Multimodal intelligent terminal radio frequency calibration method based on neural network
Through the neural network model combined with multimodal signal information, local calibration oscillation signals are determined, which solves the problem of large errors in humidity and temperature changes in the prior art, and improves the calibration accuracy and anti-interference ability of the RF circuit.
Patent Information
- Application Number
- CN202411124666.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-08-16
AI Technical Summary
The existing RF calibration methods fail to effectively utilize the multimodal signal information in the RF signal, resulting in large errors in the RF calibration effect in the environment of humidity and temperature change.
Using a multimodal intelligent terminal RF calibration method based on neural network, the signal intensity, temperature and humidity of the radio frequency signal are collected, the deviation temperature and humidity are calculated, the standard carrier component function and noise component vector are determined, the noise interference intensity is analyzed, and the neural network is combined with the The model determines the local calibration of the oscillation signal to eliminate environmental impact.
It improves the calibration accuracy of the RF circuit, reduces the interference of environmental factors on the RF signals, and enhances the anti-interference and response capabilities of the RF circuit.
Smart Images

Figure CN118646495B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of terminal radio frequency calibration, and in particular to a multi-modal intelligent terminal radio frequency calibration method based on neural network. Background Art
[0002] In industrial working environments, it is usually necessary to deploy monitoring equipment to monitor the working environment in real time. After collecting data, it is transmitted to the smart terminal via radio frequency communication. After the smart terminal receives the radio frequency signal from the monitoring equipment, it needs to demodulate the radio frequency signal to obtain the baseband signal. During the demodulation process, the local oscillation signal needs to be calibrated in real time with the carrier signal in the radio frequency signal to ensure consistency, so that a clear baseband signal can be demodulated.
[0003] Since RF circuits are easily affected by environmental factors such as humidity and temperature, the carrier in the RF transmission signal will fluctuate with the environment. Therefore, the local oscillation signal needs to be calibrated in real time to adapt to the carrier signal in the RF signal. Most of the existing RF calibration methods calibrate the signal based on the characteristic laws of the signal itself, without using multi-modal signal information such as temperature and humidity contained in the RF signal, which can easily lead to large errors in the RF calibration effect due to environmental factors. Summary of the invention
[0004] In order to solve the above technical problems, the present application provides a multi-modal intelligent terminal radio frequency calibration method based on neural network to solve the existing problems.
[0005] The multi-modal intelligent terminal radio frequency calibration method based on neural network of the present application adopts the following technical solutions:
[0006] An embodiment of the present application provides a multi-modal intelligent terminal radio frequency calibration method based on a neural network, the method comprising the following steps:
[0007] Collect the radio frequency signal received by the smart terminal, obtain the signal strength of the radio frequency signal at each time, obtain the local amplitude and local frequency of the smart terminal, and demodulate the radio frequency signal to obtain the temperature and humidity at each time;
[0008] The signal strength at each moment and the adjacent moments are combined into a signal strength vector at each moment; based on the difference between the temperature and humidity at each moment and the optimal working temperature and the optimal working humidity, the deviation temperature and the deviation humidity at each moment are obtained;
[0009] Determine the standard carrier component function at each moment based on the signal strength vector at each moment and the local amplitude and local frequency; obtain the noise component vector at each moment based on the difference between the signal strength at each moment and the corresponding standard carrier component function value in combination with filtering operation;
[0010] Analyze the noise component vector and its spectrum characteristics at each moment to determine the noise interference intensity at each moment; analyze the correlation between the noise interference intensity at each moment and adjacent moments and the deviated temperature and the deviated humidity, respectively, to determine the temperature interference intensity and the humidity interference intensity at each moment;
[0011] Based on the temperature interference intensity, the humidity interference intensity, the local amplitude and the local frequency, and the signal strength, a neural network model is used to determine the local calibration oscillation signal at each moment.
[0012] In one embodiment, the deviation temperature is the difference between the temperature at each moment and the optimal working temperature.
[0013] In one embodiment, the deviation humidity is the difference between the humidity at each moment and the optimal working humidity.
[0014] In one embodiment, the determination of the standard carrier component function includes:
[0015] The signal strength vector at each moment is used as the input of the phase detection algorithm, and the output is the signal phase at each moment, the standard carrier component function at time t for: , is the local amplitude, is the local frequency, is the signal phase.
[0016] In one embodiment, the determination of the noise component vector includes:
[0017] The corresponding time of each element in the signal strength vector at any moment is used as the independent variable of the standard carrier component function at any moment, and all dependent variables obtained at any moment are composed into the standard carrier vector at any moment; the difference between the signal strength vector at any moment and the standard carrier vector is calculated, and the difference is filtered to obtain the noise component vector at any moment.
[0018] In one embodiment, the method for determining the noise interference intensity is:
[0019] Based on the difference between the noise component vector at each moment and the standard carrier vector distribution, the relative intensity of noise at each moment is determined; the spectrum diagram of the noise component vector at each moment is obtained, and the noise interference intensity at each moment is determined by combining the difference between each frequency in the spectrum diagram and the local frequency, as well as the relative intensity of noise and the amplitude corresponding to each frequency in the spectrum diagram.
[0020] In one embodiment, the relative noise intensity at each moment is the ratio of the mean value of all elements in the noise component vector at each moment to the mean value of all elements in the standard carrier vector.
[0021] In one embodiment, the noise interference intensity at time t is The calculation method is: , is the relative noise intensity at time t, is the nth frequency in the spectrum diagram of the noise component vector at time t, is the local frequency, is the normalized value of the amplitude corresponding to the nth frequency in the spectrum diagram of the noise component vector at time t, To preset a value greater than 0, is the number of frequencies in the spectrum of the noise component vector at time t.
[0022] In one embodiment, the method for determining the temperature interference intensity and the humidity interference intensity is:
[0023] The noise interference intensity and the deviation temperature at each moment and the adjacent moments are respectively formed into a noise interference vector and a deviation temperature vector at each moment, and the temperature interference intensity at each moment is determined according to the correlation between the noise interference vector and the deviation temperature vector at each moment;
[0024] The humidity interference intensity is calculated by the same calculation method as the temperature interference intensity, and the deviation temperature is replaced by the deviation humidity to obtain the humidity interference intensity at each moment.
[0025] In one embodiment, determining the local calibration oscillation signal at each time comprises:
[0026] The temperature interference intensity, humidity interference intensity, local amplitude and local frequency, and signal strength are used as inputs of a neural network model. The output of the neural network model is the signal strength of the carrier at each moment. The signal strength of the carrier is used as a local calibration oscillation signal.
[0027] This application has at least the following beneficial effects:
[0028] The present application acquires the signal strength of the radio frequency signal at each moment by collecting the radio frequency signal received by the intelligent terminal, acquires the local amplitude and local frequency of the intelligent terminal, and obtains the temperature and humidity at each moment after demodulating the radio frequency signal; the signal strength at each moment and the adjacent moments are combined into a signal strength vector at each moment; based on the difference between the temperature and humidity at each moment and the optimal working temperature and the optimal working humidity, the deviation temperature and the deviation humidity at each moment are obtained; the degree of deviation of the temperature and humidity from the optimal working temperature and the optimal working humidity is reflected; further, based on the signal strength vector at each moment and the local amplitude and the local frequency, the standard carrier component function at each moment is determined; the standard carrier component function is conducive to distinguishing the noise component in the radio frequency signal; based on the difference between the signal strength at each moment and the corresponding standard carrier component function value, combined with A filtering operation is performed to obtain a noise component vector at each moment; the noise component vector reflects the degree of noise influence at each moment; the noise component vector at each moment and the frequency spectrum characteristics of the noise component vector are analyzed to determine the noise interference intensity at each moment, and characterize the ability of noise to affect the quality of the RF circuit carrier; the correlation between the noise interference intensity at each moment and adjacent moments and the deviation temperature and the deviation humidity is analyzed to determine the temperature interference intensity and humidity interference intensity at each moment; the anti-interference characteristics and response of the RF circuit to environmental changes are characterized; finally, based on the temperature interference intensity, humidity interference intensity, local amplitude and local frequency, and signal strength, a neural network model is used to determine the local calibration oscillation signal at each moment, eliminate the RF calibration error caused by environmental influences, and improve the quality of the RF circuit. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0030] Figure 1 A flowchart of the steps of the neural network-based multi-modal intelligent terminal radio frequency calibration method provided in this application;
[0031] Figure 2 A flow chart is determined for a local calibration oscillating signal. DETAILED DESCRIPTION
[0032] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of the multi-modal intelligent terminal radio frequency calibration method based on a neural network proposed in the present application in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.
[0033] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
[0034] The specific scheme of the multi-modal intelligent terminal radio frequency calibration method based on neural network provided by the present application is described in detail below with reference to the accompanying drawings.
[0035] An embodiment of the present application provides a multi-modal intelligent terminal radio frequency calibration method based on a neural network. Specifically, the following multi-modal intelligent terminal radio frequency calibration method based on a neural network is provided. Figure 1 , the method comprises the following steps:
[0036] Step S001, collect the radio frequency signal received by the smart terminal, obtain the signal strength of the radio frequency signal at each moment, obtain the local amplitude and local frequency of the smart terminal, demodulate the radio frequency signal to obtain the temperature and humidity at each moment; combine the signal strength at each moment with the signal strength at adjacent moments to form a signal strength vector at each moment; based on the difference between the temperature and humidity at each moment and the optimal working temperature and optimal working humidity, obtain the deviation temperature and deviation humidity at each moment.
[0037] This embodiment directly obtains the radio frequency signal emitted by the monitoring equipment from the radio frequency signal receiving device in the smart terminal; the monitoring equipment includes a temperature sensor and a humidity sensor, which can measure the multimodal monitoring data of the working environment of the monitoring equipment in real time, and the multimodal monitoring data includes the temperature, humidity and image data of the location of the monitoring equipment, and sends the multimodal monitoring data to the smart terminal via the radio frequency signal. The smart terminal obtains the working environment information of the monitoring equipment by demodulating the radio frequency signal.
[0038] Therefore, at time t-1, the RF signal received by the smart terminal is demodulated to obtain the temperature and humidity of the monitoring device at time t-1, which are approximated as the temperature and humidity of the monitoring device at time t, and are recorded as the ambient temperature at time t and the ambient humidity at time t; further, the ambient temperature at time t is subtracted from the optimal operating temperature of the RF circuit, and recorded as the deviation temperature at time t; the ambient temperature at time t is subtracted from the optimal operating humidity of the RF circuit, and recorded as the deviation humidity at time t; the optimal operating temperature and optimal operating humidity of the RF circuit are fixed parameters of the RF circuit. In this embodiment, the optimal operating temperature is 25°C and the optimal operating humidity is 10%RH.
[0039] At the same time, the signal strength of the radio frequency signal of the smart terminal is obtained at time t, which is recorded as the signal strength at time t.
[0040] The sampling interval in this embodiment is set to 1 microsecond, the sampling length of the data is TS=200, and all signal strengths collected from time t-TS+1 to time t are organized into a signal strength vector at time t in chronological order, all deviation temperatures collected from time t-TS+1 to time t are organized into a deviation temperature vector at time t in chronological order, and all deviation humidity collected from time t-TS+1 to time t are organized into a deviation humidity vector at time t in chronological order.
[0041] It should be noted that, in order to avoid the situation where the data before time t is insufficient to form the signal strength vector at time t or deviates from the temperature vector or the humidity vector, this embodiment performs analysis TS microseconds after the signal starts to be transmitted.
[0042] In addition, the local amplitude of the local oscillation signal is obtained from the smart terminal , local frequency .
[0043] Step S002, based on the signal strength vector at each moment and the local amplitude and local frequency, determine the standard carrier component function at each moment; based on the difference between the signal strength at each moment and the corresponding standard carrier component function value, combined with the filtering operation, obtain the noise component vector at each moment.
[0044] Although the working state of the RF circuit will be affected by temperature and humidity, due to the different design structures of different RF circuits, the impact of temperature and humidity on the working state of the RF circuit is also different; at the same time, as the RF circuit ages, the interference effect of temperature and humidity on the RF circuit is also different.
[0045] Therefore, this embodiment obtains monitoring data on the working environment of the RF circuit through the temperature and humidity sensor, evaluates the ability of the RF circuit to resist the influence of temperature and humidity, calibrates the local oscillation signal of the RF circuit in real time, and improves the quality of the RF circuit.
[0046] The signal strength vector is the RF signal directly received by the smart terminal. The signal can be divided into three components: standard carrier component, baseband signal component and noise component. The standard carrier component is the RF circuit carrier without the influence of temperature and humidity; the baseband signal component refers to the signal component of the transmission information contained in the RF signal; the noise component is the difference between the actual carrier component and the standard carrier component. Among the three components, the noise component contains all the information about the influence of temperature and humidity on the RF circuit, so it is necessary to extract the noise component.
[0047] The standard carrier component can be directly obtained through the basic parameters of the local oscillation signal, but the baseband signal component is related to the specific data of the radio frequency transmission, has no fixed identifiable characteristics, and is easily confused with the noise component. In this embodiment, the noise component to be extracted is mainly the noise component that causes the radio frequency circuit carrier to change, and the frequency components of the baseband signal and the radio frequency circuit carrier differ greatly, so this embodiment uses a high-pass filter to remove the interference of the baseband signal component on the noise component.
[0048] Finally, the noise component of the signal strength vector is obtained to analyze the resistance of the current RF circuit to the influence of temperature and humidity.
[0049] First, taking time t as an example, the signal strength vector at time t is used as input, and the zero-crossing phase detection method is used to output the signal phase at time t. The zero-crossing phase detection method is a well-known technology, and this application will not elaborate on it in detail. The implementer can select other feasible phase detection algorithms according to the actual situation, and this application does not limit it.
[0050] Combined with the local amplitude obtained , local frequency , then the standard carrier component function at time t is:
[0051] ;in, is the standard carrier component function at time t, is the local amplitude, is the local frequency, is the signal phase at time t. It is the carrier wave when the RF circuit is not affected by temperature and humidity under theoretical conditions.
[0052] Further, TS consecutive moments from t-TS+1 to t are taken as independent variables and substituted into the standard carrier component function at t to obtain TS dependent variables, that is, , organize the TS dependent variables into a standard carrier vector at time t in time sequence.
[0053] The difference between the signal strength vector at time t and the standard carrier vector is calculated as the carrier-free RF vector at time t.
[0054] The standard carrier component in the signal strength vector is removed from the carrier-removed RF vector, and the remaining signal contains baseband signal components and noise components. This embodiment requires analysis of the noise component, and the required noise component frequency is a noise component close to the local frequency. Therefore, the carrier-removed RF vector is used as input to a high-pass filter. In this embodiment, the setting frequency of the high-pass filter is , the output of the high-pass filter is recorded as the noise component vector at time t, which represents the noise signal in the signal strength vector that is more related to the RF circuit's resistance to the influence of temperature and humidity.
[0055] The high-pass filter frequency is set to , is because the frequency of the baseband signal in this embodiment is less than This setting can filter out the baseband signal and avoid interference with noise components. The high-pass filter is a well-known technology and will not be described in detail in this application.
[0056] Step S003, analyzing the noise component vector at each moment and the spectrum characteristics of the noise component vector to determine the noise interference intensity at each moment; analyzing the correlation between the noise interference intensity at each moment and adjacent moments and the deviated temperature and the deviated humidity, to determine the temperature interference intensity and the humidity interference intensity at each moment.
[0057] This embodiment mainly analyzes the impact of noise on the quality of the RF circuit carrier, so it is necessary to judge the impact of the noise level on the carrier at different times. The noise and carrier are analyzed from the characteristics in the frequency domain. The noise component concerned in this embodiment is the noise component that causes the actual carrier signal to offset the standard carrier component. Therefore, the more similar the frequency domain characteristics of the noise component are to the frequency domain characteristics of the standard carrier component, the stronger the impact of the noise on the carrier, and the interference intensity of the noise is analyzed.
[0058] First, taking time t as an example, the fast Fourier transform algorithm is used to obtain the spectrum diagram of the noise component vector at time t, where the horizontal axis of the spectrum diagram is frequency and the vertical axis is amplitude, and the Z-score normalization method is used to obtain the normalized values of all amplitudes in the spectrum diagram, wherein the fast Fourier transform algorithm and the Z-score normalization method are both existing public technologies, which are not described in detail in this application, and the implementer can choose other existing feasible frequency domain analysis algorithms and normalization methods according to actual conditions, and this application does not impose any restrictions on this.
[0059] The mean of all elements in the noise component vector at time t is calculated, which is recorded as the noise component vector strength. At the same time, the mean of all elements in the standard carrier vector at time t is calculated, which is recorded as the standard carrier strength. The ratio of the noise component vector strength to the standard carrier strength is recorded as the relative noise strength at time t. , which reflects the degree of noise impact at time t.
[0060] Combining the frequency domain characteristics of the noise component vector and the relative noise intensity, the noise interference intensity at time t is calculated, specifically: ;in, is the noise interference intensity at time t, is the relative noise intensity at time t, is the nth frequency in the spectrum diagram of the noise component vector at time t, is the local frequency, is the normalized value of the amplitude corresponding to the nth frequency in the spectrum diagram of the noise component vector at time t, To preset a value greater than 0 to prevent the denominator from being 0, in this embodiment The value is 1. The implementer can set it according to the actual situation. This application does not limit it. is the number of frequencies in the spectrum of the noise component vector at time t.
[0061] The greater the relative intensity of the noise, the greater the intensity of the noise relative to the intensity of the standard carrier, and the more serious the noise interference; since the standard carrier is an ideal waveform with a single frequency component, the spectrum of the standard carrier only contains the local frequency. , so The difference between the frequency components of the noise and the standard carrier is compared with the weight. The smaller the difference between the two, the closer the frequency of the noise is to the characteristic frequency of the standard carrier, and the greater the interference intensity of the noise on the standard carrier.
[0062] Secondly, the resistance of the RF circuit to temperature and humidity is analyzed. The noise interference intensity characterizes the interference situation of the RF circuit at different times. Therefore, this embodiment will coordinately calculate the deviation temperature and deviation humidity with the noise interference intensity respectively. The stronger the coordinated change ability of the deviation temperature or deviation humidity and the noise interference intensity is, the weaker the resistance of the RF circuit to temperature and humidity is.
[0063] Obtain the noise interference intensity from time t-TS+1 to time t, and form the noise interference vector at time t in time sequence;
[0064] The correlation between the noise interference vector and the deviation temperature vector at time t is taken as the temperature interference intensity at time t.
[0065] The correlation between the noise interference vector and the deviation humidity vector at time t is taken as the humidity interference intensity at time t.
[0066] In this embodiment, the correlation is calculated by using a mutual information calculation method. In another embodiment, the correlation is calculated by using a Pearson correlation coefficient. The mutual information calculation method and the Pearson correlation coefficient are both known technologies, and this application will not elaborate on them in detail.
[0067] In this embodiment, the value range of the temperature interference intensity or humidity interference intensity at time t is between 0 and 1. The larger the value, the stronger the interference ability of temperature or humidity on the RF circuit, and the weaker the resistance of the RF circuit to temperature or humidity.
[0068] The temperature interference intensity and humidity interference intensity at each moment are obtained by using the same calculation method as that for the temperature interference intensity and humidity interference intensity at moment t.
[0069] Step S004, based on the temperature interference intensity, the humidity interference intensity, the local amplitude and the local frequency, and the signal strength, a neural network model is used to determine the local calibration oscillation signal at each moment.
[0070] This embodiment calculates the temperature interference intensity and humidity interference intensity at each moment, obtains the local amplitude and local frequency of the local oscillation signal, and obtains the RF signal strength at each moment in real time. Therefore, this embodiment uses the obtained data as a characteristic value, uses a neural network to process the characteristic value, and completes the RF calibration of the smart terminal. Specifically:
[0071] The five values of temperature interference intensity, humidity interference intensity, signal strength, local amplitude, and local frequency at each moment are used as a set of data at each moment. At the same time, the signal strength of the carrier of the radio frequency circuit transmitting device is obtained at each moment, recorded as the carrier strength value, as the label of the set of data at the corresponding moment; 200 sets of data and labels are selected as training data sets. The implementer can set the number of data selected for the training data set according to the actual situation, and this application does not limit this.
[0072] This embodiment uses a training data set to train the LSTM neural network model. During the training process, the activation function uses the ReLU function, and the optimizer uses the Adam optimizer; the loss function is obtained by taking the mean square error between the local calibration oscillation signal and the carrier strength value; and finally a trained LSTM neural network model is obtained. The LSTM neural network model and the training process of the LSTM neural network model are both existing well-known technologies, and this application will not be described in detail here, and the implementer can select other existing feasible neural network models according to actual conditions, and this application does not limit this.
[0073] By obtaining the deviation temperature, deviation humidity, and RF signal in real time, the temperature interference intensity and humidity interference intensity are calculated in real time, and then the RF signal strength is obtained in real time, plus the local amplitude , local frequency , a total of 5 values. The 5 values are used as the input of the trained LSTM neural network model, and the output is the carrier strength value, which is used as the local calibration oscillation signal. The local calibration oscillation signal replaces the local oscillation signal for demodulation to obtain the baseband signal, completing the RF calibration of the smart terminal. The demodulation process is an existing well-known technology and will not be described in detail in this application. The flow chart for determining the local calibration oscillation signal is as follows: Figure 2 shown.
[0074] Since this embodiment obtains temperature and humidity information in the RF circuit environment through sensors and completes RF calibration in combination with a neural network, the RF calibration error caused by environmental changes to the RF circuit carrier is eliminated, thereby improving the quality of the RF circuit.
[0075] In summary, the present application acquires the signal strength of the radio frequency signal at each moment by collecting the radio frequency signal received by the intelligent terminal, acquires the local amplitude and local frequency of the intelligent terminal, and obtains the temperature and humidity at each moment after demodulating the radio frequency signal; the signal strength at each moment and the adjacent moment are combined into a signal strength vector at each moment; based on the difference between the temperature and humidity at each moment and the optimal working temperature and the optimal working humidity, the deviation temperature and the deviation humidity at each moment are obtained; the degree of deviation of the temperature and humidity from the optimal working temperature and the optimal working humidity is reflected, and further, based on the signal strength vector at each moment and the local amplitude and local frequency, the standard carrier component function at each moment is determined; the standard carrier component function is conducive to distinguishing the noise component in the radio frequency signal; based on the difference between the signal strength at each moment and the corresponding standard carrier component function value , combined with filtering operations, the noise component vector at each moment is obtained; the noise component vector reflects the degree of noise influence at each moment; the noise component vector at each moment and the frequency spectrum characteristics of the noise component vector are analyzed to determine the noise interference intensity at each moment, and characterize the ability of noise to affect the quality of the RF circuit carrier; the correlation between the noise interference intensity at each moment and the adjacent moments and the deviation temperature and the deviation humidity is analyzed respectively, and the temperature interference intensity and humidity interference intensity at each moment are determined; the anti-interference characteristics and response of the RF circuit to environmental changes are characterized; finally, based on the temperature interference intensity, humidity interference intensity, local amplitude and local frequency, and signal strength, a neural network model is used to determine the local calibration oscillation signal at each moment, eliminate the RF calibration error caused by environmental influences, and improve the quality of the RF circuit.
[0076] It should be noted that the above sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0077] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.
[0078] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Modifications to the technical solutions recorded in the aforementioned embodiments, or equivalent replacement of some of the technical features therein, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.
Claims
1. A multi-modal intelligent terminal radio frequency calibration method based on neural network, characterized in that: The method comprises the following steps: Collect the radio frequency signal received by the smart terminal, obtain the signal strength of the radio frequency signal at each time, obtain the local amplitude and local frequency of the smart terminal, and demodulate the radio frequency signal to obtain the temperature and humidity at each time; The signal strength at each moment and the adjacent moments are combined into a signal strength vector at each moment; based on the difference between the temperature and humidity at each moment and the optimal working temperature and the optimal working humidity, the deviation temperature and the deviation humidity at each moment are obtained; Determine the standard carrier component function at each moment based on the signal strength vector at each moment and the local amplitude and local frequency; obtain the noise component vector at each moment based on the difference between the signal strength at each moment and the corresponding standard carrier component function value in combination with filtering operation; Analyze the noise component vector and its spectrum characteristics at each moment to determine the noise interference intensity at each moment; analyze the correlation between the noise interference intensity at each moment and adjacent moments and the deviated temperature and the deviated humidity, respectively, to determine the temperature interference intensity and the humidity interference intensity at each moment; Based on the temperature interference intensity, the humidity interference intensity, the local amplitude and the local frequency, and the signal strength, a neural network model is used to determine the local calibration oscillation signal at each moment; The determination of the standard carrier component function includes: The signal strength vector at each moment is used as the input of the phase detection algorithm, and the output is the signal phase at each moment, the standard carrier component function at time t for: , is the local amplitude, is the local frequency, is the signal phase; The method for determining the noise interference intensity is: Based on the difference between the noise component vector at each moment and the standard carrier vector distribution, determine the relative intensity of noise at each moment; obtain the spectrum of the noise component vector at each moment, and determine the noise interference intensity at each moment by combining the difference between each frequency in the spectrum and the local frequency, and the relative intensity of noise and the amplitude corresponding to each frequency in the spectrum; The relative noise intensity at each moment is the ratio of the mean of all elements in the noise component vector at each moment to the mean of all elements in the standard carrier vector; Noise interference intensity at time t The calculation method is: , is the relative noise intensity at time t, is the nth frequency in the spectrum diagram of the noise component vector at time t, is the local frequency, is the normalized value of the amplitude corresponding to the nth frequency in the spectrum diagram of the noise component vector at time t, To preset a value greater than 0, is the number of frequencies in the spectrum of the noise component vector at time t.
2. The method for calibrating a multi-modal intelligent terminal radio frequency based on a neural network as claimed in claim 1, characterized in that: The deviation temperature is the difference between the temperature at each moment and the optimal working temperature.
3. The multi-modal intelligent terminal radio frequency calibration method based on neural network as claimed in claim 1, characterized in that: The deviation humidity is the difference between the humidity at each moment and the optimal working humidity.
4. The multi-modal intelligent terminal radio frequency calibration method based on neural network as claimed in claim 1, characterized in that: The determination of the noise component vector includes: The corresponding time of each element in the signal strength vector at any moment is used as the independent variable of the standard carrier component function at any moment, and all dependent variables obtained at any moment are composed into the standard carrier vector at any moment; the difference between the signal strength vector at any moment and the standard carrier vector is calculated, and the difference is filtered to obtain the noise component vector at any moment.
5. The multi-modal intelligent terminal radio frequency calibration method based on neural network as claimed in claim 1, characterized in that: The method for determining the temperature interference intensity and the humidity interference intensity is: The noise interference intensity and the deviation temperature at each moment and the adjacent moments are respectively formed into a noise interference vector and a deviation temperature vector at each moment, and the temperature interference intensity at each moment is determined according to the correlation between the noise interference vector and the deviation temperature vector at each moment; The humidity interference intensity is calculated by the same calculation method as the temperature interference intensity, and the deviation temperature is replaced by the deviation humidity to obtain the humidity interference intensity at each moment.
6. The multi-modal intelligent terminal radio frequency calibration method based on neural network as claimed in claim 1, characterized in that: The determining of the local calibration oscillation signal at each moment includes: The temperature interference intensity, humidity interference intensity, local amplitude and local frequency, and signal strength are used as inputs of a neural network model. The output of the neural network model is the signal strength of the carrier at each moment. The signal strength of the carrier is used as a local calibration oscillation signal.
Citation Information
Patent Citations
Oscillating circuitry
CN109302180A
Semiconductor integrated circuit test system
CN118330447A
Internet of Things multi-mode monitoring data optimization transmission method
CN118376284A
Optical modulation system and optical transmitter provided with the same
JP2007116326A
System for spectral shaping of vehicle noise cancellation
US20190103087A1
