Active sound cancellation system for time-varying signals

Abstract

In a fluidborne or structureborne sound system, a method and apparatus arerovided for actively canceling transient acoustic noise. A high-speed controller utilizes a weighted combination of open and closed-loop inputs to provide a correction signal to a cancellation source. The cancellation source introduces a canceling acoustic wave that is equal in amplitude but 180° out of phase with respect to the acoustic noise in order to cancel the acoustic noise within the system. The open-loop input is provided by a database containing a predicted, off-line model of the acoustic noise. The closed-loop input is provided by a combination of 1) an input signal generated by the input acoustic wave and feedback from the cancellation source as measured by an input sensor and 2) an error signal generated at the output of the system as measured by an error sensor. The weighted combination of the open and closed-loop inputs is chosen to minimize the error signal.

What is claimed is: 
     
       1. An apparatus for actively canceling acoustic noise in a fluidborne or structureborne sound system by introducing a canceling acoustic wave into said system from a cancellation source and for adaptively compensating for feedback to said input from said cancellation source, said system having an input for receiving an input acoustic wave and an output for radiating an output acoustic wave, comprising: a database having a predicted, off-line model or the canceling acoustic wave;   an input sensing means for sensing the combination of said input acoustic wave and said feedback to said input and for providing an on-line input signal indicative thereof;   an error sensing means for sensing the combination of said output acoustic wave and said canceling acoustic wave and for providing an error signal indicative thereof; and   a controller means for adaptively modeling said system based upon a model input and an error input, said model input comprising a combination of: 1) a selected percentage of said on-line input signal and 2) a selected percentage of said offline model, and said error input comprising said error signal, said controller means further outputting a correction signal to said cancellation source to introduce said canceling acoustic wave such that said error signal is minimized by said model input, wherein said selected on-line input signal percentage and said selected off-line model percentage are complementary percentages.   
     
     
       2. An apparatus as in claim 1 wherein said controller means comprises a parallel processor. 
     
     
       3. An apparatus as in claim 1 wherein said input sensing means and said error sensing means comprise accelerometers for a structureborne sound system. 
     
     
       4. An apparatus as in claim 1 wherein said input sensing means and said error sensing means are hydrophones for a fluidborne sound system. 
     
     
       5. An apparatus as in claim 1, wherein said input sensing means and said error sensing means are hot wire probes. 
     
     
       6. An apparatus as in claim 1 wherein said input sensing means and said error sensing means are laser Doppler velocimeters. 
     
     
       7. An apparatus as in claim 1 wherein said cancellation source is an omni-directional sound projector. 
     
     
       8. An apparatus as in claim 1 wherein said controller means comprises an algorithm means having said selected percentage of said on-line input signal as a first input, said selected percentage of said off-line model as a second input, said error signal as a third input and an output providing said correction signal to said cancellation source. 
     
     
       9. An apparatus as in claim 8 wherein said algorithm means uses an auto regression with moving average algorithm. 
     
     
       10. An apparatus as in claim 8 wherein said algorithm means uses an auto regression with moving average plus exogenous input algorithm. 
     
     
       11. An apparatus as in claim 1 wherein said database is generated with the aid of an auto regression with moving average algorithm. 
     
     
       12. An apparatus as in claim 1 wherein said database is generated with the aid of an auto regression with moving average plus exogenous input algorithm. 
     
     
       13. A method for actively canceling acoustic noise in a fluidborne or a structureborne sound system, said system having an input for receiving an input acoustic wave and an output for radiating an output acoustic wave, by introducing a canceling acoustic wave from a cancellation source and for adaptively compensating for feedback to said input from said cancellation source, comprising the steps of: providing a database having a predicted, off-line model of the canceling acoustic wave;   sensing a combination signal based on the combination of said input acoustic wave and feedback from said cancellation source;   sensing an error signal based on the combination of said output acoustic wave and said canceling acoustic wave from said cancellation source;   selecting a percentage of said combination signal and a percentage of said off-line model, wherein said combination signal percentage and said off-line model percentage are complementary percentages;   modeling said system with a controller means having a model input and an error input, said model input comprising a combination of: 1) said selected percentage of said combination signal and 2) said selected percentage of said off-line model, and said error input comprising said error signal; and   outputting a correction signal from said controller means to said cancellation source to introduce said canceling acoustic wave such that said error signal is minimized by said model input.   
     
     
       14. A method according to claim 13 wherein said predicted model is modeled with an auto regression with moving average algorithm. 
     
     
       15. A method according to claim 13 wherein said predicted model is modeled with an auto regression with moving average plus exogenous input algorithm. 
     
     
       16. A method according to claim 13 wherein said step of modeling is accomplished with an auto regression with moving average algorithm. 
     
     
       17. A method according to claim 13 wherein said step of modeling is accomplished with an auto regression with moving average plus exogenous input algorithm. 
     
     
       18. In a fluidborne or a structureborne sound system, said system having an input for receiving an input acoustic wave and an output for radiating output acoustic wave, a cancellation source for introducing a canceling acoustic wave into said system, an input sensor for generating an input signal from a combination of said input acoustic wave and feedback from said canceling acoustic wave, an error sensor for generating an error signal from a combination of said canceling acoustic wave and output acoustic wave, and a controller means for adaptively modeling said system based upon said input signal and said error signal whereby a closed-loop response to said system is generated, a method for actively canceling acoustic noise in said system comprising the steps of: providing said controller means with a predicted off-line model of the canceling acoustic wave; selecting a percentage of said off-line model and a percentage of said closed-loop response, wherein said off-line model percentage and said closed-loop response percentage are complementary percentages;   combining said selected percentages of said off-line model and said closed-loop response to form a weighted combination; and   outputting a correction signal to said cancellation source based upon said weighted combination whereby said cancellation source introduces said canceling acoustic wave into said system.