Adaptive control system with a corrected-phase filtered error update

Claim 2

An adaptive control system having a system input and a system output, the system comprising: an adaptive filter that inputs one or more reference signals and one or more error input signals, and outputs one or more correction signals; one or more output transducers, each output transducer inputting one of the correction signals and outputting a control signal that combines with a system input; one or more error sensors that senses a system output and outputs an error signal; an error signal filter that inputs each error signal and outputs one or more filtered error signals, wherein the error signal filter includes a delayed Hermitian transpose of a model of an auxiliary path between the output of the adaptive filter and the input of the error signal filter; and, an error signal correlator that inputs each reference signal and each filtered error signal and outputs the one or more error input signals.

Claim 3

A system as recited in claim 1 wherein the error signal filter also includes a summer for each filtered error signal.

Claim 4

A system as recited in claim 1 further comprising a delay element that delays each reference signal before the reference signal is input to the error correlator.

Claim 5

A system as recited in claim 1 wherein the error correlator has at least one multiplier.

Claim 6

A system as recited in claim 4 wherein the adaptive filter has a channel corresponding to a relationship between each reference signal and each correction signal, and the error correlator has a separate multiplier to update each filter element.

Claim 7

A system as recited in claim 1 further comprising one or more input sensors, each input sensor outputting one of the reference signals.

Claim 8

A system as recited in claim 1 wherein the one or more of the error signals are the one or more reference signals.

Claim 9

A system as recited in claim 7 further comprising a delay element that delays each reference signal before the signal inputs the error correlator as a regressor.

Claim 10

A system as recited in claim 1 further comprising a copy of the model of the auxiliary path and a summing junction: wherein the one or more correction signals input the copy of the model, and one or more filtered correction signals output the copy of the model and are summed with the one or more of the error signals at the summing junction to form the one or more reference signals.

Claim 11

A system as recited in claim 9 further comprising a delay element that delays each reference signal before the reference signal inputs the error correlator as a regressor.

Claim 12

A system as recited in claim 3 wherein the delay element delays each reference signal an amount of time so that the delayed reference signal is in phase within 90° of the correlated filtered error signal.

Claim 13

A system as recited in claim 3 wherein the delay element delays each reference signal an amount of time that is an estimate of the effective response time of the auxiliary path.

Claim 14

A system as recited in claim 3 wherein the delay element delays each reference signal the same amount of time as a delay of the delayed Hermitian transpose.

Claim 15

A system as recited in claim 3 wherein the model of the auxiliary path has a series of tap weights in the time domain corresponding to each path between each adaptive filter output and each input to the error signal filer, each series of tap weights has the same fixed number of tap weights, and the delay element delays the reference signal an amount of time essentially equal to an amount of time represented by the fixed number of tap weights.

Claim 16

A system as recited in claim 1 wherein the model of the auxiliary path has a series of tap weights in the time domain corresponding to each path between each adaptive filter output and each input to the error signal filter, and the error signal filter has a filter channel for each model channel, the filter channel having a shifted and reversed series of tap weights.

Claim 17

A system as recited in claim 15 comprising not more than one output transducer and not more than one error sensor wherein the model has not more than one model channel, and the filter channel is a delayed complex conjugate of the model channel in the frequency domain.

Claim 18

A system as recited in claim 15 wherein the model channels are also transposed to form the filter channels in the error signal filter.

Claim 19

A system as recited in claim 17 wherein the error signal filter further has a summer corresponding to each filtered error signal.

Claim 20

A system as recited in claim 15 wherein each series of tap weights has the same fixed number of tap weights and the delay is determined according to the fixed number of tap weights.

Claim 21

A system as recited in claim 1 wherein the model of the auxiliary path is an adaptive model that adapts on-line, and the control system further comprises a random noise source.

Claim 22

A system as recited in claim 1 wherein the model of the auxiliary path models the auxiliary path over one or more selected frequencies only.

Claim 23

A system as recited in claim 1 wherein the adaptive filter is an FIR filter.

Claim 24

A system as recited in claim 1 wherein the adaptive filter is an IIR filter.

Claim 25

A system as recited in claim 1 wherein the adaptive control system is implemented in an active acoustic attenuation system, and the system input is an input acoustic wave, the control signal is a canceling acoustic wave and the system output is an output acoustic wave.

Claim 26

A system as recited in claim 24 wherein the active acoustic attenuation system is an active sound attenuation system, and the output transducer is a loudspeaker, and the error sensor is an error microphone.

Claim 27

A system as recited in claim 24 wherein the active acoustic attenuation system is an active vibration attenuation system, and the output transducer is a shaker and the error sensor is an accelerometer.

Claim 28

An adaptive control method in a system having a system input and a system output comprising the steps of: adaptively generating n correction signals by processing m reference signals through an adaptive filter having n×m channels; generating a control signal from each correction signal, the n control signals combining with the system input to yield the system output; sensing the system output and generating p error signals in response thereto; filtering the p error signals to generate n filtered error signals by inputting the p error signals to an error signal filter having n×p channels, the error signal filter being the delayed Hermitian transpose in the frequency domain of an auxiliary path model having p×n channels; correlating each of the m reference signals with each of the n filtered error signals to generate a corresponding error input signal for each of the n×m channels in the adaptive filter; and adapting each of the n×m channels in the adaptive filter from the corresponding error input signal.

Claim 29

An adaptive control method in a system having a system input and a system output comprising the steps of: adaptively generating n correction signals by processing m reference signals through an adaptive filter having n×m channels; generating a control signal from each correction signal, the n control signals combining with the system input to yield the system output; sensing the system output and generating p error signals in response thereto; filtering the p error signals to generate n filtered error signals by inputting the p error signals to an error signal filter having n×p channels, the error signal filter being the delayed Hermitian transpose in the frequency domain of an auxiliary path model having p×n channels; correlating each of the m reference signals with each of the n filtered error signals to generate a corresponding error input signal for each of the n×m channels in the adaptive filter; and adapting each of the n×m channels in the adaptive filter from the corresponding error input signal.

Claim 30

A method as recited in claim 27 further comprising the steps of: delaying each of the m reference signals before correlating the reference signal with each of the n filtered error signals.

Claim 31

A method as recited in claim 27 wherein each of the n×p channels in the delayed Hermitian transpose is formed by taking the complex conjugate of the corresponding p×n channel in the n×p auxiliary path model.

Claim 32

A method as recited in claim 27 which is implemented in the time domain, wherein the p×n channels in the auxiliary path model are each a series of tap weights, and each of the n×p channels in the error signal filter is formed by reversing the order of the tap weights.

Claim 33

A method as recited in claim 30 further comprising the step of: delaying each of the m reference signals before correlating the reference signal with each of the n filtered error signals.

Claim 34

A method as recited in claim 28 which is implemented in the time domain, wherein the p×n channels in the auxiliary path model are each a series of N c tap weights, and each reference signal is delayed the same amount of time as represented by N c taps.

Claim 35

A method as recited in claim 28 wherein each of the reference signals is delayed an amount of time that is an estimate of the effective response time of the auxiliary path.

Claim 36

A method as recited in claim 27 wherein the p×n auxiliary path model is generated adaptively on-line using a random noise source.

Claim 37

A method as recited in claim 27 wherein the p×n auxiliary path model models the auxiliary path over one or more selected frequencies only.

Claim 38

A method as recited in claim 27 where in the adaptive filter is an FIR filter.

Claim 39

A method as recited in claim 27 wherein the adaptive filter is an IIR filter.

Claim 40

A method as recited in claim 27 wherein the system is an active acoustic attenuation system, and the system input is an input acoustic wave, the n control signals are each a canceling acoustic wave, and the system output is an output acoustic wave.

Claim 41

A method as recited in claim 27 wherein the active acoustic attenuation system is an active sound attenuation system, and each of the n canceling acoustic waves is generated using a loudspeaker, and each of the p error signals is generated using a microphone.

Claim 42

A method as recited in claim 39 wherein the active acoustic attenuation system is an active vibration attenuation system, and each of the n canceling acoustic waves is generated using a shaker, and each of the p error signals is generated using an accelerometer.

Claim 43

A method as recited in claim 27 further comprising the step of measuring each of the m reference signals with an input sensor.

Claim 44

A method as recited in claim 41 further comprising the step of: delaying each of the reference signals before correlating the reference signal with each of the n filtered error signals.

Claim 45

A method as recited in claim 27 further comprising the step of providing at least one of the error signals to the adaptive filter as at least one of the reference signals.

Claim 46

A method as recited in claim 43 further comprising the step of: delaying each of the reference signals before correlating the reference signal with each of the n filtered error signals.

Claim 47

A method as recited in claim 27 further comprising the steps of: filtering the n correction signals through a copy of the p×n auxiliary path model to generate p filtered correction signals; and adding at least one of the p filtered correction signals to the corresponding error signal to general p reference signals; providing the p reference signals to the adaptive filter for processing.

Claim 48

A method as recited in claim 45 further comprising the step of: delaying each of the reference signals before correlating the reference signal with each of the n filtered error signals.

Claim 49

A method as recited in claim 27 wherein the adaptive filter outputs not more than one correction signal, not more than one error signal is generated and the error signal filter has not more than one channel which is determined by taking a delayed complex conjugate of the auxiliary path model.