Electrical audio signal processing systems and devices – Acoustical noise or sound cancellation – Within cabin or compartment of vehicle
Reexamination Certificate
1998-12-23
2001-05-08
Mei, Xu (Department: 2644)
Electrical audio signal processing systems and devices
Acoustical noise or sound cancellation
Within cabin or compartment of vehicle
C381S071110, C381S071120, C702S190000, C244S017130
Reexamination Certificate
active
06229898
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to an active control system to reduce ambient vibrations. More particularly, this invention relates to producing a control signal, which reflects adjustments to estimated vehicle structural dynamics, using signal filtering. The control signal is transmitted to one or more vibration generators for reducing ambient vibrations in a vehicle.
2. Brief Description of the Art
Vibrations can reach undesired levels in the interior of a vehicle, such as a helicopter or an airplane resulting in an unpleasant environment in the vehicle. This is particularly true for helicopters, which vibrate in response to loads generated by the rotor blades as they support and propel the aircraft through the air. In the past, helicopter manufacturers have employed a variety of vibration control approaches using principles of mechanical isolation and absorption to render the aircraft ride acceptable with respect to comfort. These approaches, which may be designated as “passive”, attack vibration by modifying the inherent structural dynamics of the aircraft, “de-tuning” its response to prevalent frequencies in the disturbing load signature. An alternate approach employs powered actuators, which apply vibratory loads to the structure in such a way as to produce a vibration field which nullifies the ambient vibration. In the arena of vibration control this type of approach is designated as “active” in the sense that an actuator is commanded to actively generate the nullifying vibration field.
A key component to the successful implementation of an active vibration control system is the proper determination of commands to the actuators. When the relationship between actuator forces and the vibratory response to those forces is known a priori, the proper selection of commands to the actuators can employ a “deterministic” control algorithm. Oftentimes, however, the relationship between actuator forces and vibratory response is either not known a priori, or varies as the structural dynamics are affected by fuel burn-off, cargo re-distribution, or in the case of helicopters, due to changes in the frequency of rotor loading as rotor speed changes during maneuvers. In such cases, “on-line system identification” may be employed to use information acquired during the course of making control changes to establish and/or modify an estimate of the relationship between actuator forces and vibratory response. This estimate is then used during subsequent control iterations to reduce vibration.
A typical environment of a control system in which the present invention is suitably employed is that of a minimum variance controller such as that described in NASA Contractor Report 3821, “Refinement and Evaluation of Helicopter Real-Time Self-Adaptive Active Vibration Controller Algorithms” on pages 1-30. NASA Contractor Report 3821 is hereby incorporated by reference in its entirety herein. The controller uses the approximation that there is a quasi-static linear relationship between actuator commands and system response, which may be expressed by the transfer matrix relationship:
Z
i
=&tgr;(
U
i
−U
i−1
)+
Z
i−1
+V+E,
where:
Z
i
is the n dimensional measured vibration response at time i;
Z
i−1
is the n dimensional state vector of vibration response at time i−1;
U
i
is the m dimensional command to actuators at time i;
U
i−1
is the m dimensional command to actuators at time i−1;
V is the n dimensional change in measured vibration due to change in external disturbances;
E is the change in measured vibration due to change in measurement noise; and
&tgr; is the current estimate of the n by m dimensional transfer matrix.
The minimum variance controller is obtained by minimization of the performance criteria:
J=E
{(
Z
i
−Z
opt
)
T
W
z
(
Z
−Z
opt
)+U
i
T
W
U
U
i
+&Dgr;U
i
T
W
&Dgr;U
&Dgr;U
i
}
where:
J=the performance index (a scalar);
E=expected value;
Z
opt
=the vector of desired vibration at sensor locations (typically zeroes);
Z=the vector of measured vibration at sensor locations;
W
Z
=diagonal weighting matrix on output (vibration) parameters;
U=command input to actuators;
W
U
=diagonal weighting matrix that constrains the amplitude of actuator commands;
&Dgr;U=change in command input to actuators;
W
&Dgr;U
=is a diagonal weighting matrix that constrains the rate of change in command inputs;
T=vector or matrix transpose; and
i=a counter of discrete time increments.
Since performance index J includes measured output parameters and control inputs, each output parameter and control input can be individually weighed to make it more or less important than the other elements.
Using a deterministic controller in a local model, the optimal change in command input to the actuator, &Dgr;U, for the ith rotor revolution is as follows:
&Dgr;
U
i
=−D[W
U
U
i−1
+&tgr;
T
W
z
(
Z
i−1
−Z
opt
)]
where:
&Dgr;U
i
=the optimal control input required to minimize the performance index; and
D=(&tgr;
T
W
z
&tgr;+W
U
+W
&Dgr;U
)
−1
.
The other variables are the same as described above.
When J is minimized, the solution to optimal controller provides information for determining the next control step. For improved control performance an accurate estimate of the matrix &tgr; is essential. In a real time application, it is helpful to employ a Kalman filter to track the elements of &tgr;. Previous attempts to accurately track the values &tgr; because the Kalman filter has been unable to adequately suppress the corrupting influence of system and measurement noise. For example, U.S. Pat. No. 4,819,182 issued to King discloses a method and apparatus for reducing vibration of a helicopter fuselage using a plurality of actuators, which are oscillated at a frequency corresponding to the exciting frequency. A plurality of accelerometers generate signals representing dynamic acceleration. The actuators are controlled by a processor. This patent does not disclose modifying a covariance value as a means for estimating system dynamics. U.S. Pat. No. 4,819,182 is hereby incorporated by reference in its entirety herein.
In many applications, the implementation of the Kalman filter algorithm assumes that measurement noise and system noise are independent from one another, have a Gaussian distribution, and have zero mean. When these assumptions are valid, a standard Kalman filter approach can be effective in reducing the corrupting influence of noise on the estimation of the relationship between actuator forces and vibration. When these assumptions are not valid, standard approaches are ineffective, resulting in poor estimation, and consequently decreased performance of the control system using the estimate for control decisions. The object of the present invention is to address the shortcomings of the standard Kalman filter approach in this regard.
BRIEF SUMMARY OF THE INVENTION
An object of the present invention is to provide a control system that utilizes dynamically validated and updated data to generate a control signal that is used for ambient vibration reduction. Accordingly, one embodiment of the instant invention is drawn to a system having at least one sensor, mounted in a vehicle, for generating successive first and second ambient vibration signals. These ambient vibration signals correspond to a first and second sensed condition. One or more actuators are mounted on the vehicle, for generating a vibration-reducing actuator force that is applied to the vehicle. A data module is coupled to the sensor and the actuator for storing data. This data correlates at least one actuator force with at least one sensor response to the force, and outputs the stored data. A control circuit, including a Kalman filter, receives the first and second ambient vibration signals in succession from the sensor and receives output data from th
Fitzpatrick ,Cella, Harper & Scinto
Mei Xu
Sikorsky Aircraft Corporation
LandOfFree
Active vibration control system using on-line system... 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 vibration control system using on-line system..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Active vibration control system using on-line system... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2510871