Pulse or digital communications – Testing – Signal noise
Reexamination Certificate
1999-02-18
2003-02-04
Pham, Chi (Department: 2631)
Pulse or digital communications
Testing
Signal noise
C379S296000
Reexamination Certificate
active
06516027
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to data communications, specifically to an apparatus and method for allocating bits among carrier tone subchannels (bins) in a discrete multitone modulation (DMT) communication system. The present invention allows DMT communication to proceed at a data rate via communication equipment which is less complicated, less costly, and more efficient than prior DMT communication devices.
Data transmission modulation schemes are broken into many categories, the two major of which are single carrier modulation and multicarrier modulation. Single carrier modulation employs the use of a single frequency upon which data is modulated. For example, a single carrier modem might use a 2 KHz carrier and quadrature amplitude modulation (QAM) or phase shift keying (PSK) modulation to “superimpose” the data on the carrier. Although used in many applications and environments, single carrier modulation modems, such as V.34 modems, are often used on voiceband systems such as POTS lines (plain old telephone system). Further, bit allocation among carrier tones is not an issue in single carrier modulation systems since only one carrier tone is used. As such, the invention is not directed to single carrier systems.
The other category of modulation transmission techniques is multicarrier modulation. In multicarrier modulation communications, the transmission channel is partitioned into a number of narrowband subchannels. These narrowband transmission subchannels typically correspond to a series of contiguous frequency bands in which each subchannel uses a distinct carrier frequency. This distinct carrier frequency is also referred to as a “tone”, and the subchannel is referred to as a “bin”. The capacity of the entire transmission channel is the sum of the capacities of each individual subchannel. The method for assigning data information and energy to each of the subchannels in a multichannel transmission environment is called “loading”. The distribution of data information among the channels is referred to as “bit allocation”.
Multicarrier modulation techniques for data communication are not new, and have been known and used for the past 50 years. However, prior to the proliferation of digital communication technologies and efficient methods for digital signal processing, multicarrier modulated communication systems often relied on a group of modems with differing carrier frequencies whose signals were added together to form a composite signal. As shown in
FIG. 1
, a data stream would enter a serial to parallel signal splitter
2
whose output would feed a plurality of modems
4
. The modems would each be set to modulate at a different carrier frequency. The analog output of each modem would be added together and transmitted along the communication channel
8
. The composite signal would be separated and demodulated at the receiving end via a complementary set of equipment.
This technique required multiple modems, was not reliable and did not take into account the different operating and performance characteristics of each individual carrier frequency. This resulted in an inefficient use of the communication channel resources. For example, in a typical communication environment, signals with higher frequencies attenuate quicker and are more subject to noise than signals at low frequencies. Thus, it was possible that a subchannel with a high signal-to-noise ratio could be under utilized, while a subchannel with a low signal-to-noise ratio could be over utilized resulting in an unnecessary degradation in overall transmission channel performance.
Subsequent DMT multicarrier modulation equipment made use of digital signal processing techniques including Fast Fourier Transforms and Inverse Fast Fourier Transforms. Digital signal processing allowed a single DMT communication device to be used to modulate all subchannels, thereby improving reliability and lowering the cost of communications. These devices, however, were still constrained by the limitations of their predecessors, namely, inefficient use of the communication channel resources.
A preferred approach is to load each subchannel based on the individual transmission characteristics of that subchannel. Better subchannels, should carry more information than poorer quality subchannels. This allows an efficient use of the communication channel resources.
Transmission channels are typically characterized by the channel's margin, signal-to-noise ratio gap (hereinafter SNR gap), and capacity. All are related concepts. The margin is the amount of additional signal-to-noise ratio in excess of the minimum required to achieve a given performance level for a particular type of modulation scheme with a particular SNR gap. The SNR gap is a function of a chosen probability of transmission error and the modulation and coding techniques. The SNR gap measures the inefficiency of the transmission method with respect to the best possible performance, assuming an additive white Gaussian noise channel. The SNR gap is often constant over a wide range of transmission rates which may be transmitted by the particular modulation coding technique. The channel capacity refers to the maximum data rate capable of being transmitted on a particular channel. The optimum line coding technique has a SNR gap of zero dB. Although such an optimum line code requires infinite decoding/encoding delay and is infinitely complex, it has become practical at typical Digital Subscriber Line (DSL) speeds to implement modulation methods to achieve SNR gaps as low as 1-2 dB. Therefore, one factor to be considered during the bit allocation process is the transmission quality of each subchannel, in order to maximize the bit allocation for each subchannel.
Another factor to be considered during the bit allocation process is gain scaling. Gain scaling allows an individual subchannel to be amplified or attenuated based on a set of parameters. Certain DSL technologies and standards such as the Asymmetric Digital Subscriber Line (ADSL) T1.413 standard specify a scaling range, such as ±2.5 dB, subject to a maximum total energy level for the entire channel. However, these standards do not specify how to implement the gain scaling factor.
Bit allocation in a discrete multitone transmission environment is a well-studied problem, relating to the development of a number of different bit allocation algorithms. Optimum bit allocation can be achieved using the well-known “water-filling” algorithm. However, there are two major drawbacks with “water-filling”. First, the resulting calculated bit distribution is not necessarily a set of integers. This may result in complications during implementation. Second, the “water-filling” algorithm maximizes the aggregate bit rate for a given margin and bit error rate. However, maximizing the margin at a given bit error rate and a given bit rate is desirable in many applications.
The loading goal of bit-loading is generally classified as either rate-adaptive loading or margin-adaptive loading. In rate-adaptive loading, the data rate is maximized subject to the constraint of fixed total power. In margin-adaptive loading, the goal is to maximize the margin at a fixed data rate.
There are at least two popular strategies employed to try to achieve these goals. The first, Chow's algorithms, suggests an on/off energy distribution, and only approximately solve the problems associated with achieving the rate-adaptive and margin-adaptive loading goals. The second category is Greedy algorithms. This well-known class of algorithms can solve the loading problems exactly. For example, the Hughes-Hartog and Campello algorithms belong to the family of Greedy algorithms. Greedy algorithms work to allocate bits to the subchannels in an incremental way such that a subchannel is always chosen for the next bit to be allocated according to the least cost in energy. In other words, the best subchannel gets the first bit, and a table of incremental next energies for the next bit must be computed and the process continued until
Choudhary Prashant
Kapoor Samir
NEC USA Inc.
Ostrolenk Faber Gerb & Soffen, LLP
Pham Chi
Tran Khai
LandOfFree
Method and apparatus for discrete multitone communication... 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 apparatus for discrete multitone communication..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for discrete multitone communication... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3115688