Communications: directive radio wave systems and devices (e.g. – Directive – Including a steerable array
Reexamination Certificate
2003-07-10
2004-12-28
Gregory, Bernarr E. (Department: 3662)
Communications: directive radio wave systems and devices (e.g.,
Directive
Including a steerable array
C342S372000, C342S374000, C367S122000, C367S123000, C381S092000
Reexamination Certificate
active
06836243
ABSTRACT:
TECHNICAL FIELD
The invention relates to a system for processing a signal being emitted from a target signal source into a noisy environment, wherein said target signal source is located in a target signal source direction &phgr;
s
with regard to the position of a transducer array, the system comprising: the transducer array having M transducers for receiving said signal being mixed with noise, each of the transducers generating a corresponding transducer output signal, respectively and a beamformer for receiving and filtering said M transducer output signals in order to generate at least one output signal y′
i
(n) i=1 . . . N, said beamformer comprising filter coefficients defining a predetermined filtering characteristic, e.g. a desired look direction.
The invention further relates to a method for processing a signal being emitted from a target signal source into a noisy environment, carried out by said system.
BACKGROUND OF THE INVENTION
Beamforming systems are typically used to improve the quality of a received signal by processing the signals received by an array of multiple transducers. Transducer array signal processing can be utilized to enhance the performance of the receiving system capturing the desired signal that has been emitted into a noisy environment. Beamforming methods enable steering of the maximum sensitivity (look direction) of the transducer array towards any desired signal source by changing the beamforming filter coefficients. Typical applications can be found in radio communications, radar signal processing, underwater acoustics, and speech acquisition for teleconferencing and hands-free systems.
It is known in the art that according to principle of reciprocity the signal sources and receivers are interchangeable and by changing the direction of the signal flow and replacing the receiving transducers with transmitting transducers a corresponding signal can be emitted from the transducer array to direction &phgr;
s
.
Different beamforming methods have been studied widely in the literature. One of the most frequently studied methods is adaptive beamforming in which the filter coefficients are adjusted according to the received signal characteristics. A well known adaptive beamforming method has been presented by Frost (Frost O. L., “An algorithm for linearly constrained adaptive array processing”, Proc. IEEE, Vol. 60, No. 8, pp. 926-935, August 1972). Frost's method has been further developed by Griffiths and Jim (Griffiths L. J., Jim C. W., “An alternative approach to linearly constrained adaptive beamforming”, IEEE Trans. Antennas Propag., Vol. AP-30, No. 1, pp. 27-34, January 1982). The fundamental problem of these adaptive beamformers is that the adaptive filters are designed to cancel the noise term in the beamformer output signal but, in practice, the noise estimate also contains a component that correlates with the desired signal. Therefore, the adaptive filters do not only attenuate the noise but they also cause unpredictable distortion of the desired signal. Said correlation is typically caused by multi-path propagation, misaligned look direction, or improper modeling and variations in the propagation medium.
In order to alleviate the desired signal distortion it is more favorable to design a fixed beamforming filter that is optimized for a given application.
The advantage of a fixed beamforming implementation is that the fixed filter coefficients can be optimized based on a priori knowledge of the source and the medium as well as the desired performance criteria so that the filtering performance is deterministic.
Such a design can be formulated as a spatio-temporal filter design problem where the array geometry (transducer positions) and the temporal sampling interval define a spatio-temporal sampling grid. It is known in the art that both the spatio-temporal sampling grid and the corresponding beamformer filter coefficients can be optimized without suffering from the distortion of the desired signal (Kajala M., Hämäläinen M., “Broadband beamforming optimization for speech enhancement in noisy environments”, IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, October 1999). Such a system and method will now be described in more detail by referring to
FIGS. 10
to
12
.
The system according to
FIG. 10
comprises a transducer array
10
, which comprises several transducers
10
-j, j=1 . . . M. The transducers receive a signal emitted from a signal source
20
and being mixed or superimposed with noise. Transducers can be e.g. microphones to receive and transform pressure signals into electrical signals.
Said signal source is located in a target signal source direction &phgr;
s
with regard to the position of the transducer array
10
. As illustrated in
FIG. 11
said target signal source direction &phgr;
s
generally represents the 3-dimensional position of the target signal source in space, e.g. in Cartesian coordinates &phgr;
s
=(X
s
, Y
s
, Z
s
) or in spherical polar coordinates &phgr;
s
=(r
s
, &phgr;
s
, &thgr;
s
), relative to the transducer array
10
which is in
FIG. 11
assumed to be located in the center of the coordinate system. If signals come from the far field, they can be modeled as plane waves. In that case, the distance r
s
becomes redundant and the signal source direction &phgr;
s
reduces to &phgr;
s
=(&phgr;
s
, &thgr;
s
).
When receiving said signal the transducers
10
-
1
. . .
10
-M of said transducer array
10
according to
FIG. 10
generate analogue signals which are subsequently sampled by a multichannel analogue to digital converter
15
. The sampled signals x
1
(n) . . . x
M
(n) are fed into an elementary beamformer
30
. Said beamformer is represented by a filter bank comprising M FIR filters
35
-
1
. . .
35
-M. An output signal y′(n) of said beamformer
30
is formed by adding the output signals x′
1
(n) . . . x′
M
(n) of said FIR filters, respectively.
The beamformer
30
is adapted to steer the beam shape of the transducer array
10
to any particular look direction &phgr;
i
, by using a filter with an appropriate set of fixed filter coefficients. Said look direction &phgr;
i
of the beamformer does in general not necessarily coincide with the target signal source direction &phgr;
s
.
In the case that the signal is emitted from said signal source
20
into a noisy environment the sampled transducer signals x
1
(n) . . . x
M
(n) represent a noisy signal. At the same time, if the look direction &phgr;
i
to which the directivity of the beamformer
30
is correctly adjusted coincides with the target signal source direction &phgr;
s
, the noisy components in the transducer signals are substantially suppressed. The output signal y′(n) of the filter bank
30
estimates the signal emitted from said signal source
20
; i.e. the signal to noise ratio SNR or any other appropriate measure of quality of the filtered signal y′(n) is maximized. Due to the fact that the beamformer
30
in
FIG. 10
comprises only one set of fixed filter coefficients, the successful use of said system requires apriori knowledge about the target signal source direction &phgr;
s
. In order to enhance the beamformer output signal y′(n), it is necessary to align the look direction &phgr;
i
of the beamformer with said target signal source direction &phgr;
s
by providing an appropriate set of fixed filter coefficients.
The sequential alignment of the elementary beamformer's directivity to several predetermined look directions &phgr;
i
, i=1 . . . N, e.g. to track a moving signal source, a respective set of fixed filter coefficients for each predetermined look direction &phgr;
i
has to be stored in a memory and implemented in said filter before usage. This would be very ineffective. Consequently, the system according to
FIG. 10
is not appropriate for processing signals emitted from different signal sources located in different target directions &phgr;
s,p
p=1 . . . P with respect to the transducer array
10
.
FIG
Hāmäläinen Matti
Kajala Matti
Gregory Bernarr E.
Mull Fred H.
Nokia Corporation
Ware Fressola Van Der Sluys & Adolphson
LandOfFree
System and method for processing a signal being emitted from... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for processing a signal being emitted from..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for processing a signal being emitted from... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3334544