Multi-band signal recovery system design oriented to communication and inductance integration
By using sparseness estimation and exponential step adaptive matching tracking algorithm in the modulated broadband converter system, combined with the adaptive adjustment channel number module, the problems of unknown sparseness and low algorithm efficiency in broadband spectrum perception are solved, and high efficiency is achieved , accurate signal recovery and spectrum perception.
Patent Information
- Application Number
- CN202411251300.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-07
- Publication Date
- 2024-12-03
AI Technical Summary
The existing modulated broadband converter systems have problems such as unknown or variable sparseness, low algorithm efficiency, weak adaptive adjustment capabilities, and inability to meet the requirements of high-precision and high-efficiency reconstruction in broadband spectrum perception scenarios.
Using the exponential variable step size adaptive matching tracking (ES-EVssAMP) algorithm based on sparse estimation and the adaptive adjustment channel number module, the step size is quickly iterated through the sparse estimation module and the exponential function adaptive adjustment, combining structural adaptive Adjust the number of channels to achieve sparse sampling and signal reconstruction.
It improves the overall operating efficiency and flexibility of the algorithm, can position the spectral position more accurately, realizes more accurate signal reconstruction, and reduces the sampling rate and waste of communication resources.
Smart Images

Figure FT_1 
Figure FT_2 
Figure FT_3
Abstract
Description
Technical Field
[0001] The invention belongs to the field of multi-band signal restoration, and in particular relates to a multi-band signal restoration system design based on synaesthesia integration. Background Art
[0002] As communication technology develops from 5G to 6G, people begin to envision a new expectation of all-round coverage, environmental intelligence, sensory connection, and synaesthesia fusion. A notable feature of 6G technology is its multi-domain integration, which means that it not only covers ground networks, but also extends to the air and space. Satellite Internet of Things constitutes a core part of this vision, and with the help of satellite's global coverage capability, it realizes cross-regional and cross-border information exchange. Satellite Internet of Things, composed of multiple satellites, has a wide coverage range and can achieve global network interconnection. In order to achieve the purpose of multi-domain communication, satellites need to have more advanced perception functions to ensure the stability and reliability of communication links. On the other hand, due to the large number of nodes connected, this increases the use of spectrum resources, making the already scarce wireless spectrum resources even scarcer. From another perspective, the amount of data generated by satellite Internet of Things is extremely large, covering a variety of communication data types, so large-scale data processing and analysis are required to extract valuable information and insights from these massive data. Therefore, Integrated Sensing and Communication (ISAC) has become a hot topic in the current research field of wireless communication technology. ISAC has the synergy between communication and spectrum positioning perception, which helps to better allocate various communication resources and enhance network stability. It can also improve the efficiency of data transmission and overall network performance, and reduce data latency and packet loss. As the satellite Internet of Things architecture gradually matures, the interawareness integration framework still faces the challenge of scarce spectrum resources. Therefore, in order to improve the efficiency of spectrum use, it is necessary to adopt broadband spectrum sensing technology in a wider frequency range.
[0003] In addition, as the wireless communication environment becomes more and more complex and the spectrum range continues to expand, the traditional sampling speed has brought great pressure to the front-end equipment of the receiver. Existing digital receiving equipment can no longer meet the actual needs of broadband signal processing. In order to meet this challenge, a modulated wide band converter (MWC) system based on compressed sensing theory was designed, which can effectively compress and sample sparse signals. Its main advantages include: it can support sub-Nyquist sampling, it can move multiple signals at the same time to achieve the effect of down-conversion, it has relatively low requirements for the matching degree of the system and signal model, and it is very suitable for various application scenarios. MWC can not only solve the problems of spectrum perception, but also realize distortion-free communication. It can not only locate multi-band signals at the perception level, but also further restore the original signal at the communication level. Therefore, the MWC system can have the functions of communication and perception at the same time, which meets the technical characteristics of ISAC and applies the MWC system to the application of synaesthesia integration.
[0004] The Orthogonal Matching Pursuit (OMP) algorithm in the MWC system requires known sparsity conditions for reconstruction, but in actual conditions, the sparsity is unknown or changes. In response to this problem, some scholars have proposed the Sparsity Adaptive Matching Pursuit (SAMP) algorithm and made many improvements to this algorithm. From the perspective of algorithm and structure, the implementation process of the MWC system is overly dependent on prior conditions, and the improved algorithm rarely considers the issues of algorithm efficiency and algorithm flexibility. Due to the limitation of fixed step size, the algorithm cannot meet the requirements of high-precision and high-efficiency reconstruction. At the same time, in actual broadband spectrum sensing scenarios, due to the dynamic changes in the number of signals, its adaptive adjustment ability is relatively weak. Due to multiple factors such as the time-varying nature of the channel and the interference of noise, the performance of the system is significantly reduced, and the parameters of the model cannot be updated in time according to the real-time changes in the environment. Therefore, this limits the application of the MWC system in broadband spectrum sensing scenarios.
[0005] Based on the above problems, the present invention improves the algorithm and structure on the basis of the aliasing modulation broadband converter. The algorithm proposes an exponential variable step size adaptive matching pursuit (ES-EVssAMP, TheEstimation of the Sparsity-Exponential Variable Step Size Adaptive Matching Pursuit) algorithm based on sparsity estimation, and the structure designs an adaptive aliasing modulation broadband converter system. This system does not take the sparsity of the signal as a known standard. Its main purpose is to improve the sampling efficiency and sparsely sample the signal. In view of the fact that the signal is full of a lot of noise, the compressed sensing theory becomes an effective means to achieve sparse sampling. The control system uses the corresponding point error of the sparse sampling and the reconstructed signal as its core control parameter. When the reconstruction error fails to meet the established standard, the system will be adjusted by increasing the number of channels until the established requirements are met. In a scenario where the number of signal sub-bands is small but the recovery speed is fast, the amount of data processed can be reduced by reducing the number of channels. In spectrum sensing technology, it can also meet the communication and perception needs of synaesthesia integration. Summary of the invention
[0006] The purpose of the present invention is to provide a multi-band signal recovery system for synaesthesia integration. The ES-EVssAMP algorithm used mainly targets the algorithm efficiency and fixed step-by-step iteration problems of the SAMP algorithm, adds a sparsity estimation module and improves the fixed step-by-step iteration; and performs adaptive channel adjustment on the modulation broadband converter structure. To achieve the above purpose, the specific improvements are as follows:
[0007] 1) A sparsity estimation module is added to the algorithm. This module can reduce the iterative process to a few steps for fast iteration, and determine the conversion of large and small scale iterations through the designed criteria, solving the problems of slow iteration speed and low algorithm efficiency.
[0008] 2) Improvements are made to the original algorithm that requires a fixed step size for progressive iteration. The function characteristics of the exponential function are used to adaptively adjust the step size, which solves the problem of not being able to meet both high precision and high efficiency due to the influence of the step size, and also reflects the flexibility of the algorithm.
[0009] The steps of the improved algorithm (ES-EVssAMP) are as follows:
[0010] S1. Input sensor matrix A m×n , observation matrix y m×1 , step size S, exponential function parameter β, sparsity estimation parameter δ k , a first threshold parameter θ1 and a second threshold parameter θ2, wherein the observation matrix is any column in the sensor matrix;
[0011] S2. Initialization parameters, including iteration stage stage = 1, iteration number t = 1, sparsity estimate K0 = S, residual vector r0 = y, initial index set Candidate index set And the initial support set
[0012] S3. Calculate the initial index set Λ0=max(A T y,K0), where the max(a,b) function means taking the b largest elements in the matrix a;
[0013] S4. Calculate the estimated sparsity, if Then K0=K0+S, return to S3; otherwise K1=K0, Execute S5;
[0014] S5. Calculate a new initial index set Λ t =max(|A T r t-1 |,K t );
[0015] S6. Merge and update candidate index set C t =F t-1 ∪Λ t ;
[0016] S7. Obtain the minimum estimated signal value
[0017] S8. Get the current support set
[0018] S9. Update the final estimate and residual of the signal
[0019] S10. If || r c ||2≤θ1, then proceed to step 9, otherwise proceed to S13;
[0020] S11. If || r c If ||2≤θ2, stop iteration and output the result, otherwise proceed to S12;
[0021] S12. If || r c ||2≥||r t-1 ||2 K t+1 =K t +S,stage=stage+1F t new =F t , t=t+1, otherwise stage=stage+1 F tnew =F t , t=t+1. Then return to S5;
[0022] S13. Output
[0023] 3) An adaptive passband adjustment module is introduced in the structure. By comparing the sampling points and the corresponding signal points after reconstruction, the error is used to dynamically adjust the number of channels in the system, thereby achieving the purpose of automatically adjusting the number of channels.
[0024] A computer device of the present invention comprises a processor and a computer program running on the processor, wherein the processor is used to execute the steps of the multi-band signal recovery system.
[0025] The present invention solves the problems caused by the original SAMP algorithm for reconstruction and recovery:
[0026] 1) Improve the overall operating efficiency of the algorithm based on the sparsity estimation module.
[0027] 2) The exponential function adaptive variable step size can improve the flexibility of the algorithm, so that the accuracy can be adjusted according to the specific requirements of the actual situation while improving the efficiency of the algorithm. It can locate the spectrum position more accurately and reconstruct the signal more accurately.
[0028] 3) Structural improvements can reduce the sampling rate and provide an appropriate number of channels to prevent waste of communication resources. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 is a schematic diagram of the structure of a modulated broadband converter;
[0030] Figure 2 It is the basic structure of the controller;
[0031] Figure 3 It is the process of improving the system for spectrum sensing;
[0032] Figure 4 yes Figure 5 Schematic diagram of the comparison of OMP, SAMP, SPAMP, K-Means SAMP, and ES-EVssAMP reconstruction algorithms in various aspects;
[0033] Figure 5 It is the MWC recovery spectrum with 30, 40 and adaptive channel numbers. DETAILED DESCRIPTION
[0034] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.
[0035] The idea of the present invention is to reconstruct the signal using an aliasing modulation broadband converter, and considering the requirements for high efficiency and high precision of signal reconstruction, an exponential variable step size adaptive matching pursuit (ES-EVssAMP) algorithm based on sparsity estimation is used, in which a first set step size and a second set step size are added, and the algorithm can be quickly iterated by using the first set step size to update the current sparsity, and the current sparsity can be slightly adjusted by the second set step size. In terms of structure, an adaptive channel number adjustment module is introduced, and the number of channels of the system is dynamically adjusted by comparing the sampling point and the corresponding signal point after reconstruction, thereby achieving the purpose of automatically adjusting the number of channels.
[0036] The broadband sparse analog signal is a multi-band signal. x(t) is a signal with limited bandwidth, continuous square integrability and limited energy. Its main feature is that all its energy is concentrated in one or more separated frequency bands, which can simulate real signals. Its spectrum is bounded, and the Fourier transform of x(t) is as follows:
[0037]
[0038] Assuming that all sub-bands of X(jω) have equal bandwidth BHz, the number is K, and W / 2 is the bandwidth of x(t), then KB is much smaller than 2 times of the signal bandwidth. The purpose is to make the spectrum complement aliasing and provide a prerequisite for signal recovery, otherwise the signal cannot be recovered. The modulation broadband converter system can convert high-frequency signals into low-frequency signals for processing and sampling. The system block diagram is as follows Figure 1 shown.
[0039] x(t) is the original signal, p i (t) is the mixing sequence, whose value jumps between ±1, and the jump frequency is set equal to the Nyquist sampling frequency. H(f) is an ideal low-pass filter, and the sampling period is T p , the cut-off frequency is ω s / 2=2π / T p , the Fourier transform of the mixing sequence is:
[0040]
[0041] When the signal passes through the mixing sequence, the relationship between the time domain and the frequency domain is as follows:
[0042]
[0043] Because the premise of time domain and frequency domain conversion is that the signal is bounded, after filtering and time-frequency conversion, the output yi The DTFT transformation of (n) is:
[0044]
[0045] Where L0 represents all possible non-zero terms of X(f) in the positive frequency part within the signal bandwidth, that is:
[0046]
[0047] Formula (4) converts the sampling sequence y into i The spectrum of [n] is linked to the spectrum of the unknown signal X(f) to construct a classical reconstruction formula similar to compressed sensing, which is rewritten in matrix form.
[0048]
[0049] y(f) is an m×1 vector, where the i-th element is z(f) is a L×1 vector, where L=2L0+1 and the i-th element is
[0050]
[0051] A is an m×L matrix, and the element in the i-th row and l-th column is
[0052] The broadband signal is modulated by the function p i After the action of (t), the period extension is realized in the frequency domain. After passing through the low-pass filter, the signal energy of each sub-band signal in the baseband part is obtained. At this time, the sub-band signals are mixed with each other. The data stream y[n] is obtained by low-speed sampling to obtain Q
[0053]
[0054] where y[n]=[y1[n],…,y m [n]] T Indicates that at sampling time nT s The corresponding information sequence. This expression connects the time domain and the frequency domain, and realizes the construction of the frame structure through the information sequence y[n]. The matrix Q is decomposed as follows.
[0055] Q=VV H (9)
[0056] A corresponding frame structure V can be obtained. The signal support set information is obtained through the compressed sensing theory. Through the obtained signal support set information, the mixed sub-band signals are separated to obtain the spectrum information of the original signal, thereby finding the spectrum holes and completing the broadband spectrum sensing process.
[0057] The MWC system can directly obtain the signal support set, and can directly use existing ADC equipment in hardware implementation. At the same time, it does not need time synchronization, which makes the implementation simpler. At the same time, the emergence of the MWC prototype has attracted widespread attention and has become one of the preferred technologies for realizing broadband spectrum sensing based on sub-Nyquist sampling. It is also the main reason why the present invention takes MWC as the research object.
[0058] In terms of algorithm, it is difficult for the SAMP algorithm to ensure high efficiency and high accuracy in the signal reconstruction process. In this invention, we propose an ES-EVssAMP algorithm. The algorithm uses sparsity estimation and adaptive variable step size modules to improve the iterative process.
[0059] Since the sparsity is usually unknown in practical applications, the fixed step size is usually set to 1, which is accurate but takes a lot of time and reduces the reconstruction efficiency. Therefore, an adaptive sparsity method based on the RIP condition is proposed. The criterion is as follows: for any signal with a sparsity of K, the perception matrix A satisfies the RIP condition, when δ k ∈(0,1), if K0≥K, then the following equation is satisfied: where supp is the true support set and y is the observation matrix.
[0060]
[0061] (10) can be used as the condition for terminating the iteration in the sparsity estimation. In order to avoid overestimation, the initial value of K0 is set to 1. Otherwise, K0 = K0 + S. Repeat this process until the formula is valid. The sparsity estimation is used as the termination condition for large-scale iteration. The reconstruction of traditional signals uses the estimated sparsity as the known sparsity after realization, which indirectly improves the processing performance of the algorithm.
[0062] When we use large-scale iterations to approach true sparsity, we need to perform small-scale iterations to get closer to true sparsity. In this case, a variable step size module is used. When the support set size approaches true sparsity, the step size is reduced to improve signal reconstruction accuracy. This process requires the step size to gradually decrease and the step size is greater than 0, so an exponential function is chosen. Consider the following exponential function:
[0063] y=β x ,β∈(0,1),x≥0 (11)
[0064] Where β is a fixed constant, (11) is monotonically decreasing and the function value is greater than 0 when the conditions in the formula are met. Therefore, the exponential function can be used to adjust the step size. The initial step size is selected as a larger value, and the length of the support set is K = K + S. The stage represents the number of times the step size is reduced. When the step size needs to be reduced, the length of the support set is Indicates rounding up. When the sparsity estimate is close to the true sparsity, the support set length becomes smaller and smaller as the number of stages increases until the final iteration condition is met. By selecting the step size, the efficiency of the algorithm is guaranteed and the flexibility of the algorithm is improved. The modified algorithm is detailed in the content of the invention.
[0065] Secondly, an adaptive channel number adjustment structure is added to the structure. In the application environment of interawareness integration, although the number of users accessed will change, the number of subbands remains unchanged for a period of time. Due to the limited system bandwidth, when the number of channels is too large, it will affect indicators such as channel utilization and bit error rate, thereby causing resource waste and reducing system performance. Therefore, based on MWC-based broadband spectrum sensing, the present invention proposes a method for sparse sampling of signals (lower than Nyquist sampling). By comparing the sampling points and the corresponding signal points after reconstruction, the error can be used to dynamically adjust the number of channels of the system, thereby achieving the purpose of automatically adjusting the number of channels. The process of channel addition is shown in the figure.
[0066] The error calculation method is based on the mean square error:
[0067]
[0068] Where z i is the coefficient sampling point; Represents the point after the corresponding sampling point is reconstructed, and N is the number of samples to be compared. The structural block diagram of the main controller is shown in the figure. The figure shows the process of the system for broadband spectrum sensing:
[0069] The system can achieve the purpose of adaptively adjusting the number of channels through the above process. Although the number of channels may be increased more than the most suitable number of channels in this process, it has no effect on the result and can find the suitable number of channels more quickly.
[0070] In order to verify the improved algorithm and structural performance, a noisy electrical signal is simulated using formula (13). i ,B i ,f i ,σ i Represents the i-th energy coefficient, bandwidth, carrier spectrum and delay time. These values are randomly selected. The maximum bandwidth B is set to 50MHz, and the Gaussian white noise is n(t). The signal-to-noise ratio is 10dB. In order to ensure that there is no loss of generality, the following parameters are set. The period length of the AMWC system sequence is M=195, and the period length is T p The number of sub-bands N is 6. In each signal [-f NYQ / 2,f NYQ / 2] uniformly randomly selects carriers, f NYQThe carrier spectrum of each sub-band is evenly distributed in the entire bandwidth where the signal is located. In the ES-EVssAMP algorithm, we set the following parameter values. The two threshold parameters θ1 and θ2 are 10 -6 and 0. The base β of the exponential function in the variable order module is 0.5, and the δ of the sparsity estimation part is k is 0.5. The initial step size of the ES-EVssAMP algorithm is 5.
[0071]
[0072] The support set reconstruction success rate of the algorithm can be calculated using (14). O ,S R represent the support sets of the original signal and the reconstructed signal respectively.
[0073]
[0074] The results of the reconstruction of OMP, SAMP, SPAMP, K-means SAMP and ES-EVssAMP algorithms in MWC are shown in the figure. It can be seen from the figure that the SAMP step size is large. Although the number of signal iterations is small and the reconstruction time is short, the effectiveness of the reconstruction support set is still low, resulting in substandard reconstruction results. As the number of SAMP steps decreases, although the reconstruction success rate is higher and the effect is improved, the excessive number of signal iterations will lead to longer reconstruction time. Although the SPAMP and K-means SAMP algorithms are similar to the enhanced versions in reconstruction accuracy, they exceed the ES-EVssAMP algorithm in reconstruction time. Under the same conditions, the ES-EVssAMP algorithm can reconstruct signals faster and more accurately. Its success rate is similar to that of the SAMP algorithm, and it requires the least number of steps. This is because the SAMP algorithm lacks a mechanism to estimate sparsity. SPAMP and K-means SAMP algorithms have difficulty adapting to sparsity due to constant step size, weak selection and insufficient regularization. The ES-EVssAmp algorithm integrates a module for estimating sparsity and variable step size into adaptive dynamic processing, which improves the accuracy of the support set length, achieves sparsity, and improves the efficiency of operation. Combining these two techniques improves the usability of the algorithm, and the results confirm the effectiveness of signal recovery.
[0075] At the same time, for the structure, the system sets a total of 100 channels, the initial channels are 30, and the number of channels increases to 10. The ES-EVssAMP algorithm is used to recover the multi-band signal and locate the spectrum. The spectrum results are shown below. It can be seen that the system can accurately realize broadband spectrum perception without manually adjusting the number of channels. The channel can be adaptively adjusted to 70 to restore the spectrum more accurately. In the case of poor channel quality or large noise, although some small side lobes will appear in the system, the user's usage frequency can be accurately sensed without interfering with the user, thereby ensuring accurate signal transmission. If the number of channels is too small, the system cannot accurately sense the user's usage status and effectively restore the signal. Therefore, when the number of sub-bands increases, the number of channels must also increase to ensure that the user is not interfered with, which further confirms the practicality of this system.
[0076] In the application environment of synaesthesia integration, the number of users accessed is not fixed, and the number of sub-bands may be fixed only within a period of time. Therefore, based on the spectrum perception based on aliasing modulation broadband converter, the present invention introduces a method of sparse sampling of signals, compares the sampling points with the corresponding signal points after reconstruction, and controls the number of channels of the system according to the error, so as to achieve the purpose of adaptively adjusting the number of channels; by analyzing the increase in the number of channels, determining the optimal increase number, and finally conducting verification experiments, simulation experiments prove that the combination of ES-EVssAMP algorithm can not only quickly and accurately restore the spectrum of broadband sparse signals, but also flexibly determine the number of channels. It further verifies the significance of the improved algorithm and improved structure for the design of multi-band signal recovery system for synaesthesia integration.
[0077] The present invention has the following characteristics and significant improvements:
[0078] 1. The algorithm uses a sparsity estimation module and an exponential function adaptive variable step size module. The sparsity estimation shortens the original iteration process to a few steps, greatly improving the efficiency of the algorithm and becoming a condition for large-scale iteration. When the sparsity estimation is close to the actual sparsity, a small-scale iteration is performed, and the exponential function adaptive variable step size module automatically adjusts the step size to ensure the reconstruction accuracy and greatly improve the flexibility of the algorithm.
[0079] 2. Use dual threshold iteration termination conditions to prevent the algorithm from overfitting while achieving higher accuracy.
[0080] 3. Based on MWC-based broadband spectrum sensing, a method for sparse sampling (lower than Nyquist sampling) of signals is proposed. By comparing the sampling points and the corresponding signal points after reconstruction, the error can be used to dynamically adjust the number of channels in the system, thereby achieving the purpose of automatically adjusting the number of channels.
Claims
1. An adaptive sparsity method based on the RIP condition is proposed. The criterion is as follows: for any signal with sparsity K, the perception matrix A satisfies the RIP condition when δ k ∈(0,1), if K0≥K, then the following equation is satisfied: where supp is the true support set and y is the observation matrix. It can be used as the condition for terminating the iteration in sparsity estimation. To avoid overestimation, the initial value of K0 is set to 1. Otherwise, K0 = K0 + S. Repeat this process until the formula is valid. Use sparsity estimation as the termination condition for large-scale iteration. The reconstruction of traditional signals uses the estimated sparsity as the known sparsity after realization, which indirectly improves the processing performance of the algorithm.
2. When large-scale iterations are used to approach true sparsity, small-scale iterations need to be performed to get closer to true sparsity. In this case, a variable step size module is used. When the support set size approaches true sparsity, the step size is reduced to improve signal reconstruction accuracy. This process requires the step size to be gradually reduced and the step size is greater than 0, so an exponential function is chosen. Consider the following exponential function: y=β x ,β∈(0,1),x≥0 (2) Where β is a fixed constant, which is monotonically decreasing and has a function value greater than 0 when the conditions in the formula are met. Therefore, the step size can be adjusted using this exponential function. Therefore, the step size can be adjusted using this exponential function. The initial step size is selected as a larger value, and the length of the support set is K = K + S. The stage indicates the number of times the step size is reduced. When the step size needs to be reduced, the length of the support set is Indicates rounding up. When the sparsity estimate is close to the true sparsity, the support set length becomes smaller and smaller as the number of stages increases until the final iteration condition is met. The efficiency of the algorithm is guaranteed and the flexibility of the algorithm is improved by selecting the step size.
3. Use dual-threshold iteration termination conditions. Perform residual vector calculation on the MWC system transmission signal based on the initial sparsity estimate. If the currently calculated residual vector value is greater than the first threshold and greater than or equal to the residual vector value calculated in the last iteration, increase the current sparsity estimate according to the first set step size. If it is less than or equal to the first threshold, greater than the second threshold and greater than or equal to the residual vector value calculated in the last iteration, increase the current sparsity estimate according to the second set step size. Recalculate the residual vector again using the increased sparsity estimate until the calculated residual vector value is less than or equal to the second threshold, and the iteration ends. Use the corresponding sparsity estimate at the end of the iteration as the final sparsity estimate, perform signal estimation according to the final sparsity estimate, and output the final estimated signal. The first set step size is greater than the second set step size. This prevents the algorithm from overfitting while making the algorithm more accurate.
4. Based on MWC-based broadband spectrum sensing, the present invention proposes a method for sparse sampling (lower than Nyquist sampling) of signals. By comparing the sampling points and the corresponding signal points after reconstruction, the error can be used to dynamically adjust the number of channels of the system, thereby achieving the purpose of automatically adjusting the number of channels. The process of channel addition is shown in the figure. The error calculation method is based on the mean square error: Where z i is the coefficient sampling point; Represents the point after the corresponding sampling point is reconstructed, and N is the number of samples for comparison. The structural block diagram of the main controller is shown in the figure. The figure shows the process of the system for broadband spectrum sensing: The system can achieve the purpose of adaptively adjusting the number of channels through the above process. Although the number of channels may be increased more than the most suitable number of channels in this process, it has no effect on the result and can find the suitable number of channels more quickly.
