Error Signal Differential Term Feedback Enhanced Variable Step Size FxLMS Algorithm for Piezoelectric Active Vibration Control

Li, Weiguang; Wang, Wei; Li, Bin; Yang, Zhichun

doi:https://doi.org/10.1155/2020/8832467

Shock and Vibration

On this page

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

Special Issue

Condition Monitoring and Fault Diagnosis of Electrical Drive Systems

View this Special Issue

Research Article | Open Access

Volume 2020 | Article ID 8832467 | https://doi.org/10.1155/2020/8832467

Error Signal Differential Term Feedback Enhanced Variable Step Size FxLMS Algorithm for Piezoelectric Active Vibration Control

Weiguang Li,¹Wei Wang,¹Bin Li,¹and Zhichun Yang¹

Academic Editor: Muhammad Irfan

Received30 Jul 2020

Revised12 Sept 2020

Accepted19 Oct 2020

Published06 Nov 2020

Abstract

FxLMS (Filtered-x Least Mean Square) algorithm is widely used in the field of AVC (active vibration control) for its good convergence and strong adaptability. However, the convergence rate and steady-state error are mutually restricted for the fixed step FxLMS algorithm. Increasing step size to accelerate the convergence rate will result in larger steady-state error and even cause control divergence. In this paper, a new DVSFxLMS (error signal Differential term feedback Variable Step size FxLMS) algorithm is proposed by establishing nonlinear function between and error signal, while using differential term of the error signal as the feedback control function. Subsequently, a DVSFxLMS controller is designed to carry out the AVC simulation and experiments on cantilever beam with PSA (piezoelectric stack actuator). Simulation and experimental results show that the proposed DVSFxLMS algorithm has faster convergence rate and smaller steady-state error than the traditional FxLMS algorithm, which also has strong antinoise ability and adaptive control ability to quickly track the variable external disturbance.

1. Introduction

Adaptive filter technology has wide range applications in the field of digital signal processing. Compared with a conventional filter, an adaptive filter can adjust characteristics of the filter online according to adaptive filter technology, and obtain the best performance filter by finding the appropriate weight coefficients [1]. The adaptive filter is often divided into two separate parts: one part is a digital filter and the other part is an adaptive algorithm. The digital filter adjusts weight coefficients through the adaptive algorithm to improve its signal processing performance. The schematic diagram of adaptive filtering is shown in Figure 1.

When applying adaptive filtering technology to the field of AVC, a FIR (finite impulse response) filter is usually used as a feedforward controller, called the adaptive controller. The weight coefficients of the adaptive controller are adjusted according to the adaptive algorithm that usually is LMS algorithm, so that the mean square value of the vibration response measured by sensor converges to the minimum direction. The developed FxLMS vibration control algorithm has advantages of good convergence and strong adaptability, and does not depend on the accurate model of controlled structure, which has become one of the hotspots of AVC algorithm research [2–5]. However, in the FxLMS control algorithm, the convergence rate and steady-state error of LMS algorithm are greatly affected by the step size . Generally, the larger the step size, the faster the convergence rate, but the steady-state error will also increase; reducing step size can reduce the steady-state error, thereby improving the convergence accuracy, but a smaller step size will significantly reduce the convergence and tracking rate of LMS algorithm [6]. Therefore, the traditional fixed step LMS algorithm contradicts the adjustment of step size in terms of convergence rate, tracking rate, and convergence accuracy. In order to solve this contradiction, it is necessary to adjust the step size of LMS algorithm in real time during control process, which is called the variable step size LMS algorithm [7–9]. Gitlin et al. [10] proposed to reduce step size with the increase in number of the algorithm iterations, so as to achieve the purpose of variable step size, but this adjustment rule is only applicable to the time-invariant systems. Qin and Ouyang [11] proposed a variable step size SVSLMS (Sigmoid function Variable Step size LMS) algorithm based on the sigmoid function. By establishing the nonlinear relationship between the step size and error signal, which effectively compensated for the shortcomings in [10]. However, with continuous convergence of the SVSLMS algorithm, the step size changes more and more drastically, which greatly affects the steady-state performance of the SVSLMS algorithm. Gao and Xie [12] proposed a simpler variable step size algorithm , which overcomes the shortcomings of the sigmoid function in steady-state step adjustment process and also achieves faster convergence rate and smaller steady-state error. However, the current step size of this algorithm is only related to the current error signal, ignoring the effect of previous iteration error signal on the current step size, so it has a certain negative impact on the steady-state error and the convergence rate [13–15]. At the same time, under low SNR (signal noise ratio) environment, the convergence effect of this algorithm is not very ideal, which greatly restricts its application range [16]. In addition, most of the existing variable step size LMS algorithms are mainly used in the fields of system identification [17, 18] and active noise control [19, 20], which are different from their application in the active control of structural vibration [21, 22]. In the field of AVC, the system may be interfered by strong external noise, which makes the algorithm generate larger step size and output too large control signals, causing system instability or even damage. Therefore, when the variable step size FxLMS algorithm is applied to the active control of structural vibration, the robustness to noise interference needs to be considered [23].

In this study, a new variable step size FxLMS algorithm is developed which is called DVSFxLMS algorithm, by establishing nonlinear function between the step size and the error signal, while the differential term of the error signal is adopted as feedback control function in original algorithm, so that the current step size of the proposed algorithm is related to error signal rate. At the same time, the correlation value of the error signal is used instead of the square of the error signal to adjust step size. Subsequently, the DVSFxLMS controller is designed to actively control the vibration response of cantilever beam with PSA. The AVC simulation and experiments of piezoelectric cantilever beam under harmonic excitation and harmonic excitation with superimposed noise are carried out, and the adaptability of the DVSFxLMS algorithm is also studied. Simulation and experimental results show that compared with the traditional fixed step FxLMS algorithm, the proposed DVSFxLMS algorithm has faster convergence rate, smaller steady-state error, which also has strong antinoise ability and adaptive control ability to quickly track the variable external disturbances. The block diagram of the proposed research is shown in Figure 2.

2. Design of DVSFxLMS Controller

The LMS algorithm proposed by Widrow and Hoff [24] in 1960 is widely used in the fields of system identification [25], signal processing [26], and adaptive control [27] because of its advantages such as small calculation, easy implementation, and great stability. The LMS algorithm based on the steepest descent method can be summarized as the following iterative process:where is the filter input signal vector of length L; is the N weight coefficients of the filter at time n; is the step size, which determines the steady-state performance and convergence rate of the LMS algorithm. The convergence condition of the LMS algorithm iswhere is the maximum eigenvalue of the autocorrelation matrix of .

The variable step size algorithm can solve the contradiction between step size, convergence rate, and convergence accuracy in the traditional fixed step LMS algorithm. The principle of the variable step size LMS algorithm is to use a larger step size to obtain a faster convergence rate at the initial stage of the convergence or the system changes suddenly; when the algorithm converges to steady state, a smaller step size is used to reduce the steady-state error. At the same time, the calculation amount of the algorithm should be as small as possible, and the parameters that need to be adjusted should be as few as possible to enhance the practicality of the algorithm. Therefore, based on the algorithm in [13], and combining the advantages of the algorithm in [16], while a feedback control function inversely proportional to the differential term of error signal is introduced, a new variable step size LMS algorithm is proposed in this paper:

The proposed algorithm contains two parameters p and β. Among them, the parameter , which controls the shape of the variable step size function; the parameter β > 0, which controls the value range of the variable step size function. From equation (2), ; therefore, . Compared with the original algorithm, equation (3) considers the difference term of the error signal, which reflects the correlation between the step size and the error signal rate. At the same time, the correlation value of the error signal is used instead of the square of the error signal to adjust step size, which can further improve the antinoise ability of the variable step size algorithm [16].

The DVSFxLMS vibration control algorithm thus constructed takes the vibration response of the controlled structure caused by external disturbance as starting point, and requires the control signal to drive the actuator to generate control force or moment on the controlled structure, so that the control response will cancel out the response caused by external disturbances at observation points, so as to achieve the purpose of eliminating or reducing the vibration level of the controlled structure. Figure 3 shows the structural diagram of the DVSFxLMS controller.

In Figure 3, is the structural vibration response at time n when no control signal is applied. The channel from actuator to sensor is called the control channel, and is the model obtained from the offline identification of the control channel. is the structural vibration response caused by the control signal through the control channel model . is the structural vibration response error signal.

In summary, the iterative process of the DVSFxLMS vibration control algorithm is obtained as follows:

3. AVC Simulation of DVSFxLMS Algorithm

3.1. Piezoelectric Cantilever Beam

In the current simulation, an aluminum alloy cantilever beam is controlled structure, and its dimension is 900 mm × 20 mm × 5 mm. As shown in Figure 4, based on the maximum modal strain energy criterion, the PSA is installed at the root of the cantilever beam to establish vibration control simulation system, in which the observation point A is at the tip of the cantilever beam.

The dimension of PSA is 170 mm × 20 mm × 37 mm, and its total weight is 190 g. The piezoelectric stack used therein is PI™ PICMA® P-840.60, and the parameters of this piezoelectric stack are shown in Table 1 [28]. The PSA is glued to controlled structure through epoxy resin to realize the axial actuation of the piezoelectric stack into a pair of actuating bending moments on the controlled structure.

3.2. Offline Identification of Control Channel

Figure 5 shows the schematic diagram of the control channel offline identification. Applying the excitation signal to the unknown control channel, then generating an output response , and applying the same excitation signal to the adaptive filter, the filter output is . Then and are subtracted to get the identification error signal . The LMS algorithm adjusts the weight coefficients of the filter according to the error signal, and finally makes the output of the filter close to the output response of the control channel. At this time, the characteristic of the adaptive filter can be used as an estimate of the control channel.

After the input and output data are obtained, the least square method is used to identify the FIR model for control channel. Whether the results of the model identification meet the needs can be qualitatively evaluated from following two requirements: firstly, comparing the frequency response function of the experimental model and the identification model, which requires the amplitude and phase as consistent as possible; secondly, the iterative curves of weight coefficient are required to smoothly converge during the identification process, and the identification error gradually decreases. If the above requirements are not met, it is necessary to readjust the step size or the order of the FIR model in LMS identification algorithm.

In the control channel offline identification experiment, 28–32 Hz narrow-band random signal is input to the control channel, the output signal of control channel is measured by the acceleration sensor A, and the step size is set to 5E – 9; then, a 1500 order FIR filter model is obtained by the LMS algorithm identification. The identification results are shown in Figure 6. It can be seen from the figure that the frequency response function of the control channel model obtained by offline identification is in good agreement with the experimental curve, and the weight coefficients of the adaptive filter converge smoothly, while the MSE (mean square error) of the identification model gradually approaches zero, which indicates that the FIR filter model identified in the control frequency band can truly reflect the dynamic characteristics of the control channel.

(a)

(b)

(c)

3.3. Simulation of DVSFxLMS Control Algorithm

In order to verify the effectiveness of the proposed DVSFxLMS control algorithm (DVSFxLMS in legends), the traditional fixed step FxLMS control algorithm (fixed FxLMS in legends) is used as a comparison group to compare and study its vibration suppression effect in the active control of piezoelectric cantilever beam harmonic vibration. The control channel model is obtained by offline identification process in Section 3.2. The simulation results are shown in Figure 7. It can be seen from Figure 7 that the DVSFxLMS control algorithm has better vibration suppression performance, compared with the FxLMS control algorithm, it has faster convergence rate and smaller steady-state error.

(a)

(b)

4. AVC Experiment Verification of DVSFxLMS Algorithm

The AVC experimental diagram of the piezoelectric cantilever beam is shown in Figure 8. The external excitation signal is generated by the Quanser real-time system, and the external excitation signal passes through the output board and amplified by the power amplifier (MB YE5872A), and then input into the electromagnetic exciter (MB Dynamics) to excite the cantilever beam. Structural vibration response is collected by the acceleration sensor (PCB 333B30, sensitivity: 100 mV/g), enters the DVSFxLMS controller through the acceleration signal conditioner (PCB 482C) and Quanser input board, then the control voltage calculated by the DVSFxLMS controller passes through Quanser output board, which is amplified by the power amplifier (PA-V-M4) and drive the PSA to control the cantilever beam, thereby achieving active control of the cantilever beam vibration response. The setup of AVC experiment using PSA is shown in Figure 9.

4.1. AVC Experiment under Harmonic Excitation

The external excitation is selected near the natural frequency of the second-order bending mode of the cantilever beam, which is 30 Hz, and the control channel model is obtained by offline identification process described in Section 3.2. After the vibration response of the cantilever beam reaches steady state, the controller is turned on at 10 s, and the experiment lasts for 50 s. Among them, the adaptive filter order of the traditional fixed step FxLMS controller is set to 32, and the step size in the adaptive algorithm is adjusted to 2E − 6. Meanwhile, the adaptive filter order of the DVSFxLMS controller is set to 32, and the step size parameters are adjusted to β = 5E − 6 and . The superiority of the DVSFxLMS control algorithm proposed over the traditional fixed step FxLMS control algorithm is compared and studied.

The experimental results are shown in Figure 10. From Figures 10(a) and 10(c), it can be seen that the convergence time of the fixed step FxLMS controller is 19.5 s. After the structural vibration reaches steady state, the peak acceleration at the tip of the cantilever beam is decreased by 93.7%. At the same time, the convergence time of the DVSFxLMS controller designed is 8.1 s, and the control effect reaches 95.1%, which is better than fixed step FxLMS controller. As shown in Figure 10(b), at the initial period, the output voltage of the DVSFxLMS controller is higher than that of the fixed step FxLMS controller, and then quickly reaches the optimal value to ensure high level of the vibration suppression performance. Figure 10(d) shows the iterative curves of the adaptive filter weight coefficients of the DVSFxLMS controller, and the filter weight coefficients converge quickly and smoothly. In summary, compared with the fixed step FxLMS control algorithm, the DVSFxLMS control algorithm proposed has the characteristics of fast convergence rate and good steady-state vibration suppression performance, which effectively solve the constraints of the convergence rate, steady-state error and step size of the fixed step FxLMS control algorithm.

(a)

(b)

(c)

(d)

4.2. AVC Experiment under Harmonic Excitation with Superimposed Noise

In order to verify the ability of the DVSFxLMS control algorithm proposed to resist noise interference, this section conducts experimental research on AVC under the harmonic excitation with superimposed noise.

Excitation signal consists of 30 Hz sine signal superimposed 28–32 Hz zero-mean Gaussian white noise, and the variance of the superimposed noise σ² is set to 1, 0.25, and 0.01, respectively. The controller parameter settings are the same as described in Section 4.1. Turn on the controller at 10 s, and the experiment lasts for 50 s. The experimental results of AVC under the harmonic excitation with superimposed noise are shown in Figures 11 and 12.

(a)

(b)

(c)

(a)

(b)

(c)

It can be seen from Figure 11 that under the harmonic excitation with superimposed noise, the DVSFxLMS controller can all converge to steady state in about 8 s. During the steady state, the DVSFxLMS controller can also adaptively adjust the filter parameters according to the changes of the error signal to suppress vibration response of the controlled structure, which shows the ability to resist noise interference.

The acceleration response PSD (power spectral density) contrast at point A in the open/closed loop state of the control system is shown in Figure 12. Under the harmonic excitation of superimposed noise signals with variances of 1, 0.25, and 0.01, the DVSFxLMS controller can effectively reduce vibration level of the controlled structure. The lower the variance of the superimposed noise, the better effect of the active structural vibration control. Finally, the peak value of the PSD spectrum is decreased by 21.5 dB, 24.9 dB, and 30.5 dB, and the RMS (root mean square) of the acceleration response is decreased by 89.5%, 92.2%, and 95.5%, respectively, which verifies that the proposed DVSFxLMS control algorithm has strong antinoise ability.

4.3. AVC Experiment under Variable External Excitation

In order to investigate the adaptability of the DVSFxLMS control algorithm to variable external disturbance, this section conducts an adaptive control experiment on the changes of the excitation signal amplitude, phase, and frequency. The control system starts to use 30 Hz sine signal as the excitation, and the controller is turned on at 10 s. After vibration response of the structure is attenuated and stabilized, the excitation signal amplitude increases to 1.5 times of the original amplitude at 40 s, the phase of excitation signal changes at 70 s, and the frequency of excitation signal increases by 5% at 100 s; that is, it becomes 31.5 Hz. The experiment time lasts for 150 s, and the controller parameters are consistent with Section 4.1.

Figure 13(a) shows the acceleration response at point A, and Figure 13(b) shows the output voltage of the DVSFxLMS controller. It can be seen that after the controller is turned on, the structural vibration response quickly decays to steady state. When the excitation signal amplitude, phase, or frequency changes, the structural vibration response rapidly increases. Then, the DVSFxLMS controller adjusts the output voltage according to the changes of the structural vibration response, and drives the PSA to generate actuating bending moment, so that the structural vibration response quickly decays to steady state again. The adjustment process takes 9.3 s, 5.1 s, and 13.4 s, respectively, which realizes the adaptive control of the structural vibration response when the external excitation changes. It can be seen from Figure 13(c) that when the excitation signal parameters change, the DVSFxLMS controller can quickly adjusts the weight coefficients of the adaptive filter according to the DVSFxLMS algorithm to reach steady state, which ensures good vibration suppression performance. The experimental results of adaptive vibration control show that the DVSFxLMS control algorithm proposed has fast tracking rate and strong adaptive control ability to variable external disturbance.

(a)

(b)

(c)

5. Conclusion

In the present study, a new variable step size FxLMS algorithm (DVSFxLMS algorithm) is proposed by establishing nonlinear function between the step size and the error signal, while using the difference term of error signal into the original algorithm as a feedback control function. At the same time, the correlation value of the error signal is used instead of the square of the error signal to adjust step size. Subsequently, DVSFxLMS controller is designed, and PSA is used to conduct AVC simulation and experiments on a cantilever beam. Results show that the proposed DVSFxLMS control algorithm can effectively suppress the vibration response of the cantilever beam. The control effect of single frequency harmonic excitation is 95.1%, and the convergence time is 8.1 s, while that of the fixed step FxLMS controller is 19.5 s. Compared with the traditional fixed step FxLMS control algorithm, the convergence rate is faster, and the steady-state vibration suppression performance is better. The DVSFxLMS controller can still achieve good control effects under the harmonic excitation with superimposed noise. The vibration response of the cantilever beam converges to steady state in about 8 s. Under the harmonic excitation of superimposed noise signals with variances of 1, 0.25, and 0.01, the peak value of the PSD spectrum of the acceleration response at observation point is decreased by 21.5 dB, 24.9 dB, and 30.5 dB, and the RMS is reduced by 89.5%, 92.2%, and 95.5%, respectively, which shows that the DVSFxLMS control algorithm has strong antinoise ability. When the amplitude, phase, or frequency of the excitation signal changes, the DVSFxLMS controller can adjust the control voltage accordingly, quickly suppress the structural vibration response in about 8 s, and achieve good control effect, which indicates that the proposed DVSFxLMS control algorithm has strong adaptive control ability to quickly track the variable external disturbance.

A new DVSFxLMS control algorithm is proposed in this paper, and the DVSFxLMS controller is designed to effectively suppress the vibration response of the piezoelectric cantilever beam. Considering that the physical characteristics and system characteristics of the controlled structure used in this paper are relatively stable, and therefore, a reliable identification result is obtained by using the offline identification strategy of control channel based on the LMS algorithm. In order to further improve the practicability and applicability of the algorithm proposed, and make it also suitable for time-varying structures, an adaptive controller with online identification function of control channel can be studied based on the variable step size LMS algorithm [3, 29]. In addition, when faced with large and complex controlled structures, the single-channel controller has certain limitations. The multichannel FxLMS controller can be designed based on the variable step size LMS algorithm [30, 31], while considering the control channel coupling phenomenon to further optimize the structure of variable step size FxLMS controller to be applied to actual engineering structures.

Data Availability

The MATLAB .mat data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (grant no. 11502208).

References

P. S. R. Diniz, “Fundamentals of adaptive filtering,” Springer International, vol. 694, pp. 13–78, 2013.
View at: Google Scholar
S. Elliott, I. Stothers, and P. Nelson, “A multiple error LMS algorithm and its application to the active control of sound and vibration,” Institute of Electrical and Electronics Engineers Transactions on Acoustics, Speech, and Signal Processing, vol. 35, no. 10, pp. 1423–1434, 1987.
View at: Publisher Site | Google Scholar
W. Niu, C. Zou, B. Li, and W. Wang, “Adaptive vibration suppression of time-varying structures with enhanced FxLMS algorithm,” Mechanical Systems and Signal Processing, vol. 118, pp. 93–107, 2019.
View at: Publisher Site | Google Scholar
L. H. Yang, S. Y. Liu, H. P. Zhang et al., “Hybrid filtered-x adaptive vibration control with internal feedback and online identification,” Shock and Vibration, vol. 2018, Article ID 9010567, 15 pages, 2018.
View at: Publisher Site | Google Scholar
J. C. O. Marra, E. M. O. Lopes, J. J. d. Espíndola, and W. A. Gontijo, “Hybrid vibration control under broadband excitation and variable temperature using viscoelastic neutralizer and adaptive feedforward approach,” Shock and Vibration, vol. 2016, Article ID 5375309, 12 pages, 2016.
View at: Publisher Site | Google Scholar
T. Aboulnasr and K. Mayyas, “A robust variable step-size LMS-type algorithm: analysis and simulations,” Institute of Electrical and Electronics Engineers Transactions on Signal Processing, vol. 45, no. 3, pp. 631–639, 1997.
View at: Publisher Site | Google Scholar
B. Huang, Y. Xiao, J. Sun, and G. Wei, “A variable step-size FXLMS algorithm for narrowband active noise control,” Institute of Electrical and Electronics Engineers Transactions on Audio, Speech, and Language Processing, vol. 21, no. 2, pp. 301–312, 2013.
View at: Publisher Site | Google Scholar
D. Bismor, K. Czyz, and Z. Ogonowski, “Review and comparison of variable step-size LMS algorithms,” The International Journal of Acoustics and Vibration, vol. 21, pp. 24–39, 2016.
View at: Publisher Site | Google Scholar
B. Jalal, X. Yang, Q. Liu, T. Long, and T. K. Sarkar, “Fast and robust variable-step-size LMS algorithm for adaptive beamforming,” Institute of Electrical and Electronics Engineers Antennas and Wireless Propagation Letters, vol. 19, no. 7, pp. 1206–1210, 2020.
View at: Publisher Site | Google Scholar
R.D Gitlin and S.D Weinstein, “On the design of gradient algorithms for digitally implemented adaptive filters,” Institute of Electrical and Electronics Engineers Transactions on Circuit Theory, vol. 20, no. 2, pp. 125–136, 1973.
View at: Publisher Site | Google Scholar
J. Qin and J. Ouyang, “A new variable step size LMS adaptive filtering algorithm,” Journal of Data Acquisition and Processing, vol. 12, no. 3, pp. 171–194, 1997.
View at: Google Scholar
Y. Gao and S. Xie, “A variable step size LMS adaptive filtering algorithm and its analysis,” Journal of Electronics, vol. 28, no. 8, pp. 1094–1097, 2001.
View at: Google Scholar
G. Han, F. Wang, H. Zhao et al., “A variable step size LMS adaptive filtering algorithm and its application,” Journal of North University of China, vol. 38, p. 2, 2017.
View at: Google Scholar
W. Ao, W. Q. Xiang, Y. P. Zhang et al., “A new variable step size LMS adaptive filtering algorithm,” in Proceedings of the International Conference on Computer Science and Electronics Engineering, pp. 265–268, Institute of Electrical and Electronics Engineers, Kuala Lumpur, Malaysia, April 2012.
View at: Google Scholar
Y. Li and X. N. Wang, “A modified VSLMS algorithm,” in Proceedings of the International Conference on Advance Communication Technology, pp. 615–618, Institute of Electrical and Electronics Engineers, Bangalour, India, January 2011.
View at: Google Scholar
K. Xu, J. Hong, and G. Yue, “A modified LMS algorithm for variable step size adaptive filters,” Journal of Circuits and Systems, vol. 9, no. 4, pp. 115–117, 2004.
View at: Google Scholar
G. Wang, H. Zhao, and P. Song, “Robust variable step-size reweighted zero-attracting least mean m-estimate algorithm for sparse system identification,” Institute of Electrical and Electronics Engineers Transactions on Circuits and Systems II: Express Briefs, vol. 67, no. 6, pp. 1149–1153, 2019.
View at: Publisher Site | Google Scholar
J. Mohammad, S. Mohammad, H. Aykut, and O. Kukrer, “A zero-attracting variable step-size LMS algorithm for sparse system identification,” Signal, Image and Video Processing, vol. 9, 2013.
View at: Google Scholar
P. Song and H. Zhao, “Filtered-x least mean square/fourth (FXLMS/F) algorithm for active noise control,” Mechanical Systems and Signal Processing, vol. 120, pp. 69–82, 2019.
View at: Publisher Site | Google Scholar
H. Guo, Y. S. Wang, N. N. Liu, R. P. Yu, H. Chen, and X. T. Liu, “Active interior noise control for rail vehicle using a variable step-size median-LMS algorithm,” Mechanical Systems and Signal Processing, vol. 109, pp. 15–26, 2018.
View at: Publisher Site | Google Scholar
Y. Fang, X. Zhu, Z. Gao, J. Hu, and J. Wu, “New feedforward filtered-x least mean square algorithm with variable step size for active vibration control,” Journal of Low Frequency Noise, Vibration and Active Control, vol. 38, no. 1, pp. 187–198, 2019.
View at: Publisher Site | Google Scholar
X. Zhu, Y. Fang, J. Hu, Z. Gao, and Z. Miao, “Analysis and validation of the active vibration control of flexible piezoelectric beam with fx-VSSLMS algorithms,” Journal of Vibration, Measurement and Diagnosis, vol. 40, no. 2, pp. 215–221, 2020.
View at: Google Scholar
Y. Fang, X. Zhu, J. Hu, Z. Gao, and Z. Miao, “A VSSFxLMS algorithm and its application in multiple DOF micro-vibration control,” Journal of Vibration Engineering, vol. 33, no. 3, 2020.
View at: Google Scholar
B. Widrow and M. E. Hoff, “Adaptive switching circuits,” IRE Wescon Convention Record, vol. 4, pp. 96–104, 1960.
View at: Publisher Site | Google Scholar
J. Nagumo and A. Noda, “A learning method for system identification,” IEEE Transactions on Automatic Control, vol. 12, no. 3, pp. 282–287, 1967.
View at: Publisher Site | Google Scholar
J. Yi and S. Huaizong, “A new variable step size LMS adaptive filtering algorithm and its simulation,” Signal Processing, vol. 26, no. 9, pp. 1385–1388, 2010.
View at: Google Scholar
D. Meng, P. Xia, and L. Song, “MIMOMH feed-forward adaptive vibration control of helicopter fuselage by using piezoelectric stack actuators,” Journal of Vibration and Control, vol. 24, no. 23, pp. 5534–5545, 2018.
View at: Publisher Site | Google Scholar
P-840 Preloaded Piezo Actuators, https://www.physikinstrumente.com/en/products/linear-actuators/nanopositioning-piezo-actuators/p-840-preloaded-piezo-actuators-101000/#specification.
Q. Huang, Z. Gao, S. Gao, Y. Shao, and X. Zhu, “Research of active vibration control algorithm based on online control channel identification,” in Proceedings of the Third International Conference on Measuring Technology and Mechatronics Automation, Shanghai, China, January 2011.
View at: Google Scholar
Y. Shao, Z. Gao, S. Gao, J. Yi, and X. Zhu, “FXLMS algorithm based multi channel active vibration control of piezoelectric flexible beam,” in Proceedings of the 8th World Congress on Intelligent Control And Automation, Jinan, China, July 2010.
View at: Google Scholar
J.-H. Shin, W.-J. Jung, S.-R. Bae, S.-K. Lee, and M. K. Kwak, “Multi-input multi-output fxLMS algorithm for active vibration control of structures subjected to non-resonant excitation,” Transactions of the Korean Society for Noise and Vibration Engineering, vol. 28, no. 3, pp. 330–338, 2018.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2020 Weiguang Li et al. 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

Views

511

Downloads

691

Citations