Feature Extraction and Signal Enhancement Based on Unsaturation Piecewise Tri-Stable Stochastic Resonance Driven by Lévy Noise

Zhao, Shuai; Shi, Peiming

doi:https://doi.org/10.1155/2023/7071092

Security and Communication Networks

On this page

Abstract Introduction Discussion Conclusion Data Availability Conflicts of Interest Acknowledgments References Copyright Related Articles

Research Article | Open Access

Volume 2023 | Article ID 7071092 | https://doi.org/10.1155/2023/7071092

Feature Extraction and Signal Enhancement Based on Unsaturation Piecewise Tri-Stable Stochastic Resonance Driven by Lévy Noise

Shuai Zhao^1,2and Peiming Shi¹

Academic Editor: Keke Tang

Received07 Dec 2022

Revised11 Jan 2023

Accepted16 Jan 2023

Published24 Jan 2023

Abstract

As science advances and machines become larger and more sophisticated, it is vital to determine whether there is a mechanical failure without damaging the device. Stochastic resonance (SR), as a widely used method, can effectively extract the periodic signal in the noise and then realize the identification of mechanical faults, which is very important for safety and property protection. Most studies on SR are based on additive white Gaussian noise (AWGN) as the driving source; however, there are many kinds of noise in reality. In this paper, based on the previously proposed piecewise tri-stable SR (PTSR) model, the influence of parameters on the mean of signal-to-noise ratio gain (SNR-GM) of the system driven by Lévy noise is studied. It is also verified by simulation signals and actual signals that PTSR can also achieve feature extraction and signal enhancement with Lévy noise as the driving source, which proves that PTSR can be applied in a wider range of conditions.

1. Introduction

Benzi et al. proposed and explained SR in 1981 [1]. As SR is a phenomenon formed by the interaction of periodic force, random force, and nonlinear system, it can be applied in many disciplines [2–8]. Since SR can transfer the energy of noise to periodic signal, it shows excellent performance in the field of feature extraction and signal enhancement [9–15]. The improvement of potential function, the change of SR order, the adjustment of SR parameters, and the extension of SR model have been studied by many scholars [16]. Yan et al. studied the behavior of particles in a fractional SR system, analyzed the first moment of the system, and gave the steady-state conditions [17]. In reference [18], the white noise-driven neural network SR is analyzed at three levels of dynamics, and the network is composed of two modules. Qiao et al. modified the traditional SR model, successfully improved the output saturation problem, and applied it to feature extraction [19]. Yan et al. added a time delay term to the system on the basis of fractional SR and used fractional Gaussian noise to drive it [20]. In reference [21], quantum SR first appeared in real experiments. Yu et al. added a time delay to the neural network SR and investigated the oscillating transitions of the system [22]. Li and Shi studied the method of extracting nonstationary signals using SR and applied it to bearing signals [23].

As an important part of SR, noise is of great significance to study the drive of SR system based on different noises. Lévy noise, as a colored noise, is also used by some scholars to drive SR system. Zhu et al. also studied a network-based SR system, but Levy noise was used to drive the SR [24]. In reference [25], a new potential function is presented, using Lévy noise as the driving source, and the characteristics of the system are analyzed. Han and Shi used colored noise to drive multistable SR and studied the first-passage time of the system in detail [26].

In order to overcome output saturation, we have previously proposed the PTSR method. In order to make the PTSR model more universal in terms of feature extraction and signal enhancement, a Lévy noise-driven PTSR method has been proposed. Section 2 introduces and describes the PTSR model and how the parameters of the PTSR system relate to the shape of the potential function. In Section 3, the Lévy noise generation method is studied and the PTSR model is driven by it. Meanwhile, the PTSR system can still perform feature extraction and signal enhancement with the help of simulation signals. In Section 4, the relationship between parameters and SNR-GM of PTSR system driven by Lévy noise is studied. In Section 5, engineering signals are used to verify the ability of PTSR system to extract features and enhance signals even when Lévy noise is the driving source. Section 6 discusses the possible research direction in the future. Finally, Section 7 gives the conclusions.

2. The PTSR System and Parameters

In prior research, a PTSR was invented, and its potential function is shown as follows [27]:where , , and are positive number and . Figure 1 displays its potential function. It can be seen that the potential function of PTSR has three stable points and two unsteady points, and .

In reference [28], it is found, the adjustment of system parameters has a great effect on the shape of the potential function, because the change of the shape of the potential function will greatly change the transition rate of the internal particles. The study of reference [28] is based on AWGN, and the evaluation standard is the SNR proposed in reference [29], but in this paper, the noise becomes Lévy noise, and the evaluation standard is different from that so it is still important to study the effect of parameters on the shape of the potential function. Now, we redraw the potential for the PTSR system and keep the other two parameters constant and only change the size of the other one at a time, as shown in Figure 2.

(a)

(b)

(c)

It can be seen from Figure 2(a) that increasing the size of will make the left and right potential wells rise and become deeper, and at the same time, the left and right potential walls of the two unsteady points become steeper, which to some extent will prevent particles from jumping between the three potential wells. It can be observed from Figure 2(b) that if the parameter is increased, the left and right potential wells will also rise, but the depth will be shallower. It can be concluded that the movement of particles between potential wells becomes easier. Finally, it can be seen from Figure 2(c) that with the increase of , the height and depth of the left and right potential well will decrease and become deeper, while the middle potential well will become steeper, which will reduce the motion rate of the particles.

3. PTSR System Driven by Lévy Noise

3.1. The Simulation of Lévy Noise

Lévy noise is also known as stable noise. As a typical non-Gaussian noise, there is usually no explicit expression for its distribution function and probability density function.

3.1.1. Distribution of Lévy Noise

In this paper, in order to simulate the generation of Lévy noise, we use the Chambers-Mallows-Stuck (CMS) [30, 31] algorithm to obtain the random variable of Lévy noise. The characteristic equation and expression of its steady-state distribution are shown in equations (2) and (3):where , ,, and

The random variable follows the following distribution:where and satisfy equations (5) and (6), respectively:

In equation (4), and obey interval distribution and exponential distribution, respectively, and they satisfy the interval and the mean 1.

3.1.2. Solving the Langevin Equation

The Langevin equation of SR system driven by Lévy noise is as follows:where is the output, is the periodic signal, is the Lévy noise, and is the noise amplification factor.

The Langevin equation of SR generally needs to be solved by using fourth-order Runge—Kutta. Driven by Lévy noise, the CMS algorithm and Runge—Kutta are combined, as follows:where , , and are the th sampled values of the output and input signals, respectively, and is the step size. Since, the final particle will have an infinite trajectory due to oscillation, truncation should be carried out [32, 33], when , should be implemented. Setting that, and , on the premise that only is changed, Lévy noise is drawn in Figure 3.

(a)

(b)

(c)

(d)

(e)

(f)

According to Figure 3, when is relatively small, the number of Lévy noise pulses is relatively small, and the absolute value of the peak value is generally relatively high. With increasing , the number of Lévy noise pulses increases gradually, and the mean value of the absolute value of the peak decreases gradually. When approaches , the distribution of levy noise is mainly concentrated around the X-axis.

3.2. Simulation of Feature Extraction and Signal Enhancement

After simulating Lévy noise, it is necessary to verify whether the PTSR system driven by Lévy noise can perform feature extraction and signal enhancement on input signals. A sinusoidal signal of amplitude and characteristic frequency is mixed with Lévy noise. We set the sampling frequency , , ,, and and apply the signal through the PTSR system, The original signal and output signal are shown in Figure 4.

(a)

(b)

(c)

(d)

From Figures 4(a) and 4(b), it can be found that the periodic signal is chaotic after Lévy noise is mixed, and the characteristic frequency is also submerged in other chaotic frequencies, meanwhile, it is also marked that the amplitude is 0.7983. It can be found in Figure 4(c) that the signals become relatively clear and some periodicity is restored, indicating that the PTSR system driven by Lévy noise can filter some chaotic noises from the perspective of filtering. Finally, it is shown in Figure 4(d) that the amplitude of the characteristic frequency is the highest among all frequencies and the value is 0.853. It indicates that the PTSR system successfully carries out feature extraction under Lévy noise and produces signal enhancement.

4. SNR-GM of PTSR System Driven by Lévy Noise

There are many evaluation criteria for SR performance. As an indicator of signal enhancement, SNR gain [34] is widely concerned. At the same time, due to the randomness of noise, SNR-GM is used as the standard to measure the performance of SR in this paper, and , which is as follows:where is given by the following equation:where and are the input and output SNR, which can be obtained by the following formula:where is fast Fourier transform (FFT) of input single, is FFT of output single , and is length of sequence.

4.1. Effect of System Parameter m, p, and q on SNR-GM

Input periodic signals are set with amplitude and characteristic into the PTSR system with from 0 to 5 and set , , and . Also, we set the arithmetic parameter to and . Under the premise that the value of only one parameter is changed each time and the other two parameters remain unchanged, the SNR-GM curve obtained is shown in Figure 5. It could be seen that all curves in the three images show a trend of first rising and then falling, which is also an important basis for judging whether SR occurs in the system. It can be observed from Figures 5(a) and 5(c) that if other parameters are fixed and only the size of or is changed, the SNR-GM of the system will increase with the increase of or . On the contrary, it can be seen from Figure 5(b) that the SNR-GM of the system will decrease with the increase of . The abovementioned conclusion is consistent with the conclusion summarized in Figure 2. Particles are easier to move, so they will generate higher energy and thus have higher SNR-GM. It is worth noting here that the abovementioned analysis was set based on the abovementioned parameters. SR is a complex system, and different results may appear under the setting of other parameters, therefore, it is necessary to continue to study the influence of system parameter , , and on SNR-GM under different values of . Input the same signal from the abovementioned experiment into the signal, and set parameters , , , , and . When the value of ranges from 1.5 to 3.5, the value of is changed, and the SNR-GM curve obtained is shown in Figure 6.

(a)

(b)

(c)

As can be seen from Figure 6, parameter has no effect on the rise of SNR when takes any value, and all curves show a downward or flat trend, which is consistent with the previous judgment on parameter in some cases. Next, the experiment is conducted for parameters and , and the input signal remains unchanged, except for parameter , which is set as 0.8 and 0.5, respectively, other parameters remain unchanged. SNR-GM curves were obtained under the premise that values ranged from 0.5 to 0.8 and values ranged from 0.8 to 1.1, as shown in Figures 7 and 8. It can be seen from Figure 7 that parameter only has a rising or flat effect on SNR-GM, while Figure 8 shows that parameter has a falling or flat effect on SNR-GM, which is consistent with the previous conclusion under certain conditions.

Meanwhile, by observing Figures 6–8, it is found that parameter also has a great influence on SNR-GM, Therefore, we should continue to research the effect of and on SNR-GM.

4.2. Effects of and on SNR-GM

We input the same signal into the PTSR system as before, with the parameters set to , , , and from 0 to 1.5 and draw the SNR curve under different values of as shown in Figure 9.

As shown in Figure 9, when increases gradually, SNR-GM will gradually increase; however, when the value of exceeds 1, SNR-GM will rapidly decrease and further decrease with increasing . Meanwhile, when the value of exceeds 1, the value of , which enables the system to obtain the highest SNR-GM, will obviously change. That is to say, when is less than 1, a mainly affects the level of SNR; when is greater than 1, in addition to the size of SNR-GM, it also has a great impact on the position of . Next, the influence of on SNR-GM is discussed. The same signal is still used, and parameters , , and are set at the same time. When is from 0 to 6, SNR-GM curves with different values of are drawn, as displayed in Figure 10. It could be seen from Figure 10 that, like the curves in other figures, all curves show a trend of first rising and then falling, which still proves that SR phenomenon occurs in the PTSR system driven by Lévy noise. At the same time, increasing the absolute value of , increases the SNR-GM of the system. If SNR-GM is greater when is negative then when is positive, is the same absolute value.

5. The Experimental Result

The experimental platform shown in Figure 11 is used to extract signals and verify the ability of enhancement for the PTSR system driven by Lévy noise. This equipment is a medium-speed outer ring bearing failure platform of a company and its and , and Figure 12 shows the original signal. Figure 12 shows that both the time domain and frequency domain of the original signal are chaotic, and the amplitude at the characteristic frequency is equal to 0.2526. The envelope spectrum of the original signal is given in Figure 13. It can be observed from Figure 13 that the characteristic frequency of the fault can be indicated, but the noise interference is very high. The original signal is then mixed with Lévy noise as shown in Figure 14, and the mixed signal passes through the PTSR system. First, scale transformation [35] is carried out and is set; then, the secondary sampling frequency and . For optimal parameters, particle swarm optimization (PSO) is used to search to maximize the output SNR-GM, and the output singles are shown in Figure 15.

(a)

(b)

(a)

(b)

(a)

(b)

It can be seen from Figure 13(a) that the time-domain signal recovers part of its periodicity, indicating that the PTSR system driven by Lévy noise can successfully remove noise from the perspective of filtering. Meanwhile, in Figure 13(b), the amplitude of feature frequency is the highest, indicating that feature extraction has been successfully carried out by the system. Finally, the amplitude at the characteristic frequency has increased from 0.2526 to 1.826, indicating successful signal enhancement in the Lévy noise-driven PTSR system.

In order to further verify the adaptation degree of Lévy noise in PTSR system, the abovementioned signals are processed by other SR systems. Standard tri-stable SR [36] and classical bistable SR are selected in this paper, and their potentials are and , respectively. It should be noted that due to the randomness of noise, signals still need to pass through the PTSR system in this experiment. All parameters are consistent with the previous experimental settings. The system parameters of the three SR systems are automatically searched through PSO, and the frequency-domain diagrams of the three output signals are obtained, as shown in Figure 16.

(a)

(b)

(c)

It can be seen from Figure 16 that all three SR systems can use Lévy noise to achieve feature extraction and signal enhancement, but the enhancement results are, respectively, from the original 0.2526 to 0.8612, 1.272, and 1.736. The experimental results show that PTSR system has better adaptation to Lévy noise and stronger ability to utilize Lévy noise.

6. Discussion

Based on the previously proposed PTSR, Lévy noise is used as the driving source to verify that the PTSR method can still perform feature extraction and signal enhancement. The relationship between system parameters and SNR-GM is also described. In the follow-up study, we can continue to study whether PTSR still has the abovementioned ability under the driving of other noises. However, it is worth investigating whether there are other more effective criteria besides SNR-GM.

7. Conclusion

Through the study of the PTSR system driven by Lévy noise, the following conclusions are obtained:(1)The selection of system parameters will greatly affect the performance of SR system, so the relationship between parameters and SNR-GM is important. When Lévy noise drives, within a certain range, the increase of will increase SNR-GM, while and have opposite effects, and these effects correspond to the change of potential function.(2) and also have a strong influence on the SNR-GM of the system and will change with their values, plus or minus.(3)Through the verification of simulation signals and actual engineering signals, it is found that the PTSR system driven by Lévy noise can be used for feature extraction and signal enhancement.(4)The fault type cannot be predicted in advance, so the capability of feature extraction and signal enhancement is an important criterion to evaluate the performance of SR. When Lévy noise is the driving source, the PTSR system has better adaptation and signal enhancement ability compared to STSR and SR systems.

Data Availability

Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data are not available.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

The study was funded by the National Natural Science Foundation of China (Grant no. 61973262) and the Central Government Guides Local Science and Technology Development Fund Projects (Grant no. 216Z2102G).

References

R. Benzi, A. Sutera, and A. Vulpiani, “The mechanism of stochastic resonancehe mechanism of stochastic resonance,” Journal of Physics A: Mathematical and General, vol. 14, no. 11, pp. 453–457, 1981.
View at: Publisher Site | Google Scholar
M. Dykman and P. V. E. Mcclintock, “What can stochastic resonance do?” Nature, vol. 391, no. 6665, 1998.
View at: Publisher Site | Google Scholar
R. Dutta, A. Das, N. G. Stocks, and DG. Morgan, “Stochastic resonance-based electronic nose: Stochastic resonance-based electronic nose: A novel way to classify bacteria novel way to classify bacteria,” Sensors and Actuators B: Chemical, vol. 115, no. 1, pp. 17–27, 2006.
View at: Publisher Site | Google Scholar
F. Marino, M. Giudici, S. Barland, and S Balle, “Experimental evidence of stochastic resonance in an excitable optical system,” Physical Review Letters, vol. 88, Article ID 040601, 2002.
View at: Publisher Site | Google Scholar
P. S. Burada, G. Schmid, D. Reguera, M. H. Vainstein, J. M. Rubi, and P. Hänggi, “Entropic stochastic resonance,” Physical Review Letters, vol. 101, no. 13, Article ID 130602, 2008.
View at: Publisher Site | Google Scholar
S. Bai, F. Duan, F. Chapeau-Blondeau, and D. Abbott, “Generalization of stochastic-resonance-based threshold networks with Tikhonov regularization,” Physical Review E, vol. 106, no. 1, 2022.
View at: Publisher Site | Google Scholar
B. B. Dan, M. H. Li, T. Tang, X. P. Qi, Z. J. Zhu, and Y. M. Ouyang, “Spatial-temporal stochastic resonance model for dim-small target detection,” IEEE Geoscience and Remote Sensing Letters, vol. 19, pp. 1–5, 2022.
View at: Publisher Site | Google Scholar
W. Moon, L. T. Giorgini, and J. S. Wettlaufer, “Analytical solution of stochastic resonance in the nonadiabatic regime,” Physical Review E, vol. 104, no. 4, Article ID 044130, 2021.
View at: Publisher Site | Google Scholar
S. L. Lu, Q. B. He, and J. Wang, “A review of stochastic resonance in rotating machine fault detection,” Mechanical Systems and Signal Processing, vol. 116, pp. 230–260, 2019.
View at: Publisher Site | Google Scholar
B. Wang, F. Y. Gao, M. K. Gupta, G. Krolczyk, P. Gardoni, and Z. X. Li, “Risk analysis of a flywheel battery gearbox based on optimized stochastic resonance model,” Journal of Energy Storage, vol. 52, Article ID 104926, 2022.
View at: Publisher Site | Google Scholar
J. H. Yang, C. Yang, X. Z. Zhuang, H. G. Liu, and Z. L. Wang, “Unknown bearing fault diagnosis under time-varying speed conditions and strong noise background,” Nonlinear Dynamics, vol. 107, no. 3, pp. 2177–2193, 2022.
View at: Publisher Site | Google Scholar
S. B. Jiao, X. X. Qiao, S. Lei, and W. Jiang, “A novel parameter-induced adaptive stochastic resonance system based on composite multi-stable potential model,” Chinese Journal of Physics, vol. 59, pp. 138–152, 2019.
View at: Publisher Site | Google Scholar
P. M. Shi, D. Z. Yuan, D. Y. Han, Y. Zhang, and R. R. Fu, “Stochastic resonance in a time-delayed feedback tristable system and its application in fault diagnosis,” Journal of Sound and Vibration, vol. 424, pp. 1–14, 2018.
View at: Publisher Site | Google Scholar
S. Wang, P. J. Niu, Y. F. Guo, and F. Z. Wang, “Early diagnosis of bearing faults using decomposition and reconstruction stochastic resonance system,” Measurement, vol. 158, Article ID 107709, 2020.
View at: Publisher Site | Google Scholar
Z. J. Qiao, J. Liu, and X. Ma, “Double Double stochastic resonance induced by varying potential-well depth and widthtochastic resonance induced by varying potential-well depth and width,” Journal of the Franklin Institute, vol. 358, no. 3, pp. 2194–2211, 2021.
View at: Publisher Site | Google Scholar
Z. J. Qiao, Y. G. Lei, and N. Li, “Applications of stochastic resonance to machinery fault detection: Applications of stochastic resonance to machinery fault detection: A review and tutorial review and tutorial,” Mechanical Systems and Signal Processing, vol. 122, pp. 502–536, 2019.
View at: Publisher Site | Google Scholar
Z. Yan, J. L. G. Guirao, and T. Saeed, “Analysis of stochastic resonance in coupled oscillator with fractional damping disturbed by polynomial dichotomous noise,” Nonlinear Dynamics, vol. 110, no. 2, pp. 1233–1251, 2022.
View at: Publisher Site | Google Scholar
X. L. Yang, N. Li, and Z. K. Sun, “Extended analysis of stochastic resonance in a modular neuronal network at different scales,” Nonlinear Dynamics, vol. 98, no. 2, pp. 1029–1039, 2019.
View at: Publisher Site | Google Scholar
Z. J. Qiao, Y. G. Lei, J. Lin, and F. Jia, “An adaptive unsaturated bistable stochastic resonance method and its application in mechanical fault diagnosis,” Mechanical Systems and Signal Processing, vol. 84, pp. 731–746, 2017.
View at: Publisher Site | Google Scholar
Z. Yan, J. L. G. Guirao, T. Saeed, H. T. Chen, and X. B. Liu, “Different stochastic resonances induced by multiplicative polynomial trichotomous noise in a fractional order oscillator with time delay and fractional Different Stochastic Resonances Induced by Multiplicative Polynomial Trichotomous Noise in a Fractional Order Oscillator with Time Delay and Fractional Gaussian Noiseaussian noise,” Fractal and Fractional, vol. 6, no. 4, 2022.
View at: Publisher Site | Google Scholar
T. Wagner, P. Talkner, J. C. Bayer, and E. P. Rugeramigabo, “Quantum stochastic resonance in an a.c.-driven single-electron quantum dot,” Nature Physics, vol. 15, no. 4, pp. 330–334, 2019.
View at: Publisher Site | Google Scholar
H. Yu, X. Guo, and J. Wang, “Multiple stochastic resonances and oscillation transitions in cortical networks with time delay,” IEEE Transactions on Fuzzy Systems, vol. 28, no. 1, pp. 39–46, 2020.
View at: Publisher Site | Google Scholar
Z. X. Li and B. Q. Shi, “A piecewise nonlinear stochastic resonance method and its application to incipient fault diagnosis of machinery,” Chinese Journal of Physics, vol. 59, pp. 126–137, 2019.
View at: Publisher Site | Google Scholar
Q. H. Zhu, J. W. Shen, and J. C. Ji, “Internal signal stochastic resonance of a two-component gene regulatory network under Lévy noise,” Nonlinear Dynamics, vol. 100, no. 1, pp. 863–876, 2020.
View at: Publisher Site | Google Scholar
L. F. He, X. C. Zhou, and T. Q. Zhang, “Stochastic resonance characteristic analysis of new potential function under Levy noise and bearing fault detection,” Chinese Journal of Physics, vol. 56, no. 2, pp. 538–560, 2018.
View at: Publisher Site | Google Scholar
D. Y. Han and P. M. Shi, “Study on the mean first-passage time and stochastic resonance of a multi-stable system with colored correlated noises,” Chinese Journal of Physics, vol. 69, pp. 98–107, 2021.
View at: Publisher Site | Google Scholar
S. Zhao, P. M. Shi, and D. Y. Han, “A novel mechanical fault signal feature extraction method based on unsaturated piecewise tri-stable stochastic resonance,” Measurement, vol. 168, Article ID 108374, 2021.
View at: Publisher Site | Google Scholar
S. Zhao, P. M. Shi, D. Y. Han, and R. R. Fu, “A novel piecewise tri-stable stochastic resonance system with time-delayed feedback and its application,” Chinese Journal of Physics, vol. 73, pp. 288–303, 2021.
View at: Publisher Site | Google Scholar
J. M. Li, X. F. Chen, Z. H. Du, Z. Fang, and Z. He, “A new noise-controlled second-order enhanced stochastic resonance method with its application in wind turbine drivetrain fault diagnosis,” Renewable Energy, vol. 60, pp. 7–19, 2013.
View at: Publisher Site | Google Scholar
J. M. Chambers, C. L. allows, and B. W. Stuck, “Correction to: “a method for simulating stable random variables”,” Journal of the American Statistical Association, vol. 71, no. 354, pp. 340–344, 1976.
View at: Publisher Site | Google Scholar
A. Weron and R. Weron, “Computer simulation of Lévy α-stable variables and processes,” Lecture Notes in Physics, vol. 457, Springer, Berlin, Germany, 1995.
View at: Publisher Site | Google Scholar
R. Weron and A. Weron, “On the Chambers-Mallows-Stuck method for simulating skewed stable random variables,” Statistics & Probability Letters, vol. 28, no. 2, pp. 165–171, 1996.
View at: Publisher Site | Google Scholar
C. Gong and Z. L. Wang, MATLAB Language Commonly Used Algorithm for Assembly, Publishing House of Electronics Industry, Beijing, China, 2008.
P. Wan, Y. J. Zhan, X. C. Li, and Y. H. Wang, “Numerical research of signal-to-noise ratio gain on a monostable stochastic resonance,” Acta Physica Sinica, vol. 60, Article ID 040502, 2011.
View at: Google Scholar
P. M. Shi, S. J. An, P. Li, and D. Y. Han, “Signal feature extraction based on cascaded multi-stable stochastic resonance denoising and EMD method,” Measurement, vol. 90, pp. 318–328, 2016.
View at: Publisher Site | Google Scholar
Z. H. Lai and Y. G. Leng, “Dynamic response and stochastic resonance of a tri-stable system,” Acta Physica Sinica, vol. 64, 2015.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2023 Shuai Zhao and Peiming Shi. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies