data dimension of each layer. The dimensionality of the model input data is (sample number, transmission spectrum data point number and characteristic number), the sample number is 64, the transmission spectrum duration is 1ms, the time interval is 1us, namely the data point number is 1000, and the characteristic number is 1. The dimension of the input data is thus (64, 1000, 1).
Relative phase of microwave signal with respect to reference signal(as labels) and the corresponding probe light transmission spectra as data sets, the data sets being randomly scrambled and divided into a training set, a validation set and a test set, for training, validating and testing the deep learning model, respectively. The verification set is used for judging whether the model is over-fitted in the training process. The probe light transmission spectrum and the corresponding phase are input into the model for model learning during training. During learning, namely training, the 'distance' between the predicted value and the true value of the model, namely the loss function, is calculated, and the derivative of the loss function to the weight of each layer is reversely propagated to update the weight of each layer. After the training is completed, i.e. after the Loss curve (Loss) of the model converges, the embodiment can use the model to predict, and perform the test by using the test set. During the test, only the probe light transmission spectrum is provided, and no corresponding phase is provided, and the performance index of the model on the test set is obtained by comparing the predicted value of the model with the actual value.
Fig. 3 shows that the loss of the training set and the verification set is converged and close after 130 generations in the training and verification process of the model, which indicates that the model is trained well and has no over-fitting and under-fitting phenomena. The training set size is 524 and the validation set size is 172. Wherein the model traverses the data of the training set all over one generation. The loss function is the mean square error between the predicted value and the true valuem is the number of samples, yiIs the true value, f (x)i) And (4) predicting the value of the model. Finally, the frequency division multiplexing signals of arbitrary different phase combinations are combined by a loudspeaker (0 or pi) are loaded on the rydberg atoms, and the original relative phase can be obtained through the prediction of a neural networkThe number of frequency division multiplexing signals used for training the neural network is the same as that used for prediction.
The model test results are shown in fig. 4, and a confusion matrix is drawn by using the prediction results of the model in the test set. The test set size was 160% and the accuracy of the model was 99.375%.
The result of transmitting the two-dimensional code is shown in fig. 5, when the training generations are 3 and 4, the information cannot be read from the two-dimensional code recovered from the model, and the corresponding accuracy rates are 51.02% and 76.87%, respectively. And at the 35 th generation, the information can be read from the two-dimensional code recovered from the model, and the accuracy at the moment is 99.32%. The accuracy is determined by dividing the number of correctly predicted waveforms by the total number of waveforms. In the embodiment, the multi-frequency microwave is used for carrying out frequency division multiplexing digital communication, a two-dimensional code is transmitted, and the two-dimensional code is coded by a letter 'USTC'. The two-dimensional code is coded and transmitted in a frequency division multiplexing binary phase shift keying (FDM-2PSK) mode, the two-dimensional code is received through a Reidberg atom and recovered by using deep learning models trained in different generations, and finally the fact that the two-dimensional code reconstructed by the deep learning models after 35 generations of training can decode the original letter 'USTC' is found, and the model accuracy is 99.32%. The hardware device for machine learning is an NVIDIA GeForce GTX 1650 display card.
Claims (5)
1. The deep learning enhanced Reedberg atom multi-frequency microwave receiver is characterized by comprising a first laser (1), wherein a light path of the first laser (1) sequentially comprises a half-wave plate (2), a polarization beam splitter (3), a first dichroic mirror (4), a rubidium bubble (5), a second dichroic mirror (7) and a differential photoelectric detector (8);
the half-wave plate (2) is used for changing the polarization direction of light and is combined with the polarization beam splitter (3) to adjust the relative light intensity of the two beams of light split by the polarization beam splitter (3);
the polarization beam splitter (3) splits the incident light of the laser (1) into two parallel beams, one beam is probe light (12), and the other beam is reference light (11); the polarization directions of the two beams of light are mutually orthogonal; the two beams of light are transmitted into the rubidium bubble (5) through the first dichroic mirror (4); the detection light (12) and the reference light (11) are finally received by the differential photoelectric detector (8);
the coupling light (10) is emitted by a second laser (9), the light path sequentially passes through a second dichroic mirror (7), a rubidium bubble (5) and a first dichroic mirror (4), the coupling light (10) enters the rubidium bubble (5) after being reflected by the second dichroic mirror (7), is reversely coincided with the detection light (12), and is reflected by the first dichroic mirror (4);
the outside of the rubidium bubble (5) is provided with a loudspeaker (6) for loading a multi-frequency microwave signal on atoms of the rubidium bubble (5).
2. The deep learning enhanced rydburg atom multi-frequency microwave receiver according to claim 1, wherein the horn (6) is loaded with multi-frequency microwave signals propagating in a direction perpendicular to the probe light (12) and the reference light (11).
3. A deep learning enhanced riedberg atom multi-frequency microwave receiver according to claim 1, characterized in that the first laser (1) emits laser light at 795 nm; the laser light emitted by the second laser (9) is 474nm coupled light (10).
4. The method for detecting a deep learning enhanced Reedberg atom multi-frequency microwave receiver according to any one of claims 1 to 3, wherein the laser emitted from the laser (1) is split into two parallel beams by a polarization beam splitter (3) and passes through a rubidium bubble (5), one beam is a detection light (12), and the other beam is a reference light (11); the coupling light (10) and the detection light (12) are reversely coincided, and atoms are excited into a Reedberg state through a two-photon process;
the detection light (12) and the reference light (11) are finally received by a differential photoelectric detector (8), and the output signal of the differential photoelectric detector (8) is the transmission spectrum of the detection light;
the microwave is loaded on atoms through a horn (6), and the propagation direction of the microwave is vertical to the probe light (12), the reference light (11) and the coupling light (10); and (3) inputting the output signal of the differential photoelectric detector (8) into a trained deep learning model to obtain a prediction result.
5. The method of claim 4, wherein the deep learning model builds a neural network using a two-way long-short term memory layer and a one-dimensional convolutional layer;
sequentially adding a batch normalization layer, an activation function layer and a one-dimensional pooling layer in front of a two-way long-short term memory layer behind the one-dimensional convolution layer;
the final fully-connected layer of the model is used to scale the output to the appropriate dimension.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110636376.1A CN113376445B (en) | 2021-06-08 | 2021-06-08 | Deep learning enhanced Reedberg atom multi-frequency microwave receiver and detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110636376.1A CN113376445B (en) | 2021-06-08 | 2021-06-08 | Deep learning enhanced Reedberg atom multi-frequency microwave receiver and detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113376445A true CN113376445A (en) | 2021-09-10 |
CN113376445B CN113376445B (en) | 2021-12-14 |
Family
ID=77576355
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110636376.1A Active CN113376445B (en) | 2021-06-08 | 2021-06-08 | Deep learning enhanced Reedberg atom multi-frequency microwave receiver and detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113376445B (en) |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107046159A (en) * | 2016-02-08 | 2017-08-15 | 日本电产艾莱希斯株式会社 | Waveguide assembly and antenna assembly and radar with the waveguide assembly |
CN109004983A (en) * | 2018-06-21 | 2018-12-14 | 上海第二工业大学 | A kind of accurate method for sensing based on phase to intensity modulated transfer principle |
CN109655828A (en) * | 2018-12-18 | 2019-04-19 | 中国人民解放军国防科技大学 | SST (stimulated Raman Scattering) depth learning inversion method for multi-frequency one-dimensional synthetic aperture microwave radiometer |
CN110133942A (en) * | 2019-04-23 | 2019-08-16 | 中国科学技术大学 | Regulate and control the multistable device and method of Rydberg atom |
CN110226184A (en) * | 2016-12-27 | 2019-09-10 | 杰拉德·迪尔克·施密茨 | For machine sensible system and method |
CN110401492A (en) * | 2018-07-27 | 2019-11-01 | 中国计量科学研究院 | A kind of radio amplitude-modulated signal method of reseptance and amplitude modulation Quantum receiver based on quantum effect |
CN110518985A (en) * | 2019-07-08 | 2019-11-29 | 清远市天之衡传感科技有限公司 | Radio digital communication system and method based on Rydberg atom frequency mixer |
GB201915420D0 (en) * | 2019-10-24 | 2019-12-11 | British Telecomm | Wireless telecomunications network |
CN111595826A (en) * | 2019-02-21 | 2020-08-28 | 中国科学技术大学 | System and method for analyzing phase diagram of polynome of rydberg atoms |
US20200295838A1 (en) * | 2019-03-11 | 2020-09-17 | Government Of The United States Of America, As Represented By The Secretary Of Commerce | Rydberg atom mixer and determining phase of modulated carrier radiation |
CN112098737A (en) * | 2020-08-27 | 2020-12-18 | 北京无线电计量测试研究所 | Method and device for measuring microwave electric field intensity |
-
2021
- 2021-06-08 CN CN202110636376.1A patent/CN113376445B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107046159A (en) * | 2016-02-08 | 2017-08-15 | 日本电产艾莱希斯株式会社 | Waveguide assembly and antenna assembly and radar with the waveguide assembly |
CN110226184A (en) * | 2016-12-27 | 2019-09-10 | 杰拉德·迪尔克·施密茨 | For machine sensible system and method |
CN109004983A (en) * | 2018-06-21 | 2018-12-14 | 上海第二工业大学 | A kind of accurate method for sensing based on phase to intensity modulated transfer principle |
CN110401492A (en) * | 2018-07-27 | 2019-11-01 | 中国计量科学研究院 | A kind of radio amplitude-modulated signal method of reseptance and amplitude modulation Quantum receiver based on quantum effect |
CN109655828A (en) * | 2018-12-18 | 2019-04-19 | 中国人民解放军国防科技大学 | SST (stimulated Raman Scattering) depth learning inversion method for multi-frequency one-dimensional synthetic aperture microwave radiometer |
CN111595826A (en) * | 2019-02-21 | 2020-08-28 | 中国科学技术大学 | System and method for analyzing phase diagram of polynome of rydberg atoms |
US20200295838A1 (en) * | 2019-03-11 | 2020-09-17 | Government Of The United States Of America, As Represented By The Secretary Of Commerce | Rydberg atom mixer and determining phase of modulated carrier radiation |
CN110133942A (en) * | 2019-04-23 | 2019-08-16 | 中国科学技术大学 | Regulate and control the multistable device and method of Rydberg atom |
CN110518985A (en) * | 2019-07-08 | 2019-11-29 | 清远市天之衡传感科技有限公司 | Radio digital communication system and method based on Rydberg atom frequency mixer |
GB201915420D0 (en) * | 2019-10-24 | 2019-12-11 | British Telecomm | Wireless telecomunications network |
CN112098737A (en) * | 2020-08-27 | 2020-12-18 | 北京无线电计量测试研究所 | Method and device for measuring microwave electric field intensity |
Non-Patent Citations (3)
Title |
---|
CHRISTOPHER L. HOLLOWAY 等: "A Multiple-Band Rydberg-Atom Based Receiver/Antenna: AM/FM Stereo Reception", 《RESEARCHGATE》 * |
HAIYANG ZOU 等: "Atomic Receiver by Utilizing Multiple Radio-Frequency Coupling at Rydberg States of Rubidium", 《APPLIED SCIENCES》 * |
廖开宇 等: "基于里德堡原子的微波传感与通信", 《中国科学》 * |
Also Published As
Publication number | Publication date |
---|---|
CN113376445B (en) | 2021-12-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10243673B2 (en) | Frequency demodulation systems and methods for optical signals | |
AU2019243976B2 (en) | Balanced optical receivers and methods for detecting optical communication signals | |
Tuzlukov | Signal detection theory | |
Allen et al. | Detecting a stochastic background of gravitational radiation: Signal processing strategies and sensitivities | |
US9891134B2 (en) | Long distance optical fiber sensing system and method | |
US3706094A (en) | Electronic surveillance system | |
CN105973450B (en) | Optical fiber Fizeau interference arrays distributed vibration sensing system and method | |
Yang | Source depth estimation based on synthetic aperture beamfoming for a moving source | |
Baltus et al. | Convolutional neural networks for the detection of the early inspiral of a gravitational-wave signal | |
Bereketli et al. | Experimental results for direction of arrival estimation with a single acoustic vector sensor in shallow water | |
Yeo et al. | Reduced‐dimension DOD and DOA estimation through projection filtering in bistatic MIMO radar with jammer discrimination | |
CN113376445B (en) | Deep learning enhanced Reedberg atom multi-frequency microwave receiver and detection method | |
WO2023069333A1 (en) | Few-mode rayleigh-based distributed fiber sensor for simultaneous temperature and strain sensing | |
CA2732685A1 (en) | System and method for magnitude and phase retrieval by path modulation | |
Tian et al. | Fully digital multi‐frequency compact high‐frequency radar system for sea surface remote sensing | |
EP2738958A1 (en) | Angular resolution of images using photons having non-classical states | |
CN106559142A (en) | Light emitting devices, launching technique, optical pickup apparatus and method of reseptance | |
WO2012138252A2 (en) | Method for transmitting information and method for detecting a signal | |
FR3027685A1 (en) | SYSTEM FOR LOCATING AN OBJECT PROVIDED WITH AN RFID TAG | |
Liang et al. | 2D-SenseNet: a simultaneous demodulation and denoising network for DAS | |
CN112051610A (en) | Vector field multi-mode surface wave frequency dispersion calculation method and system | |
Yue et al. | Real-time identification of frequency-hopping millimeter-wave signals using photonic time stretch and reservoir computing | |
Zhang et al. | MT‐BCS‐Based DoA and Bandwidth Estimation of Unknown Signals through Multiple Snapshots Data | |
Chaudhary et al. | Highly efficient photonic radar by incorporating MDM-WDM and machine learning classifiers under adverse weather conditions | |
CN114459589B (en) | Underwater acoustic signal detection method based on Redberg atomic radar |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20220315 Address after: 201804 room 105, floor 1, building 1, No. 6988, Jiasong North Road, Jiading District, Shanghai j828 Patentee after: Shanghai Texian Photoelectric Technology Co.,Ltd. Address before: 230026 room 611-338, R & D center building, international intelligent voice Industrial Park, 3333 Xiyou Road, high tech Zone, Hefei, Anhui Patentee before: Hefei Hengyuan Quantum Technology Co.,Ltd. |