Pulse or digital communications – Systems using alternating or pulsating current – Plural channels for transmission of a single pulse train
Reexamination Certificate
2000-02-22
2001-04-24
Chin, Stephen (Department: 2734)
Pulse or digital communications
Systems using alternating or pulsating current
Plural channels for transmission of a single pulse train
C375S222000, C375S225000
Reexamination Certificate
active
06222888
ABSTRACT:
The invention relates generally to a high-speed rate adaptive communications system which utilizes an improved combination of bit/energy initializations and bit fine tunings effected on a data transmission signal consisting of a multitude of sub-channels. The present invention has specific applicability to Asymmetric Digital Subscriber Loops (ADSL) and similar environments.
BACKGROUND OF THE INVENTION
Remote access and retrieval of data and information are becoming more desirable and common in both consumer and business environments. As data and information transfer is becoming more and more voluminous and complex, using traditional data links such as voice-band modems is too slow in speed. For example, the use of the Internet to locate and access information is increasing daily, but the retrieval of typical graphics, video, audio, and other complex data forms is generally unsatisfactorily slow using conventional voice-band modems. In fact, the slow rate of existing dial-up analog modems frustrates users, and commerce and interaction using the Internet would have been even higher were it not for the unacceptable delays associated with present day access technology. The ability to provide such desired services as video on demand, television (including HDTV), video catalogs, remote CD-ROM's, high-speed LAN access, electronic library viewing, etc., are similarly impeded by the lack of high speed connections.
Since copper lines are widely available and developed, solutions to the high speed access problem have been focused on improving the performance of voice-band modems, which operate at the subscriber premises end over a 3 kHz voice band and transmit signals through the public switching telephone network (PSTN). The phone company network treats them exactly like voice signals. These modems presently transmit up to 33.6 kb/s over 2-wire telephone line, even though the practical speed was 1.2 kb/s only twenty years ago. The improvement in voice band-modems over the past years has resulted from significant advances in algorithms, digital signal processing, and semiconductor technology. Because such modems are limited to voice bandwidth (3.0 kHz), the rate is bound by the Shannon limit, around 30 kb/s. A V.34 modem, for example, achieves 10 bits/Hz, a figure that approaches the theoretical Shannon limits. There is a considerable amount of bandwidth available in copper lines, however, that has gone unused by voice-band modems, and this is why a proposal known as Asymmetric Digital Subscriber Loop (ADSL) was suggested in the industry as a high-speed protocol/connection alternative. The practical limits on data rate in conventional telephone line lengths (of 24 gauge twisted pair) vary from 1.544 Mb/s for an 18,000 foot connection, to 51.840 Mb/s for a 1,000 foot connection. Since a large proportion of current telephone subscribers fall within the 18,000 foot coverage range, ADSL can make the current copper wire act like a much “bigger pipe” for sending computer bits and digital information (like movies and TV channels), while still carrying the voice traffic. For example, an ADSL modem can carry information 200 times faster than the typical voice band modem used today.
ADSL is “asymmetric” in that more data goes downstream (to the subscriber) than upstream (back from the subscriber). The reason for this is a combination of cost, cross-talk, speed demands, and performance. For example, twisted pair wiring coupling increases with the frequency of the signal. If symmetric signals in many pairs are used within a cable, the data rate and line lengths become significantly limited by the coupling noise. Since the preponderance of target applications for digital subscriber services is asymmetric, asymmetric bit rate is not perceived to be a serious limitation at this time. Therefore, the ADSL standard proposes up to 6 Mb/s for downstream, and up to 640 kb/s for upstream. For example, video on demand, home shopping, Internet access, remote LAN access, multimedia access, and specialized PC services all feature high data rate demands downstream, to the subscriber, but relatively low data rates demands upstream. The principal advantage is that all of the high speed data operations take place in a frequency band above the voice band, leaving Plain Old Telephone Service (POTS) service independent and undisturbed, even if an ADSL modem fails. ADSL further provides an economical solution for transmission of high bandwidth information over existing copper line infrastructures.
Specifically, the T1E1.413 ADSL standard divides the available transmission bandwidth into two parts. At the lower 4 kHz band, ordinary (POTS) is provided. The bulk of the rest bandwidth in the range from 4 kHz to about 1 MHz is for data transmission in the downstream direction, which is defined to be from the exchange to the subscriber. The upstream control channel uses a 160 kHz band in between. The signals in each channel can be extracted with an appropriate band-pass filter.
According to the T1E1.413 ADSL standard, a line code scheme called Discrete Multi-one (DMT) is used, which divides the spectrum from 4 kHz to 1.1 MHz into 255 4.3125 kHz channels. Each channel uses Quadrature Amplitude Modulation (QAM) to carry 2 to 15 bits/QAM symbol. This results essentially in overall performance which is equivalent to around two hundred V.34 modems used in parallel on the same line. Because each channel can be configured to a different bit rate according to the channel characteristics, it can be seen that DMT is inherently “rate-adaptive” and extremely flexible for interfacing with different subscriber equipment and line conditions. In typical DMT implementations, such as shown in U.S. Pat. No. 5,479,447 to Chow et. al., transmission power to the individual channels is initially configured based on the noise power and transmission loss in each band. In this way, channels with less noise and attenuation can carry larger amounts of data, while poorer sub-channels can be configured to carry fewer bits and can even be shut down entirely. U.S. Pat. No. 5,596,604 to Cioffi et. al. shows that it is known to store relevant information for each DMT channel in a so called Bit & Energy Table. It is further known (U.S. Pat. No. 5,400,322 to Hunt et. al.) that line conditions can vary after initialization because of temperature fluctuations, interference, etc., and this can affect both the error rate and maximum data throughput. By measuring the quality of each sub-channel on an ongoing basis, an “updated” Bit & Energy Table is maintained to adaptively configure the system for maximum data throughput or error performance. In normal applications, if the quality of any particular channel degrades to the point where the error performance of the system is compromised, one or more bits on that sub-channel are automatically moved to a sub-channel that can support additional bits.
To maximize the overall bit rate from the given channel characteristics and to achieve the target bit error rate (BER), a DMT codec needs to determine the number of bits for each QAM channel and allocate the output power within the standard's output power spectrum requirements. Since the channel characteristics may vary as a function of time, the results of this bit allocation and power allocation process need to be updated constantly. The process of determining and dynamically adapting the optimal bit capacity output power for each sub-channel is generally referred to in the art as bit/energy loading and bit fine tuning respectively.
A good combined energy loading and bit fine tuning algorithm can optimize the system performance, as measured by considerations such as the overall data throughput rate, power margin, bit-error rate, and compliance with the T1E1.413 ADSL standard. Since there is a complex and nonlinear dependence between the power margin, bit allocation. and energy loading for each sub-channel, the computation may not necessarily converge. Therefore, it is critical for the energy loading and bit-fine tuning to rapidly converge and achieve the optimum
Chen Chunta
Kao Chiihsin
Liu Ming-Kang
Chin Stephen
Ghayour Mohammad
Integrated Telecom Express, Inc.
Lawt
LandOfFree
Method and circuit for controlling setup of multichannel 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 and circuit for controlling setup of multichannel system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and circuit for controlling setup of multichannel system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2517866