Data processing: artificial intelligence – Neural network – Structure
Reexamination Certificate
2000-08-11
2003-07-29
Davis, George B. (Department: 2121)
Data processing: artificial intelligence
Neural network
Structure
C706S023000, C706S014000
Reexamination Certificate
active
06601054
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates to active control of acoustic and structural vibrations with an uncertain vibration statistic or propagation path, which control is achieved by attenuating at least one primary (i.e. unwanted) vibration, at least in part, by injecting at least one specially generated cancelling (i.e. secondary) vibration. An acoustic vibration is gas-, liquid- or solid-borne. A structural vibration is usually a solid-borne longitudinal, torsional or flexural vibration; and is also called a mechanical vibration. Active control of acoustic and structural vibrations considered herein emcompasses active noise control, active sound control, active control of vibration, active noise cancellation, active vibration attenuation, active structural acoustic control, active acoustic and structural vibration control, etc. Active control of acoustic and structural vibrations is herein referred to as active vibration control (AVC) or active acoustic and structural vibration control (AASVC).
There is an enormous amount of research results on AVC (or AASVC). Most of these prior-art results and their applications can be found or tracked down from the books by P. A. Nelson and S. J. Elliott,
Active Control of Sound,
Academic Presss (1992); by S. M. Kuo and D. R. Morgan,
Active Noise Control Systems—Algorithms and DSP Implementations,
John Wiley & Sons, Inc. (1996); by C. R. Fuller,
Active Control of Vibration,
Academic Presss (1996); and by C. H. Hansen and S. D. Snyder,
Active Control of Noise and Vibration,
E & F N Spon (1997), and from the articles and journals referred to in these books. To facilitate discussion on the background of the invention, four basic AVC (i.e. AASVC) systems in the prior art for actively controlling acoustic and structural vibrations with an uncertain vibration statistic or propagation path are briefly described and their shortcomings discussed here.
For brevity, signal conditioners for converting the outputs of sensors and controllers into proper signal forms (e.g. digital signals within a certain frequency band and amplitude range) suitable for subsequent processings are omitted in the discussions in this section on the background of the present invention. An example signal conditioner, that for conditioning the output of an acoustic sensor, comprises a preamplifier, an anti-aliasing filter and an A/D converter.
A first basic AVC system, which is usually used for attenuating a broadband primary vibration, comprises a vibrational reference sensor (e.g. a microphone, hydrophone, accelerometer, velocity transducer, displacement transducer, or strain sensor), a secondary source (e.g. a loudspeaker, horn, underwater sound projector, piezoeletric actuator, electromagnetic shaker, hydraulic actuator, pneumatic actuator, proof mass actuator, electrodynamic and electromagnetic actuator, magnetostrictive actuator, or shape memory alloy actuator), an error sensor (e.g. a microphone, hydrophone, accelerometer, velocity transducer, displacement transducer, or strain sensor) and an adaptive linear filter used as a controller to drive the seconary source. The adaptive linear filter inputs the reference signal from the reference sensor and outputs a control signal for driving the seconary source. Using the error signal from the error sensor and a (mathematical) model of the secondary path (from the output of the adaptive linear filter to the output of the error sensor), an FXLMS (i.e. the filtered-x least-mean-square) algorithm is used to adjust the weights of the adaptive linear filter online to reduce the error signal from the error sensor.
A second basic AVC system, which is usually used for attenuating narrowband primary vibration whose waveform is ordinarily periodic or nearly periodic between its changes, comprises a nonvibrational reference sensor (e.g. a magnetic or optical pickup sensing the rotation of a toothed wheel), a seconary source, an error sensor and a waveform synthesizer. The nonvibrational reference sensor provides a synchronization signal for the waveform synthesizer to lock on to. Acting as a controller, the waveform synthesizer inputs the nonvibrational reference signal from the nonvibrational reference sensor and outputs a synthesized waveform for driving the seconary source. Using the error signal from the error sensor, an adaptation unit is used to adjust the synthesized waveform online to reduce the error signal from the error sensor.
A third basic AVC system uses no (vibrational or nonvibrational) reference sensor and is usually used for attenuating colored (e.g. correlated) primary vibration. It comprises a secondary source, an error sensor and an adaptive linear filter used as a controller. An estimate of the primary vibration component of the error signal from the error sensor is obtained by adding the error signal and an estimate of the secondary vibration component of the error signal, which is obtained by passing the control signal from the adaptive linear filter through a model of the secondary path. The adaptive linear filter inputs said estimate of the primary vibration component and outputs a control signal for driving the secondary source. Using the error signal from the error sensor and the model of the secondary path, an FXLMS algorithm is used to adjust the weights of the adaptive linear filter online to reduce the error signal.
A fourth basic AVC system, which is usually used for attenuating primary vibration of a large mechanical system or inside an enclosure or a large-dimension duct, comprises multiple reference sensors, multiple output actuators, multiple error sensors and an adaptive linear filter as a controller. The adaptive linear filter inputs the reference signals from the reference sensors and outputs a control signal for driving each of the output actuators. Using the error signal from each of the error sensors and the models of the secondary paths (from the output of the adaptive linear filter to each output of error sensors), a multiple-channel FXLMS (filtered-x least-mean-square) algorithm is used to adjust the weights of the adaptive linear filter online to reduce the error signals.
These basic and other AVC systems in the prior art for actively controlling acoustic and structural vibrations with an uncertain vibration statistic or propagation path suffer from at least one of the following shortcomings:
1. Use of an error sensor at each objective point: If a primary path (from the output of a reference sensor to the output of an error sensor) or a secondary path is uncertain (i.e. unknown or time-varying) and if either the weights of an adaptive linear filter or a synthesized waveform generated by a waveform synthesizer is adjusted online, one error sensor must be used at each objective point, which adds to the cost of the AVC system and may cause inconvenience. In some applications, an error sensor senses not only primary and secondary vibrations but also a third vibration that is correlated with the primary vibration and thus degrades the AVC performance.
2. Relatively slow convergence of a weight/waveform adjustment algorithm: If a primary path or a secondary path undergoes a relatively rapid change, an LMS algorithm (e.g. an FXLMS algorithm or a filtered-u recursive LMS algorithm), a RLS algorithm or an adaptation unit is not fast enough in adjusting the weights of an adaptive linear filter or the synthesized waveform generated by a waveform synthesizer used as a controller for some applications.
3. Online or frequent modelling of a secondary path: If a secondary path is time-varying, a model of it needs to be adjusted either online or offline from time to time.
4. Use of a high-order adaptive linear transversal filter: If an adaptive linear recursive filter is the natural one to be used as a controller or as a model of a secondary path, but an adaptive linear transversal filter is used instead to approximate it, the order of the adaptive linear transversal filter is usually large and requires much computation for its weight adjustment. In fact, the better the approximation i
Lo James T.
Yu Lei
Davis George B.
Maryland Technology Corporation
LandOfFree
Active acoustic and structural vibration control without... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Active acoustic and structural vibration control without..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Active acoustic and structural vibration control without... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3015358