CN117708657A - Echo signal resonance peak range detection optimization method and system - Google Patents
Echo signal resonance peak range detection optimization method and system Download PDFInfo
- Publication number
- CN117708657A CN117708657A CN202410160351.2A CN202410160351A CN117708657A CN 117708657 A CN117708657 A CN 117708657A CN 202410160351 A CN202410160351 A CN 202410160351A CN 117708657 A CN117708657 A CN 117708657A
- Authority
- CN
- China
- Prior art keywords
- frequency
- signal
- time
- echo
- peak
- 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
- 238000000034 method Methods 0.000 title claims abstract description 68
- 238000001514 detection method Methods 0.000 title claims description 32
- 238000005457 optimization Methods 0.000 title claims description 23
- 238000010897 surface acoustic wave method Methods 0.000 claims abstract description 16
- 238000000605 extraction Methods 0.000 claims abstract description 15
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000013527 convolutional neural network Methods 0.000 claims description 25
- 239000011159 matrix material Substances 0.000 claims description 23
- 230000006870 function Effects 0.000 claims description 21
- 230000008569 process Effects 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 12
- 230000009467 reduction Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 7
- 238000010408 sweeping Methods 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 230000001174 ascending effect Effects 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 3
- 230000000630 rising effect Effects 0.000 claims description 3
- 239000003638 chemical reducing agent Substances 0.000 claims description 2
- 230000003247 decreasing effect Effects 0.000 claims description 2
- 238000005316 response function Methods 0.000 claims description 2
- 238000005070 sampling Methods 0.000 claims description 2
- 230000001131 transforming effect Effects 0.000 claims description 2
- 238000010276 construction Methods 0.000 claims 1
- 230000004044 response Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000013499 data model Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 230000036039 immunity Effects 0.000 description 1
- 230000003447 ipsilateral effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a method and a system for detecting and optimizing the range of resonance peaks of echo signals, wherein the method comprises the following steps: acquiring an echo signal of a wireless passive surface acoustic wave sensor; preprocessing an echo signal to obtain a time-frequency spectrogram of the echo signal, and determining time-frequency data according to the time-frequency spectrogram; performing feature extraction on time-frequency data according to a preset CNN model containing a double-channel convolution kernel to obtain signal features; sweep the frequency value corresponding to at least one waveform characteristic point based on a preset sliding window; and positioning the peak position and the trough position of the signal characteristic by using frequency values in interval boundaries at two sides of the center point, and determining the interval moving direction of the sliding window at the peak position and the trough position to obtain the frequency peak value of the signal characteristic. And the interference of noise to the receiving peak value is avoided, and the response time of the system is quickened.
Description
Technical Field
The invention belongs to the technical field of signal detection, and particularly relates to an echo signal resonance peak range detection optimization method and system.
Background
With the development of sensing technology and the increase of market demands, passive wireless technology has been greatly developed. In some application scenarios, conventional active wireless sensors cannot work properly due to high temperature, high voltage, high rotation, strong corrosion, limited space, while passive wireless sensors have great flexibility due to their lack of power supply systems and use of wireless technology. Surface Acoustic Wave (SAW) sensors show great potential in these application scenarios due to their small size, high resolution, high immunity to interference and the possibility of long distance transmission. The resonant wireless passive surface acoustic wave sensor only needs to adopt a narrow-band radio frequency pulse query signal in measurement, however, taking a resonant SAW (SAWR) sensor which is more suitable for wireless application as an example, the echo signal has the characteristics of high carrier frequency, short effective length, low signal-to-noise ratio and the like, and great difficulty is brought to signal detection. The echo signal energy of the wireless passive SAW sensor is usually low, and the current frequency estimation algorithm has certain dependence on the signal-to-noise ratio of the signal, so that the accuracy of frequency estimation is insufficient.
Disclosure of Invention
The invention provides an echo signal resonance peak range detection optimization method and system, which are used for solving the technical problem that the accuracy is insufficient when frequency estimation is caused by certain dependence on the signal-to-noise ratio of a signal in the current frequency estimation algorithm.
In a first aspect, the present invention provides a method for detecting and optimizing a resonance peak range of an echo signal, including: acquiring an echo signal of a wireless passive surface acoustic wave sensor; preprocessing the echo signal to obtain a time-frequency spectrogram of the echo signal, and determining time-frequency data according to the time-frequency spectrogram; performing feature extraction on the time-frequency data according to a preset CNN model containing a double-channel convolution kernel to obtain signal features, wherein the signal features comprise at least one waveform feature point and a frequency value corresponding to the at least one waveform feature point; sweeping the frequency value corresponding to the at least one waveform characteristic point based on a preset sliding window, wherein the sliding window comprises a center point and interval boundaries positioned at two sides of the center point; and positioning the peak position and the trough position of the signal characteristic according to the frequency values in the interval boundaries at the two sides of the center point, and determining the interval moving direction of the sliding window at the peak position and the trough position to obtain the frequency peak value of the signal characteristic.
In a second aspect, the present invention provides an echo signal resonance peak range detection optimization system, including: the acquisition module is configured to acquire echo signals of the wireless passive surface acoustic wave sensor; the preprocessing module is configured to preprocess the echo signals to obtain a time-frequency spectrogram of the echo signals, and determine time-frequency data according to the time-frequency spectrogram; the extraction module is configured to perform feature extraction on the time-frequency data according to a preset CNN model containing a double-channel convolution kernel to obtain signal features, wherein the signal features comprise at least one waveform feature point and a frequency value corresponding to the at least one waveform feature point; the frequency sweep module is configured to sweep the frequency value corresponding to the at least one waveform characteristic point based on a preset sliding window, wherein the sliding window comprises a center point and interval boundaries positioned at two sides of the center point; the output module is configured to position the peak position and the trough position of the signal feature according to the frequency values in the interval boundaries at the two sides of the center point, determine the interval moving direction of the sliding window at the peak position and the trough position, and obtain the frequency peak value of the signal feature.
In a third aspect, there is provided an electronic device, comprising: the system comprises at least one processor and a memory communicatively connected with the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the steps of the echo signal formant range detection optimization method of any one of the embodiments of the present invention.
In a fourth aspect, the present invention also provides a computer readable storage medium having stored thereon a computer program, the program instructions, when executed by a processor, cause the processor to perform the steps of the echo signal resonance peak range detection optimization method according to any of the embodiments of the present invention.
According to the echo signal resonance peak range detection optimization method and system, time-frequency analysis preprocessing is conducted on echo signals, one-dimensional signal characteristics are classified, a two-dimensional matrix is built, two-channel signal fusion is conducted, a time-frequency data block is built, an SConv-DD convolution neural network is introduced to extract characteristic processing on two-channel input data, training parameters are reduced, and detection accuracy is improved. And carrying out characteristic analysis on the echo signals, determining a frequency sweeping range, solving the problem of characteristic value searching, and optimizing a frequency sweeping strategy. The waveform position of the section is judged by setting a sliding region, wherein the characteristic value is contained in the data characteristic of the section edge, the sliding step length and the sliding section can be automatically adjusted, the resonant peak value is sampled, the resonant frequency of the corresponding characteristic point is found out by the range section, the interference of noise to the receiving peak value is avoided, and the response time of the system is accelerated.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of an echo signal resonance peak range detection optimization method according to an embodiment of the present invention;
FIG. 2 is a block diagram of a dual channel SConv-DD convolution neural network according to one embodiment of the present disclosure;
FIG. 3 is a block diagram of a range optimization selection algorithm for resonant frequencies according to an embodiment of the present invention;
FIG. 4 is a block diagram of an echo signal resonance peak range detection optimization system according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, a flowchart of an echo signal resonance peak range detection optimization method is shown.
As shown in fig. 1, the echo signal resonance peak range detection optimization method specifically includes the following steps:
step S101, acquiring an echo signal of the wireless passive surface acoustic wave sensor.
In the step, echo signals of the SAW resonance sensor are collected, and the received signal vector is，；
Wherein,representing n transmission symbol vectors +.>For transmitting symbol vectors, < >>For receiving symbol vectors, < >>Representing a set of constellation points, each group of signal transmitting antennas being randomly selected from +.>Is selected and sent to the receiver,representing a channel matrix>The element in (a) represents the complex channel gain between the transmitting antenna and the receiving antenna, and the statistical characteristics of the elements obey a complex gaussian distribution with a mean value of 0 and a variance of 1.
To receive signalsAnd output transmit signal probability value->Building a relation mapping function between themWherein->For the input relevant signal +.>To train the parameter set of the neural network. The parameter set may be optimized by a loss function, mapping the input to the desired output.
Step S102, preprocessing the echo signals to obtain a time-frequency spectrogram of the echo signals, and determining time-frequency data according to the time-frequency spectrogram.
In the step, the collected echo signal is modulated, time-frequency analyzed and preprocessed, the processed one-dimensional sequence features are characterized, and data are formed into a group of vector sequences with the dimension of m according to the sequence:wherein each sequence. These vectors represent the sub-pixel values from->Dot start +.>Consecutive->Values.
The specific signal components are expressed as:
，
in the method, in the process of the invention,for the input relevant signal +.>Is channel noise>For transmitting signals +.>For a channel matrix>For the real part of the received signal, < > for>For receiving the imaginary part of the signal>Is the real part of the channel noise, < >>As the imaginary part of the channel noise,for transmitting the real part of the signal, < > is->For transmitting the imaginary part of the signal>Is the real part of the channel matrix, < >>Is the imaginary part of the channel matrix.
Time-frequency domain signal processing:
，
in the method, in the process of the invention,as a basis function +.>、/>All are discrete points, and are filled with->Is a time domain parameter->For wavelet transform coefficients +.>For discretizing the basis function +.>For sampling signals +.>Is the derivative of the basis function;
discretizing the echo signals according to the wavelet function, and weighting the discretized echo signals to obtain a time-frequency spectrogram of the echo signals, wherein the discretization expression of the echo signals is as follows:
，
，
the expression for weighting the discretized echo signals is as follows:
，
in the method, in the process of the invention,for complex multiplication times, ++>To collect the signal +.>For the weighting factor>Is->Discrete transformation of->For the response function, K is the fine coefficient, +.>Transforming the expression for the basis function;
step S103, carrying out feature extraction on the time-frequency data according to a preset CNN model containing a double-channel convolution kernel to obtain signal features, wherein the signal features comprise at least one waveform feature point and a frequency value corresponding to the at least one waveform feature point.
In the step, introducing a CNN model of an SConv-DD double-channel convolution kernel, and carrying out convolution noise reduction identification on the double-channel CNN model according to a time-frequency data model specifically comprises the following steps: zero padding and signal convolution are carried out on time-frequency data, and net input of a CNN model based on one-dimensional and two-dimensional matrix double-channel convolution kernels is obtained; and obtaining an output signal of the convolution layer according to the net input and the activation function of the CNN model based on the double-channel convolution model kernel. The output signal of the pooling layer is obtained by adopting a maximum pooling method, and faster and more accurate symbol detection is realized according to the rolling and pooling operation;
there are multiple hidden layers between the input layer and the output layer. From the slaveLayer transfer to->The output of a layer can be expressed as:
，
wherein,indicate->Training parameters of the layer->Indicate->Mapping function of layer.
First, theThe layer input expression is: />，/>For layer l+1Output (I)>For piecewise linear soft sign operator, < >>Weight matrix for SConv-DD convolutional neural network,>the shift matrix of the neural network is convolved for the SConv-DD.
Changing a general CNN full-connection network into a sparse connection structure to obtain an output signal of a sparse layer;
and obtaining network output obtained by adding the convolution results of each channel according to the output signals of the connected sparse layers.
The neural network structure is adopted, and the two-channel input end isAnd +.>The two vector elements are respectively used as two-channel convolution inputs, the output is added by the convolution operation result of each channel, and the structure of the two-channel SConv-DD convolution neural network is shown in figure 2.
It should be noted that, based on historical time-frequency data and historical signal features corresponding to the historical time-frequency data, iterative training is performed on a dual-channel SConv-DD convolution neural network to obtain a CNN model containing a dual-channel convolution kernel, wherein the expression of the SConv-DD convolution neural network is as follows:
，
，
in the method, in the process of the invention,transpose of channel matrix, +.>For the output of the l+1 layer, +.>For piecewise linear soft sign operator, < >>For mathematical transformation, ++>For reorganizing different input data according to different channels, the +.>Is Hadamard product, < >>Is a residual feature;
the loss function of the CNN model is as follows:
，
in the method, in the process of the invention,for loss function parameters, ++>For a channel matrix>For the input relevant signal +.>For training parameter sets, ++>For iterative index, ++>For hiding layer first layer->For the output of layer l, +.>Is a loss difference;
the training parameter set is as follows:，/>weight matrix for SConv-DD convolutional neural network,>the shift matrix of the neural network is convolved for the SConv-DD.
Further, before feature extraction is performed on the time-frequency data according to a preset CNN model containing a double-channel convolution kernel, noise reduction processing is performed on the time-frequency data according to a preset noise reduction model, wherein the expression of the noise reduction model is as follows:
，
，
，
in the method, in the process of the invention,is an intermediate signal +.>For the noise reduction function->Is a product coefficient>For the channel set +.>For variance->、All are training parameters, and are added with->For transmitting symbol vectors, < >>For the ith input noise, < >>For single channel difference, I is a diagonal matrix, < ->Is an intermediate parameter->For (I)>For the noise reducer model, ++>For the input relevant signal +.>For residual error->To accept symbol vectors>Is the variance estimate.
According to the model, various parameters are input, CNN and residual vectors are trained, so that the loss function is the mostAnd (3) small, obtaining a preliminary estimated value. Repeating training toThe output value is input as an initial value. Calculating the estimated variance->And training->Up toAnd finally outputting a signal vector value.
Step S104, sweep frequency values corresponding to the at least one waveform characteristic point are swept based on a preset sliding window, wherein the sliding window comprises a center point and interval boundaries positioned on two sides of the center point.
In the step, according to the characteristic analysis, the characteristic values corresponding to the waveform characteristic points of the signal characteristics are swept, the positions of the peaks and the troughs of the signal are positioned and the interval direction is moved according to the data characteristics of the interval boundary, and the corresponding characteristic frequencies are estimated by the swept characteristic points, so that the signal reading efficiency is improved.
Setting section widthAnd center point coordinates +.>In half width->Determining boundary conditions and positioning of wave crest and wave trough positions, determining deviation of condition adjustment range in position characteristics of the range, and defining left and right boundaries as +.>The situation of the falling position is as follows: />In harmonic ipsilateral region->Ascending area or->A descent section; />Fall at->Ascending interval +.>Fall at->A descent section; />Fall at->Ascending interval +.>Fall at->A descent section; in the above case, the characteristic value of the position determination feature point is easily determined to be inaccurate when the range of the center point is too large, and the cycle is continued when the interval includes the valley-peak feature, and the step length to be adjusted is only related to the number of iterations. And carrying out feature analysis on the signal features in each sliding window based on a feature judgment rule, wherein the expression of the feature judgment rule is as follows:
，
in the method, in the process of the invention,for signal characteristics, +.>For ascending interval, ++>For descending interval, +_>For peak, add>Is the trough of the wave>For a frequency of +.>Echo signal received signal strength index value,/-, at the time>For a frequency of +.>Echo signal received signal strength index value,/-, at the time>For a frequency of +.>Echo signal received signal strength index value,/-, at the time>For a frequency of +.>The echo signal received signal strength index value and frequency are +.>Difference of echo signal receiving signal intensity index value,/-at the time>For a frequency of +.>The echo signal received signal strength index value and frequency are +.>Difference of echo signal receiving signal intensity index value,/-at the time>Representing an and relationship.
The left shift and right shift selection of the range is determined by the signal characteristics of the current interval, and fine adjustment of a small range is judged on the signal characteristics, so that all possible selected resonant frequency ranges can be approximated in the interval. After the waveform characteristics of the current position are determined, the moving direction of the interval can be determined.
Step S105, locating the peak position and the trough position of the signal feature with the frequency values in the interval boundaries at both sides of the center point, and determining the interval moving direction of the sliding window at the peak position and the trough position, so as to obtain the frequency peak value of the signal feature.
In this step, the selection direction is determined: set the firstThe individual peaks are characterized by resonance peaks->，/>For the total number of peak features, the peak features are respectively +.>Rising edge, th->Falling edge, th->And->The trough features are surrounded, so that no matter what waveform feature the current window is located in, the number position where the feature is located is judged.
The selection position is determined by the signal characteristics in the interval, and the selectable resonance peak position is judged and approached multiple times. The section selection direction judging method comprises the following steps:
，
in the method, in the process of the invention,for the number of peak characteristic rising intervals, +.>For the peak characteristic decreasing interval number, +.>Is the number of the wave troughs,is the total number of peak features.
Judging the basis: setting a frequency intervalDetermining that the current interval coordinates are added or subtracted in the determination direction of the respective feature>Whether the latter is still the current feature. By->The value is used for judging the feature number of the current position.
，
Wherein,、/>for interval width>For the frequency interval between the troughs, related to the sensor bandwidth,/->Is the current feature number, +.>、/>Are all current features.
After the interval moving direction is judged, the moving step length and the moving range are adaptively adjusted by a dichotomy, so that the method is suitable for sequence positioning at the moment and is used for solving the frequency sweep problem.
，
Wherein,representing the direction of movement +.>Is the center point coordinate.
In order to avoid noise interference, a fault self-adaptive adjustment module is added, a threshold value is set, the floating iteration times are recorded, and the times reach the threshold valueIf the frequency characteristic is not found, the step is traced back to the moving direction, and the overall flow of the algorithm characteristic is shown in fig. 3.
In summary, according to the method, echo signals of the surface acoustic wave sensor are obtained for preprocessing, a time-frequency spectrogram of the echo signals after wavelet transformation preprocessing is obtained, and a time-frequency data model is constructed; introducing a convolutional neural network to the signals according to the preprocessed one-dimensional signal classification and two-dimensional matrix reconstruction, and extracting characteristic processing for the two-channel input data; an intelligent frame detection algorithm for double-channel SConv-DD convolution nerve noise reduction is constructed according to signal characteristics after the fusion processing of signals, meanwhile, through characteristic analysis, the problems of overlong detection time, too many scanning frequency points and low system response of received signals are solved, a sliding type area algorithm is adopted to carry out an optimal solution range detection algorithm on frequency peak characteristics in an interval, noise reduction detection is carried out on the signals, fault detection processing is carried out, and efficiency and system response speed are improved.
Referring to fig. 4, a block diagram of an echo signal resonance peak range detection optimization system according to the present application is shown.
As shown in fig. 4, the echo signal resonance peak range detection optimization system 200 includes an acquisition module 210, a preprocessing module 220, an extraction module 230, a sweep frequency module 240, and an output module 250.
The acquiring module 210 is configured to acquire an echo signal of the wireless passive surface acoustic wave sensor; the preprocessing module 220 is configured to preprocess the echo signal, obtain a time-frequency spectrogram of the echo signal, and determine time-frequency data according to the time-frequency spectrogram; the extracting module 230 is configured to perform feature extraction on the time-frequency data according to a preset CNN model including a dual-channel convolution kernel, so as to obtain signal features, where the signal features include at least one waveform feature point and a frequency value corresponding to the at least one waveform feature point; the sweep frequency module 240 is configured to sweep frequency values corresponding to the at least one waveform feature point based on a preset sliding window, wherein the sliding window comprises a center point and interval boundaries positioned at two sides of the center point; and the output module 250 is configured to locate the peak position and the trough position of the signal feature according to the frequency values in the interval boundaries at two sides of the center point, and determine the interval moving direction of the sliding window at the peak position and the trough position, so as to obtain the frequency peak value of the signal feature.
It should be understood that the modules depicted in fig. 4 correspond to the various steps in the method described with reference to fig. 1. Thus, the operations and features described above for the method and the corresponding technical effects are equally applicable to the modules in fig. 4, and are not described here again.
In other embodiments, the present invention further provides a computer readable storage medium, on which a computer program is stored, where the program instructions, when executed by a processor, cause the processor to perform the echo signal resonance peak range detection optimization method in any of the above method embodiments;
as one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions configured to:
acquiring an echo signal of a wireless passive surface acoustic wave sensor;
preprocessing the echo signal to obtain a time-frequency spectrogram of the echo signal, and determining time-frequency data according to the time-frequency spectrogram;
performing feature extraction on the time-frequency data according to a preset CNN model containing a double-channel convolution kernel to obtain signal features, wherein the signal features comprise at least one waveform feature point and a frequency value corresponding to the at least one waveform feature point;
sweeping the frequency value corresponding to the at least one waveform characteristic point based on a preset sliding window, wherein the sliding window comprises a center point and interval boundaries positioned at two sides of the center point;
and positioning the peak position and the trough position of the signal characteristic according to the frequency values in the interval boundaries at the two sides of the center point, and determining the interval moving direction of the sliding window at the peak position and the trough position to obtain the frequency peak value of the signal characteristic.
The computer readable storage medium may include a storage program area and a storage data area, wherein the storage program area may store an operating system, at least one application program required for a function; the stored data area may store data created from the use of the echo signal formant range detection optimization system, and the like. In addition, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer readable storage medium optionally includes a memory remotely located with respect to the processor, the remote memory being connectable to the echo signal formant range detection optimization system through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, as shown in fig. 5, where the device includes: a processor 310 and a memory 320. The electronic device may further include: an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 may be connected by a bus or other means, for example in fig. 5. Memory 320 is the computer-readable storage medium described above. The processor 310 executes various functional applications of the server and data processing by running nonvolatile software programs, instructions and modules stored in the memory 320, i.e. implements the echo signal resonance peak range detection optimization method of the above-described method embodiment. The input device 330 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the echo signal formant range detection optimization system. The output device 340 may include a display device such as a display screen.
The electronic equipment can execute the method provided by the embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method. Technical details not described in detail in this embodiment may be found in the methods provided in the embodiments of the present invention.
As an implementation manner, the electronic device is applied to an echo signal resonance peak range detection optimization system, and is used for a client, and includes: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executable by the at least one processor to enable the at least one processor to:
acquiring an echo signal of a wireless passive surface acoustic wave sensor;
preprocessing the echo signal to obtain a time-frequency spectrogram of the echo signal, and determining time-frequency data according to the time-frequency spectrogram;
performing feature extraction on the time-frequency data according to a preset CNN model containing a double-channel convolution kernel to obtain signal features, wherein the signal features comprise at least one waveform feature point and a frequency value corresponding to the at least one waveform feature point;
sweeping the frequency value corresponding to the at least one waveform characteristic point based on a preset sliding window, wherein the sliding window comprises a center point and interval boundaries positioned at two sides of the center point;
and positioning the peak position and the trough position of the signal characteristic according to the frequency values in the interval boundaries at the two sides of the center point, and determining the interval moving direction of the sliding window at the peak position and the trough position to obtain the frequency peak value of the signal characteristic.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on such understanding, the foregoing technical solutions may be embodied essentially or in part in the form of a software product, which may be stored in a computer-readable storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the various embodiments or methods of some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. The echo signal resonance peak range detection optimization method is characterized by comprising the following steps of:
acquiring an echo signal of a wireless passive surface acoustic wave sensor;
preprocessing the echo signal to obtain a time-frequency spectrogram of the echo signal, and determining time-frequency data according to the time-frequency spectrogram;
performing feature extraction on the time-frequency data according to a preset CNN model containing a double-channel convolution kernel to obtain signal features, wherein the signal features comprise at least one waveform feature point and a frequency value corresponding to the at least one waveform feature point;
sweeping the frequency value corresponding to the at least one waveform characteristic point based on a preset sliding window, wherein the sliding window comprises a center point and interval boundaries positioned at two sides of the center point;
and positioning the peak position and the trough position of the signal characteristic according to the frequency values in the interval boundaries at the two sides of the center point, and determining the interval moving direction of the sliding window at the peak position and the trough position to obtain the frequency peak value of the signal characteristic.
2. The method for optimizing detection of resonance peak range of echo signal according to claim 1, wherein preprocessing the echo signal to obtain a time-frequency spectrogram of the echo signal, and determining time-frequency data according to the time-frequency spectrogram comprises:
carrying out one-dimensional vector classification and two-dimensional matrix construction on the echo signals to obtain at least one-dimensional vector and two-dimensional matrixWherein->Is the real part of the channel matrix, < >>Is the imaginary part of the channel matrix;
discretizing the echo signals according to a wavelet function, and weighting the discretized echo signals to obtain a time-frequency spectrogram of the echo signals, wherein the discretized echo signals are expressed as follows:
，
，
in the method, in the process of the invention,as a basis function +.>、/>All are discrete points, and are filled with->Is a time domain parameter->For the wavelet transform coefficients,for discretizing the basis function +.>For sampling signals +.>Is the derivative of the basis function>Transforming the expression for the basis function;
the expression for weighting the discretized echo signals is as follows:
，
in the method, in the process of the invention,for complex multiplication times, ++>To collect the signal +.>For the weighting factor>Is->Is used for the discrete transformation of (a),k is a fine coefficient as a response function;
and determining time-frequency data according to the time-frequency spectrogram.
3. The method according to claim 1, wherein before the feature extraction is performed on the time-frequency data according to a preset CNN model including a two-channel convolution kernel, the method further comprises:
performing iterative training on a double-channel SConv-DD convolution neural network based on historical time-frequency data and historical signal characteristics corresponding to the historical time-frequency data to obtain a CNN model containing a double-channel convolution kernel, wherein the expression of the SConv-DD convolution neural network is as follows:
，
，
in the method, in the process of the invention,transpose of H signal, +.>For the output of the l+1 layer, +.>For piecewise linear soft sign operator, < >>For mathematical transformation, ++>For reorganizing different input data according to different channels, the +.>Is Hadamard product, < >>Is a residual feature;
the loss function of the CNN model is as follows:
，
in the method, in the process of the invention,to be damaged byLoss function parameters->For a channel matrix>For the input relevant signal +.>For training parameter sets, ++>For iterative index, ++>For hiding layer first layer->For the output of layer l, +.>Is a loss difference;
the training parameter set is as follows:，/>weight matrix for SConv-DD convolutional neural network,>the shift matrix of the neural network is convolved for the SConv-DD.
4. The method according to claim 1, wherein before the feature extraction is performed on the time-frequency data according to a preset CNN model including a two-channel convolution kernel, the method further comprises:
and carrying out noise reduction processing on the time frequency data according to a preset noise reduction model, wherein the expression of the noise reduction model is as follows:
，
，
，
in the method, in the process of the invention,is an intermediate signal +.>For the noise reduction function->Is a product coefficient>For the channel set +.>For variance->、/>All are training parameters, and are added with->For transmitting symbol vectors, < >>For the ith input noise, < >>Is a single channel difference, I is a diagonal matrix,is an intermediate parameter->Transpose of channel matrix, +.>For the noise reducer model, ++>For the input relevant signal +.>For residual error->To accept symbol vectors>Is the variance estimate.
5. The method according to claim 1, wherein after sweeping the frequency value corresponding to the at least one waveform feature point based on a preset sliding window, the method further comprises:
and carrying out feature analysis on the signal features in each sliding window based on a feature judgment rule, wherein the expression of the feature judgment rule is as follows:
，
in the method, in the process of the invention,for signal characteristics, +.>For ascending interval, ++>For descending interval, +_>For peak, add>Is the trough of the wave>Is of frequency ofEcho signal received signal strength index value,/-, at the time>For a frequency of +.>The echo signal received signal strength index value at that time,for a frequency of +.>Echo signal received signal strength index value,/-, at the time>For a frequency of +.>The echo signal received signal strength index value and frequency are +.>Difference of echo signal receiving signal intensity index value,/-at the time>For a frequency of +.>The echo signal received signal strength index value and frequency are +.>The difference in the received signal strength index values of the echo signals,representing an and relationship.
6. The method of claim 1, wherein determining a moving direction of the sliding window between the peak position and the trough position comprises:
determining the section moving direction of the sliding window at the peak position and the trough position according to a section selection direction judging rule, wherein the expression of the section selection direction judging rule is as follows:
，
in the method, in the process of the invention,for the number of peak characteristic rising intervals, +.>For the peak characteristic decreasing interval number, +.>For the number of wave troughs->Is the total number of peak features.
7. An echo signal resonance peak range detection optimization system, comprising:
the acquisition module is configured to acquire echo signals of the wireless passive surface acoustic wave sensor;
the preprocessing module is configured to preprocess the echo signals to obtain a time-frequency spectrogram of the echo signals, and determine time-frequency data according to the time-frequency spectrogram;
the extraction module is configured to perform feature extraction on the time-frequency data according to a preset CNN model containing a double-channel convolution kernel to obtain signal features, wherein the signal features comprise at least one waveform feature point and a frequency value corresponding to the at least one waveform feature point;
the frequency sweep module is configured to sweep the frequency value corresponding to the at least one waveform characteristic point based on a preset sliding window, wherein the sliding window comprises a center point and interval boundaries positioned at two sides of the center point;
the output module is configured to position the peak position and the trough position of the signal feature according to the frequency values in the interval boundaries at the two sides of the center point, determine the interval moving direction of the sliding window at the peak position and the trough position, and obtain the frequency peak value of the signal feature.
8. An electronic device, comprising: at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 6.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410160351.2A CN117708657B (en) | 2024-02-05 | 2024-02-05 | Echo signal resonance peak range detection optimization method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410160351.2A CN117708657B (en) | 2024-02-05 | 2024-02-05 | Echo signal resonance peak range detection optimization method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117708657A true CN117708657A (en) | 2024-03-15 |
| CN117708657B CN117708657B (en) | 2024-06-18 |
Family
ID=90155670
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410160351.2A Active CN117708657B (en) | 2024-02-05 | 2024-02-05 | Echo signal resonance peak range detection optimization method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117708657B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108566260A (en) * | 2018-02-01 | 2018-09-21 | 西安电子科技大学 | It is a kind of based on the concealed communication method for disturbing point multiple access |
| WO2022104868A1 (en) * | 2020-11-19 | 2022-05-27 | 华中师范大学 | Millimeter wave radar-based non-contact real-time vital sign monitoring system and method |
| CN114781457A (en) * | 2022-04-29 | 2022-07-22 | 西安电子科技大学 | Time-frequency domain interference suppression method based on automatic encoder |
| CN115060407A (en) * | 2022-06-14 | 2022-09-16 | 华东交通大学 | Flexible multi-mode sensing device for multi-dimensional wheel-rail force measurement and fusion method |
| CN116244631A (en) * | 2023-03-08 | 2023-06-09 | 重庆交通大学 | Pile foundation multi-defect identification method and system based on double-channel convolutional neural network |
| CN117111000A (en) * | 2023-03-24 | 2023-11-24 | 西安电子科技大学 | SAR comb spectrum interference suppression method based on dual-channel attention residual network |
-
2024
- 2024-02-05 CN CN202410160351.2A patent/CN117708657B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108566260A (en) * | 2018-02-01 | 2018-09-21 | 西安电子科技大学 | It is a kind of based on the concealed communication method for disturbing point multiple access |
| WO2022104868A1 (en) * | 2020-11-19 | 2022-05-27 | 华中师范大学 | Millimeter wave radar-based non-contact real-time vital sign monitoring system and method |
| CN114781457A (en) * | 2022-04-29 | 2022-07-22 | 西安电子科技大学 | Time-frequency domain interference suppression method based on automatic encoder |
| CN115060407A (en) * | 2022-06-14 | 2022-09-16 | 华东交通大学 | Flexible multi-mode sensing device for multi-dimensional wheel-rail force measurement and fusion method |
| CN116244631A (en) * | 2023-03-08 | 2023-06-09 | 重庆交通大学 | Pile foundation multi-defect identification method and system based on double-channel convolutional neural network |
| CN117111000A (en) * | 2023-03-24 | 2023-11-24 | 西安电子科技大学 | SAR comb spectrum interference suppression method based on dual-channel attention residual network |
Non-Patent Citations (3)
| Title |
|---|
| 刘伯权等: "声表面波谐振器回波信号的频率估计", 电子与信息学报, vol. 41, no. 09, 15 September 2019 (2019-09-15), pages 2087 - 2094 * |
| 张卓然;梁福来;李浩楠;吕昊;祁富贵;安强;王健琪;: "双通道SFCW生物雷达对静止人体目标探测方法的研究", 中国医疗设备, no. 09, 25 September 2016 (2016-09-25) * |
| 黄安永等: "self-similar design for stretchable wireless LC strain sensors", SENSORS ABD ACTUATORS A：PHYSICAL, vol. 224, 1 April 2015 (2015-04-01), pages 36 - 42 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117708657B (en) | 2024-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110865357B (en) | Laser radar echo signal noise reduction method based on parameter optimization VMD | |
| CN110348288B (en) | Gesture recognition method based on 77GHz millimeter wave radar signal | |
| CN103439692B (en) | STAP method based on wide symmetrical characteristic of covariance matrix | |
| CN111783558A (en) | Satellite navigation interference signal type intelligent identification method and system | |
| Du et al. | Secondary radar signal processing based on deep residual separable neural network | |
| CN110996343B (en) | Intelligent recognition system and recognition method of interference recognition model based on deep convolutional neural network | |
| CN111723329B (en) | Seismic phase feature recognition waveform inversion method based on full convolution neural network | |
| CN107024685A (en) | A kind of gesture identification method based on apart from velocity characteristic | |
| CN112949387B (en) | Intelligent anti-interference target detection method based on transfer learning | |
| WO2023010886A1 (en) | Indoor non-contact human body action recognition method and system | |
| CN101881826A (en) | Scanning-mode sea clutter local multi-fractal target detector | |
| CN108957403B (en) | Gaussian fitting envelope time delay estimation method and system based on generalized cross correlation | |
| CN109324315A (en) | Space-time adaptive based on double level block sparsity handles radar clutter suppression method | |
| CN104331583B (en) | A kind of Multifractal Modeling method based on Observed sea clutter | |
| CN105429719B (en) | Based on power spectrum and multi-scale wavelet transformation analysis high reject signal detection method | |
| Wang et al. | Multi-classification of UWB signal propagation channels based on one-dimensional wavelet packet analysis and CNN | |
| CN113341383A (en) | Radar anti-interference intelligent decision method based on DQN algorithm | |
| CN109001702B (en) | Carrier-free ultra-wideband radar human body action identification method | |
| CN115438708A (en) | Classification and identification method based on convolutional neural network and multi-mode fusion | |
| CN109212472B (en) | Indoor wireless positioning method and device in noise-oriented environment | |
| CN117708657B (en) | Echo signal resonance peak range detection optimization method and system | |
| CN115469303A (en) | Cognitive biological radar method and device for detecting human body posture and vital signs | |
| CN107765259B (en) | Power transmission line laser ranging signal denoising method based on lifting wavelet improved threshold | |
| CN109905190B (en) | Modeling method for low-frequency ground wave propagation time delay variation characteristic | |
| CN111551893A (en) | Deep learning and neural network integrated indoor positioning method |
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 |