Telephonic communications – Supervisory or control line signaling – Signal receiver
Reexamination Certificate
1999-04-27
2004-02-17
Harvey, Minsun Oh (Department: 2644)
Telephonic communications
Supervisory or control line signaling
Signal receiver
C379S283000, C379S339000, C379S142180, C379S418000, C704S262000, C704S219000, C704S216000, C702S073000
Reexamination Certificate
active
06694010
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to the art of recognizing a single frequency tone in a signal, and more particularly, the recognition of such a tone in a signal without any prior knowledge or expectation of the tone frequency.
BACKGROUND OF THE INVENTION
In echo-cancelers, it is necessary to freeze or disable the echo-cancelling operation when specific tones are propagated through the echo-canceller. One well-known case of such a tone is the AC
15
tone used for inter-PBX signalling (as set forth in Signalling System AC
15
, British Telecom Network Requirement Document No. BTNR181). PBXs using this signalling scheme do so by using a single frequency tone of 2280 Hz. It is possible for two interconnected PBXs to generate the 2280 Hz tone simultaneously. Thus, if an echo-canceller is in the exchange path of the two PBXs, the echo-canceller could eliminate the tone in one direction simply because the tone in this direction can be considered as the echo of the signal in the opposite direction.
There are many other standardized bi-directional signal tone exchanges. For example:
1) Continuity testing (1004 Hz);
2) AC
15
(2280Hz);
3) ITU-T Signalling System No.
4
(2040 Hz and 2400 Hz);
4) ITU-T Signalling System No.
5
(2400 Hz and 2600 Hz);
5) ITU-T Signalling System No.
6
and No.
7
(2010 Hz);
6) China PTT TS-
02
(2600 Hz.
These, as well as many other standardized single tone exchanges, employ numerous different single tone frequencies which must be detected for various purposes.
One way to solve the tone detection problem is to tailor multiple specific-tone detectors to detect every one of the frequencies involved. These prior art tone detectors are capable of detecting a single tone based on an expectation of what that tone should be. The drawback with these types of detectors is that their usability is restricted to the pre-programmed frequencies, i.e., one detector is required for each frequency one is trying to detect. Thus, the processing requirement grows linearly with the number of distinct frequencies, which must be detected. It would be more advantageous to have a more generic tone detector, which has the capability to recognize if a signal is a pure sinusoidal tone and, if, so, to determine the frequency thereof. Such a detector would have the distinct advantage of not having to know all the single frequencies involved in the network. Also, because such a detector can dynamically determine the frequency, various other benefits can be gained. For instance, it is possible to detect the echo protector tone used in the prologue of full-duplex voiceband data such as defined in V.
25
EPT, V.
8
, or V.
8
bis standards (see, respectively,
ITU-T Recommendation V.
25
(1996), Automatic answering equipment and general procedures for automatic calling equipment on the general switched telephone network including procedures for disabling of echo control devices for both manually and automatically established calls;
ITU-T Recommendation V.
8
(1994), Procedure for starting sessions of data transmission over the general switched telephone network; and
ITU-T Recommendation V.
8
bis (1996), Procedure for the identification and selection of common modes of operation between data circuit-terminating equipment (DCEs) and between data terminal equipment (DTEs) over the general switched telephone network and on leased point-to-point telephone-type circuits).
A number of prior art techniques have been proposed to model sinusoidal signals. These include Prony's method or Pisarenko's spectral line decomposition. Unfortunately, these two methods use fairly complex mathematical operations: Prony's method requires a polynomial root extraction and Pisarenko's technique requires calculating Eigenvalues and Eigenvectors. (See, respectively, deProny, Baron Gaspard Riche, “Essai expérimental et analytique: sur les lois de la dilatabilité de fluides elastique et sur celles de la force expansive de la vapeur de l'alkool, à différentes températures”,
Journal de l'École Polytechnique
, 1795, volume 1, cahier 22, 24-76; and
Pisarenko, V. F., “The Retrieval of Harmonics from a Covariance Function”,
Geophysical Journal of the Astronomical Society
, 1973, 33:347-366).
However, one technique, which tends to be mathematically simpler, is the auto-regressive modeling or linear prediction coding. For instance, U.S. Pat. No. 5,495,526 to Cesaro et. al. and U.S. Pat. No. 5,619,565 to Cesaro et. al. disclose single frequency tone recognition techniques based on the auto-regressive model. The drawback with the techniques outlined in these patents is the imprecision of the algorithms used to determine the auto-regressive coefficients. Secondly, these techniques use an adaptive filter, which consequently results in a comparatively long time required to detect a single tone frequency, i.e., poor convergence time.
SUMMARY OF INVENTION
The invention seeks to overcome many of the disadvantages of the prior art and more particularly provide a single frequency tone detector able to detect a tone in a signal without any prior knowledge or expectation of the tone frequency.
Broadly speaking, one aspect of the invention provides a method for recognizing a tone in a signal using at least a second order non-biased auto-regressive model defined by a first auto-regressive coefficient and a second auto-regressive coefficient. The method includes:
(a) sampling the signal at a constant sampling rate and, for each sample thereof, recursively determining a finite number of current correlation coefficients using an exponentially weighted, future sliding equivalent of the signal. The signal is time reversed such that prior to each determination of the current correlation coefficients as aforesaid, a current received sample and consecutively received previous samples of the signal are time reversed such that the current received sample is defined as an oldest sample, an oldest of the consecutively received previous samples is defined as a current sample and, where more than two consecutively received previous samples of the signal are used in the determination, all the consecutively received previous samples other than the oldest thereof are also defined to be time reversed as to their received sampling order. The current correlation coefficients are determined using pre-existing values of the correlation coefficients determined in a previous determination, the current sample and at least two consecutively received previous samples of the signal;
(b) periodically determining at least a second auto-regressive coefficient for modeling the signal using the correlation coefficients; and
(c) recognizing the presence of the tone based on the second auto-regressive coefficient.
Another aspect of the invention provides a method for detecting a tone and its frequency in a signal using at least a second order non-biased auto-regressive model defined by a first auto-regressive coefficient and a second auto-regressive coefficient. The method includes:
(a) sampling the signal at a constant sampling rate and, for each sample therof, recursively computing a finite number of current correlation coefficients using an exponentially weighted future sliding equivalent of the signal. The signal is time reversed such that prior to each computation of the current correlation coefficients as aforesaid, a current received sample and consecutively received previous samples of the signal are time reversed such that the current received sample is defined as an oldest sample, an oldest of the consecutively received previous samples is defined as a current sample and, where more than two consecutively received previous samples of the signals are used in the computation, all the consecutively received previous samples other than the oldest thereof are also defined to be reversed as to their received sampling order. The current correlation coefficients are computed using correlation coefficients computed in a previous iteration, the current sample of the signal and at least two consecutively received previous samples o
Alcatel Canada Inc.
Blake Cassels & Graydon LLP
Harvey Minsun Oh
Tran Con P.
LandOfFree
Recognition of a single frequency tone does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Recognition of a single frequency tone, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Recognition of a single frequency tone will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3301626