Browsing by Author "Liang, Keng-Wei"

Now showing 1 - 2 of 2

An Open-Loop Pole-Zero Placement Method for Active Noise Control Headphones
(2017-07-01)
Liang, Keng-Wei
;
Hu, Jwu-Sheng
;
交大名義發表
;
National Chiao Tung University
Active noise control headphones for hearing protection are effective in the low frequency range, and both analog and digital controllers are used widely. Considering the bandwidth of noise reduction, the analog controller is more beneficial due to its shorter delay compared with the digital one. Analog controllers with fixed circuit structures are preferable for practical implementation. However, fixed-order controllers with low complexity, such as a biquad filter, are difficult to design for a high reduction level. On the other hand, a high order one, getting from the H-inf-norm optimization, is costly due to a larger number of circuit components. In this paper, an H-2-norm optimization method that places the controller's pole-zero locations to meet desired robustness and performance, is proposed. To determine an initial guess with an appropriate order of the controller, a loop-shaping method by loop transfer recovery (LTR) is introduced. The experimental results show the effectiveness of the method.
Optimal Controller Design for Virtual Sensing With Independent Noise Source Measurement
(2019-01-01)
Liang, Keng-Wei
;
Hu, Jwu-Sheng
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In certain applications, when the placement of sensors at the target location is not feasible, a common practice is to obtain the target information indirectly via sensors at other locations. This technique is known as virtual sensing (or soft sensing), and involves various signal processing techniques. Virtual sensing is applied to such systems to estimate the error signal for feedback control. However, issues regarding both stability and performance degradation become significant because the estimation is not precise. In this brief, we propose a virtual sensing control (VSC) method using a feedforward controller to compensate the mismatch between an error sensor and virtual sensor under the feedback control. The results show that when an independent noise source measurement is available, a feedforward controller is more appropriate than estimation as it maintains the stability of the feedback system while achieving a good performance enhancement. For such a VSC system, the design of both feedforward and feedback controllers is coupled. In this brief, we derive an optimal controller design method to maximize the noise reduction capability. We show the advantages of the proposed VSC method by applying it to the implementation of an active noise control system.