Space electromagnetic spectrum identification method based on parameterized quantum circuit
Through the electromagnetic spectrum recognition method based on parameterized quantum circuits, quantum convolutional neural networks are used to identify complex electromagnetic spectrum in the space environment, solving the challenge of electromagnetic signal recognition in the space environment, and achieving rapid identification and efficient calculation.
Patent Information
- Application Number
- CN202410971394.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-07-19
AI Technical Summary
In space environments, accurately identifying satellite electromagnetic signals in complex electromagnetic spectrum is a challenge because ground personnel cannot obtain electromagnetic spectrum information in space in real time, affecting the execution of satellite missions.
The spatial electromagnetic spectrum recognition method based on parameterized quantum lines is adopted to construct the time-frequency image sample set through discrete short-time Fourier transform, and the electromagnetic spectrum data processing is performed using quantum convolutional neural network, using quantum circuits and Hadamar Gate Feature maps identify electromagnetic spectrum features.
It realizes the rapid and accurate identification of the electromagnetic spectrum characteristics of space satellites in complex electromagnetic environments, improves the calculation rate and information storage capabilities, and is suitable for space environments under limited hardware conditions.
Smart Images

Figure CN118937813A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electromagnetic spectrum recognition, and in particular to a spatial electromagnetic spectrum recognition method based on parameterized quantum circuits. Background Art
[0002] Situational awareness is the development direction of all countries in the field of space, and electromagnetic situational awareness is an important part of it. Whether it is the detection of the electromagnetic environment in space or the radio communication between satellites, it can only be achieved on the basis of electromagnetic spectrum recognition. In this case, it will be of utmost importance to accurately identify whether the electromagnetic signal captured in a complex electromagnetic environment is the electromagnetic signal of the satellite. In the space environment, the identification of the electromagnetic spectrum can only rely on the satellite itself. Ground personnel cannot know the electromagnetic spectrum information in space in time and issue instructions to the satellite. Therefore, there is an urgent need for mission satellites equipped with modules for distinguishing the characteristics of the electromagnetic spectrum of space to help mission satellites distinguish the electromagnetic spectrum characteristics of satellites in complex electromagnetic environments and perform the next task. Summary of the invention
[0003] The present invention provides a space electromagnetic spectrum recognition method based on parameterized quantum circuits, which is used to accurately identify the electromagnetic spectrum characteristics of space satellites in a complex electromagnetic spectrum environment under space conditions, and take the lead in electromagnetic situation awareness.
[0004] To achieve the above object, the scheme adopted by the present invention is:
[0005] Step 1, space electromagnetic spectrum data collection;
[0006] Step 2, constructing an electromagnetic spectrum time-frequency image sample set based on discrete short-time Fourier transform;
[0007] Step 3: Design of parameter quantization circuit for spatial electromagnetic spectrum time-frequency diagram;
[0008] Step 4: electromagnetic spectrum data processing based on parameter quantization circuit.
[0009] Furthermore, the specific steps of step 2 to generate an electromagnetic spectrum time-frequency image data set based on discrete short-time Fourier transform are:
[0010] The electromagnetic spectrum is converted into a time-frequency image sample set by discrete short-time Fourier transform, namely:
[0011]
[0012] Where s(k) is the discrete signal to be identified, g(t) is the window function of short time length, and * represents conjugation;
[0013] The sample set is used as a time-frequency image dataset of the spatial electromagnetic spectrum, and is divided into a training set and a test set in a certain ratio to train and test the quantum convolutional neural network for identifying the spatial electromagnetic spectrum.
[0014] Furthermore, the algorithm flow of constructing the parameter quantization circuit for identifying the spatial electromagnetic spectrum in step 3 is as follows:
[0015] S1 converts the information of the time-frequency image sample into a quantum state that can be recognized by the quantum neural network through quantum coding, which serves as the input layer of the quantum convolutional neural network;
[0016] The S2 quantum convolution kernel uses a window of size 2×2. After the input information is mapped through the Hadamard gate feature mapping, the four quantum bits are mapped to the quantum state |ψ 1 >
[0017] S3|ψ 1 >After the quantized circuit U(θ) evolves, it becomes |ψ 2 >=U(θ)|ψ1>, and measure it on the quantum bit Take the expected value
[0018] S4 moves the convolution window, scans the entire time-frequency image sample, and obtains the eigenvalue matrix composed of expected values.
[0019] Furthermore, in step 3, quantum coding refers to the process of converting the input information of the quantum neural network into a quantum state. The mapping performed by quantum computing is (x 1 ,x 2 ,...,x m ,0)→(x 1 ',x 2 ',...,x m ',y), use |ψ> 1,...,m represents a quantum state that encodes classical data. The quantum state |ψ> 1,...,m is the superposition state of 2m computational basis vectors on the m-dimensional Hilbert space, where The basis vectors are {|0,0,…,0>,...,|1,1,…,1>}, then in
[0020] Furthermore, in step 3, the quantum state is the state of the quantum bit, that is, a two-dimensional vector is used to describe the physical state of the quantum system. The quantum state is represented by |ψ>, which can change over time and is usually normalized in quantum computing.
[0021] Furthermore, in step 3, the quantized circuit U(θ) is a learnable parameterized quantum circuit, which essentially performs a unitary transformation. The parameterized quantum circuit can be represented as a Pauli matrix {σ x ,σ y ,σ z} and the linear combination of the tensor product of the identity matrix I. If there are k layers of networks, and U k (Θ) represents the unitary matrix represented by the k-th layer network, then the entire quantum circuit can be expressed as:
[0022] U(θ)=U k (Θ k ) k-1 (Θ k )…U 1 (Θ 1 ).
[0023] Furthermore, in step 3, the measurement on the quantum bit converts the quantum state output by the quantum circuit into information that can be recognized by the computer, which is a key step in the quantum neural network. In quantum computing, projection measurement is generally used, described by the observable quantity Λ. The spectral decomposition form of the observable quantity Λ is:
[0024]
[0025] Among them, λ i is the eigenvalue of the observable quantity Λ, P i is the eigenvalue λ i Projection on the feature space. After measuring the quantum state |ψ>, the probability of obtaining result i is:
[0026] p i =p(λ=λi)=<ψ|P i |ψ>
[0027] After measurement, the quantum state becomes:
[0028]
[0029] Furthermore, the specific steps of step 4 for processing electromagnetic spectrum data based on parameter quantization circuit are as follows:
[0030] The electromagnetic spectrum data processed by the parameterized quantum circuit forms a feature matrix; the feature matrix is input into the pooling layer, and the pooling method is max-pooling; the pooled electromagnetic spectrum data enters a fully connected layer, and then the loss is calculated using the loss function, and the optimizer is fed back to the parameterized quantum circuit and the fully connected layer to update the parameters, and the output layer uses the Softmax activation function;
[0031] After the quantum convolutional neural network is trained with a time-spectrum image training set generated by short-time Fourier transform of the spatial electromagnetic spectrum, the training data set is compared and analyzed with the test data set to identify the spatial electromagnetic spectrum.
[0032] The present invention has the following advantages:
[0033] Due to the characteristics of quantum superposition and entanglement, quantum neural networks can perform parallel operations, have a faster computing rate than traditional neural networks, and n quantum bits can store 2 n bits of information, with larger information storage space.
[0034] Due to the structural advantages of the quantum neural network, the electromagnetic spectrum recognition method occupies less memory than traditional neural networks and is more able to meet the limited hardware conditions caused by the complex space environment. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Figure 1 The workflow diagram of the spatial electromagnetic spectrum identification method based on parameterized quantum circuits;
[0036] Figure 2 This is the algorithm flow chart of quantum convolutional neural network;
[0037] Figure 3 This is the structure diagram of the quantum convolutional neural network. DETAILED DESCRIPTION
[0038] Specific steps of a spatial electromagnetic spectrum recognition method based on quantum neurons:
[0039] Step 1: Space electromagnetic spectrum data collection
[0040] The original data of the electromagnetic spectrum is collected and sampled by the ground terminal of the space satellite to obtain the original samples of the electromagnetic spectrum in space;
[0041] Step 2: Construction of electromagnetic spectrum time-frequency image sample set based on discrete short-time Fourier transform
[0042] The electromagnetic spectrum is converted into a time-frequency image sample set by discrete short-time Fourier transform, namely:
[0043]
[0044] Where s(k) is the discrete signal to be identified, g(t) is the window function of short time length, and * represents conjugation;
[0045] The sample set is used as a time-frequency image dataset of the spatial electromagnetic spectrum, and is divided into a training set and a test set in a certain ratio to train and test a quantum convolutional neural network for identifying the spatial electromagnetic spectrum.
[0046] Step 3: Parameter quantization circuit design for electromagnetic spectrum time-frequency diagram
[0047] Due to the characteristics of quantum superposition and entanglement, quantum neural networks can perform parallel operations, have a faster computing rate than traditional neural networks, and n quantum bits can store 2 n Bit information, with larger information storage space;
[0048] The quantum convolutional neural network used is mainly a parameterized quantum circuit;
[0049] The parameterized quantum circuit algorithm process is as follows:
[0050] S1 converts the information of the time-frequency image sample of the electromagnetic spectrum into a quantum state that can be recognized by the parameterized quantum circuit through quantum encoding, which serves as the input layer of the quantum convolutional neural network;
[0051] The S2 quantum convolution kernel uses a window of size 2×2. After the input information is mapped through the Hadamard gate feature mapping, the four quantum bits are mapped to the quantum state |ψ 1 >
[0052] S3|ψ 1 >After the quantized circuit U(θ) evolves, it becomes |ψ 2 >=U(θ)|ψ 1 > and perform measurements on the quantum bit Take the expected value
[0053] S4 moves the convolution window to scan the time-frequency image samples of the entire spatial electromagnetic spectrum to obtain an eigenvalue matrix composed of expected values;
[0054] Step 4: Electromagnetic spectrum data processing based on parameter quantization circuit
[0055] The electromagnetic spectrum data processed by the parameterized quantum circuit forms a feature matrix, which is input into the pooling layer. The pooling can be expressed by the following formula:
[0056]
[0057] in, is the jth output feature image after pooling operation in the lth layer of the network, is the input image of the jth pooling layer in the lth layer of the network, It is a pooling operation mode, and the pooling mode used is max-pooling;
[0058] The pooled electromagnetic spectrum data enters a fully connected layer, and then the loss is calculated using the loss function. The optimizer feeds back to the parameterized quantum circuit and the fully connected layer to update the parameters. The output layer uses the Softmax activation function, which can be expressed by the following formula:
[0059]
[0060] Where x is the input data;
[0061] The quantum convolutional neural network is trained using a time-spectrum image training set generated by short-time Fourier transform of the spatial electromagnetic spectrum, and the training data set is compared and analyzed with the test data set to obtain the recognition probability of the spatial electromagnetic spectrum;
[0062] The trained quantum convolutional neural network ultimately achieves the goal of identifying the spatial electromagnetic spectrum.
Claims
1. A spatial electromagnetic spectrum recognition method based on parameterized quantum circuits, characterized in that: (1) Space electromagnetic spectrum data collection The original data of electromagnetic spectrum is collected and sampled by satellite ground terminals to obtain the original samples of space electromagnetic spectrum; (2) Construction of electromagnetic spectrum time-frequency image sample set based on discrete short-time Fourier transform The spatial electromagnetic spectrum is converted into a time-frequency image sample set by discrete short-time Fourier transform, namely: Where s(k) is the discrete signal to be identified, g(t) is the window function of short time length, and * represents conjugation; The sample set is used as a time-frequency image dataset of the spatial electromagnetic spectrum, and is divided into a training set and a test set in a certain ratio to train and test a quantum convolutional neural network for identifying the spatial electromagnetic spectrum. (3) Parameter quantization circuit design for electromagnetic spectrum time-frequency diagram Due to the characteristics of quantum superposition and entanglement, quantum neural networks can perform parallel operations, have a faster computing rate than traditional neural networks, and n quantum bits can store 2 n Bit information, with larger information storage space; The quantum convolutional neural network used is mainly a parameterized quantum circuit; The parameterized quantum circuit algorithm process is as follows: S1 converts the information of the time-frequency image sample of the spatial electromagnetic spectrum into a quantum state that can be recognized by the parameterized quantum circuit through quantum encoding, which serves as the input layer of the quantum convolutional neural network; The S2 quantum convolution kernel uses a window of size 2×2. After the input information is mapped through the Hadamard gate feature mapping, the four quantum bits are mapped to the quantum state |ψ1>; S3|ψ1> evolves through the quantized circuit U(θ) to |ψ2>=U(θ)|ψ1>, and the quantum bit measurement is performed on it. Take the expected value S4 moves the convolution window to scan the time-frequency image samples of the entire spatial electromagnetic spectrum to obtain an eigenvalue matrix composed of expected values; (4) Electromagnetic spectrum data processing based on parameter quantization circuits The electromagnetic spectrum data processed by the parameterized quantum circuit forms a feature matrix, which is input into the pooling layer. The pooling can be expressed by the following formula: in, is the jth output feature image after pooling operation in the lth layer of the network, is the input image of the jth pooling layer in the lth layer of the network, It is a pooling operation mode, and the pooling mode used is max-pooling; The pooled electromagnetic spectrum data enters a fully connected layer, and then the loss is calculated using the loss function. The optimizer feeds back to the parameterized quantum circuit and the fully connected layer to update the parameters. The output layer uses the Softmax activation function, which can be expressed by the following formula: Where x is the input data; The quantum convolutional neural network is trained using a time-spectrum image training set generated by short-time Fourier transform of the spatial electromagnetic spectrum. The output results of the training data set and the test data set are compared and analyzed to obtain the recognition probability of the spatial electromagnetic spectrum. The trained quantum convolutional neural network ultimately achieves the goal of identifying the spatial electromagnetic spectrum.
2. A method for identifying spatial electromagnetic spectrum based on parameterized quantum circuits as claimed in claim 1, characterized in that: The quantum coding is to convert the input information into quantum state. The mapping performed by quantum computing is: (x1,x2,…,x m ,0)→(x1',x2',…,x m ',y) Use |ψ> 1,...,m represents a quantum state that encodes classical data. The quantum state |ψ> 1,...,m is 2 on m-dimensional Hilbert space m The superposition state of the computational basis vectors, where The basis vectors are {|0,0,…,0>,…,|1,1,...,1>}, then in 3. The method for spatial electromagnetic spectrum recognition based on parameterized quantum circuits according to claim 1, characterized in that: The quantum state is the state of a quantum bit, that is, a two-dimensional vector describing the physical state of a quantum system. The quantum state is represented by ψ>, which can change over time and is usually normalized in quantum computing.
4. The spatial electromagnetic spectrum identification method based on parameterized quantum circuits according to claim 1, characterized in that: The quantized circuit U(θ) is a learnable parameterized quantum circuit that essentially performs a unitary transformation. The parameterized quantum circuit can be represented as a Pauli matrix {σ x ,σ y ,σ z } and the linear combination of the tensor product of the identity matrix I, if there are t layers of networks and U t (Θ) represents the unitary matrix represented by the t-th layer network, then the entire quantum circuit can be expressed as: U(Θ)=U t (I t )U t-1 (I t-1 )…U 1 (Θ1).
5. The method for spatial electromagnetic spectrum identification based on parameterized quantum circuits according to claim 1, characterized in that: The measurement on the qubit is the process of converting the quantum state output by the quantum circuit into information that can be recognized by the computer, which is a key step in the quantum neural network. In quantum computing, projection measurement is generally used, described by the observable quantity Λ. The spectral decomposition form of the observable quantity Λ is: Among them, λ i is the eigenvalue of the observable quantity Λ, P i is the eigenvalue λ i Projection on the feature space. After measuring the quantum state |ψ>, the probability of obtaining result i is: p i =p(λ=λ i )=<ψ|P i |ψ> After measurement, the quantum state becomes:
Citation Information
Patent Citations
Radar signal recognition method based on quantum particle swarm convolutional neural network
CN109597043A
Electromagnetic spectrum signal coding feature recognition method based on convolutional neural network
CN115270880A
Quantum convolutional neural network construction method and system based on quantum state amplitude transformation
CN115828999A
Image classification method of hybrid quantum neural network model based on PQC
CN119580278A
Generating mixed states and finite-temperature equilibrium states of quantum systems
WO2021062357A1
