Pure delay estimation

Details Pure delay estimation Pure delay estimation

2000-05-09

2004-02-03

Barnie, Rexford (Department: 2643)

Telephonic communications

Echo cancellation or suppression

C379S406060, C379S406080, C370S286000

Reexamination Certificate

active

06687372

ABSTRACT:

TECHNICAL FIELD
The present invention relates generally to pure delay estimation in telecommunication systems.
BACKGROUND
In telecommunication systems echo is generated at a hybrid circuit that transform a 4-wire connection into a 2-wire connection and vice versa. This echo is removed by an echo canceller that estimate the impulse response of the hybrid circuit and uses this estimate to model and cancel the echo. The impulse response estimate typically comprises a dispersive part, which corresponds to the actual echo generation, and a pure delay, which corresponds to the transmission delay from the echo canceller to the hybrid circuit and back to the echo canceller.
Many operators are interested in measurements of such a pure delay for obtaining source information for network planning. Delay measurements could be performed by applying test signals to the links of interest and measuring the response times for these signals. However, such a method would require extra equipment in the system and the measured lines would not be available for normal traffic during measurements.
Methods for determining pure delay estimates from an estimate of the echo path impulse response (estimated by echo canceller filter coefficients) are described in [1-5].
Reference [1] describes a two step method that first determines a filter tap that has exceeded a first, low threshold (“where things start to happen”). A second step starts at this position and determines a second filter tap that exceeds a second, higher threshold. The dispersive part of the impulse response is assumed to start a predetermined number of taps before this second tap and to last a predetermined number of taps from this start position. The method assumes that the first coefficients of the impulse response are zero. However, this assumption is not always true, which may lead to erroneous decisions. For example, the estimated impulse response may contain oscillations with rather long periods, due to filter coefficient estimation using telephone band speech signals in the presence of background noise from the near end side.
In the method described in [2] the dispersive part of the echo path is assumed to start a fixed number of taps before the filter coefficient with maximum value. This method is not especially accurate in any case and it just breaks down if there are multiple echoes.
Reference [3] describes a method for finding center of the echo path by computing cross-correlation function between the input (RIN) and echo (SIN) signals and determines the center of echo path dispersive part as the maximum of this cross-correlation function. The dispersive part is assumed to start half the filter length before its center, which is not an especially accurate method. Also, it should be noted that the delay estimation by cross-correlation is well justified only if the input signal (RIN) is white noise or if the echo path impulse response is just a single impulse. None of the above assumptions are normally true in network echo cancellation.
Reference [4] describes a method that determines regions of the filter impulse response where the energy is concentrated. This is done by summing the absolute values of filter taps within a window of length L and comparing each sum to a fixed threshold. The window slides over the impulse response, and the window positions that have a sum that exceeds the fixed threshold are considered to belong to echo containing regions. The remaining regions are considered to be delays. This method gives a rather coarse delay measurement.
Reference [5] is similar to reference [4] in that it describes a method that determines where in the filter impulse response the energy is concentrated. In this case the average energy of a window is compared to a fixed threshold.
Methods for detecting abrupt changes in signals are also described in [6]. However, these methods are based on step functions and are therefore not suitable for echo impulse response estimates.
SUMMARY
An object of the present invention is to provide an accurate method and system for performing pure delay measurements by using equipment already present in the telecommunication system and without blocking measured lines for normal traffic.
This object is solved by a method and a system in accordance with the appended claims.
Briefly, the present invention uses processing power already present in the system, such as network echo cancellers or control units, to estimate the pure delay of a just completed call. At the end of a call the filter coefficients of the echo canceller represent an estimate of the impulse response of the hybrid circuit, from which the pure delay may be determined. The present invention uses a method based on hypothesis testing to obtain the pure delay from this estimate. Thus, no extra equipment is required and the measured lines are free for normal traffic also during delay measurements.


REFERENCES:
patent: 4587382 (1986-05-01), Yang
patent: 4736414 (1988-04-01), Montagna et al.
patent: 4764955 (1988-08-01), Galand et al.
patent: 4823382 (1989-04-01), Martinez
patent: 5414766 (1995-05-01), Cannalire et al.
patent: 5418849 (1995-05-01), Cannalire et al.
patent: 5675644 (1997-10-01), Sih
patent: 5737410 (1998-04-01), Vähätalo et al.
patent: 5920548 (1999-07-01), El Malki
patent: WO 95/17784 (1995-06-01), None
patent: WO 97/23055 (1997-06-01), None
patent: WO 98/15067 (1998-04-01), None
Van Trees, Harry L., Massachusetts Institute of Technology;Detection, Estimation, and Modulation Theory, Part 1 Detection, Estimation, and Linear Modulation Theory, pp. 18-47.
Basseville, M. et al.;Detection of Abrupt Changes: Theory and Application, Change Detection Algorithms,Prentice-Hall, Englewood Cliffs, NJ, Apr. 1993, pp. 25-65.
Eriksson, A., et al.;Ericsson Echo Cancellers—a Key to Improved Speech Quality, Ericsson Review No. 1, 1996, pp. 25-33.
Part I Changes in the Scalar Parameter of an Independent Sequence; XP-002129388, pp. 25-65.
Yip, P. et al.;An Adaptive Multiple Echo Canceller for Slowly Time-Varying Echo Paths. IEEE Transactions on Communications 38 (1990) Oct., No. 10, New York, US, XP000176867, pp. 1693-1698.
Sieben, S., European Search Report, App. No. SE 9901692, Feb. 1, 2000, pp. 1-4.
Bengtsson, R.; International-Type Search Report, App. No.99016925, Search Request No. SE 99/00579, Jan. 19, 2000, pp. 1-6.

Affiliated with

Also associated with

Rating

Pure delay estimation does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Pure delay estimation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Pure delay estimation will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3295910