Electrical computers: arithmetic processing and calculating – Electrical digital calculating computer – Particular function performed
Reexamination Certificate
1999-06-04
2002-04-30
Mai, Tan V. (Department: 2121)
Electrical computers: arithmetic processing and calculating
Electrical digital calculating computer
Particular function performed
Reexamination Certificate
active
06381623
ABSTRACT:
BACKGROUND OF THE INVENTION
FIELD OF THE INVENTION
The invention lies in the telecommunications field. More specifically, the invention relates to a method for adaptive filter adjustment in a QAM/CAP system, and to a corresponding CAP receiver.
Reference is had to the article by Chen et al. entitled “Design of Digital Carrierless AM/PM Transceivers” from AT&T and Bellcore, TEI.1.4/92-149, Aug. 19, 1992. The article provides an introduction to the design of digital, carrierless CAP transceivers and receivers. It will not be necessary to repeat that information herein and, accordingly, the basic structures will be described only briefly here.
It has been found that, with large amounts of data, a frequency-division multiplex method is better than a baseband duplex channel method (common-frequency method) with echo compensation for duplex data transmission via subscriber access lines, despite the greater bandwidth required. The crosstalk that is dominant in the relevant speed range can be suppressed by selective filtering. Thus, for rapid data transmission in the telephone subscriber access area, single-carrier methods such as QAM (quadrature amplitude modulation) and CAP (carrierless amplitude/phase modulation) are currently being discussed under the keyword VDSL (very high bit rate digital subscriber line). They allow the various frequency bands to be achieved for the outbound and inbound return directions in a simple manner by suitable selection of the carrier and mid-frequencies. The data rates in this case vary over a range from about 2 Mbit/s to 50 Mbit/s. Both symmetrical and asymmetric operation are intended to be possible. Symmetrical operation means equal data rates in both directions, which is generally required for commercial applications. Asymmetric operation means different data rates in the two directions. The latter is generally sufficient in private communication (high data rate to the subscriber, low data rate from the subscriber to the service provider).
Until now, the discrete multitone method has been standardized in the ADSL Standard (ADSL=asymmetric digital subscriber line). However, it can be assumed that single-carrier methods using CAP technology will also be standardized in the pertinent application area.
In a CAP system, the binary data are assembled into groups of L bits at the transmission end, and are supplied to a coder. The coder allocates a point in a two-dimensional signal space to each of the 2
L
combinations. The point is defined by the cartesian coordinates (x, y). The signal point may also be regarded as a point in the complex number planes.
A CAP modulator will be considered first of all. The signal points to be transmitted, defined by the two cartesian coordinates a
k
and b
k
at the time kT, are sampled at the clock rate (symbol rate) f
T
=1/T and are passed to separate sub-channels. Each of the sub-channels contains a transmission filter. After this, the signals in the two sub-channels are added, and the transmission signal is passed to the transmission channel.
In a corresponding CAP receiver, which forms the matching element to the CAP transceiver, the received input signal is subjected to A/D conversion at a predetermined sampling rate. After the digital A/D conversion, digital level regulation is carried out, in order to adjust the level of the received/wanted signal to a constant value so that, as far as possible, it is independent of the transmission line and of the crosstalk of the transmission signal in the system, as well as all other out-band interference. After digital level regulation, the samples are passed to a pair of parallel reception filters. In this case, the coefficients of the filter pair must be adjusted adaptively, starting with a fixed coefficient set, which is designed for a specific line length. After reception filtering, the signal values are sampled at the symbol rate, and are passed to a decision maker. The decision maker has the task of assigning received data to each received value pair. Once adjustment has been carried out and if the interference is sufficiently low, these received values match the transmitted data values, apart from a constant delay and an additional rotation of the complex data vector through ±90° or ±180°.
A QAM/CAP system offers the possibility to carry out the channel distortion elimination with the pair of filters that are required in any case for demodulation. These are designed as a Hilbert transformer. The coefficients of the filter pair must be adjusted adaptively for this purpose.
In actual transmission systems, the signal at the input of the receiver is in principle composed of two elements, namely the wanted band signal with in-band interference as well as out-band interference. Out-band interference relates primarily to echoes from the system's own transmitter, whose spectral elements are in the adjacent frequency band, and signals in the low-frequency range, which originate from services that use the same transmission frequency, such as ISDN.
In-band interference causes symbol errors directly, depending on the intensity and the stepped nature of the transmission. The errors can in turn be reduced to a limited extent by suitable coding measures. Out-band interference, on the other hand, can be largely suppressed by suitable selective filters. For complexity reasons, it is worthwhile to carry out both analog initial filtering and digital filtering with the reception filter pair in the QAM/CAP system.
In unfavorable applications, for example in the case of long transmission lines with a high cable losses, the proportion of out-band interference may be greater than the wanted signal despite analog initial filtering at the input of the digital receiver. The reception filter pair must thereby compensate for the linear distortion in the channel and filter out the remaining out-band interference by a sufficiently high level of stop-band attenuation.
The superimposed out-band interference, however, has a negative effect on the process of adjusting the filter pair, since, for coefficient adjustment, the input signal must be correlated with a suitable error criterion.
SUMMARY OF THE INVENTION
It is accordingly an object of the invention to provide a method for adaptive filter adjustment in a QAM/CAP system, which overcomes the above-mentioned disadvantages of the heretofore-known devices and methods of this general type and which ensures that the filter pair in the QAM/CAP system settles in a stable manner despite out-band interference, and of providing a corresponding CAP receiver.
With the foregoing and other objects in view there is provided, in accordance with the invention, a method for adaptive filter adjustment of a filter pair of a QAM/CAP receiver, which comprises:
providing an A/D converter, a digital level regulator, an adaptive reception filter pair, and a decision maker connected downstream thereof in a signal flow direction;
defining a correction value as a product of a distortion elimination error, a delayed input value, and a manipulated variable &ggr;; and
forming an n
th
filter coefficient of a filter in an i
th
adjustment step of the filter from the filter coefficient of a preceding adjustment step i-l reduced by the correction value.
In other words, the objects are satisfied with an adaptive filter adjustment of the coefficients of the adaptive filter pair in that the n
th
filter coefficient of a filter in the i
th
adjustment step of the filter is formed by the filter coefficient of the previous adjustment step reduced by a correction value. The correction value is the product of the distortion elimination error, the delayed input value, and the manipulated variable &ggr;.
In accordance with an added feature of the invention, the manipulated variable &ggr; is the reciprocal value of a power of two, i.e., it is &ggr;=2
−&eegr;
.
In accordance with an additional feature of the invention, the manipulated variable &ggr; is controlled as a function of an instantaneous mean absolute error.
In accordance with another feat
Greenberg Laurence A.
Lerner Herbert L.
Mai Tan V.
Siemens Aktiengesellschaft
Stemer Werner H.
LandOfFree
Method for adaptive filter adjustment in a QAM/CAP 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 Method for adaptive filter adjustment in a QAM/CAP system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for adaptive filter adjustment in a QAM/CAP system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2851019