(12) United States Patent
`Hwang et al.
`
`(161616111616;
`(45) Date of Patent:
`
`US 6,590,893 B1
`Jul. 8, 2003
`
`US006590893B1
`
`(54) ADAPTIVE TRANSMISSION SYSTEM IN A
`NETWORK
`
`(75) Inventorsi Chien-Meen Hwallg, San Jose, CA
`(Us); Eugen Gersholl, San 105% CA
`(US); Maged F. Barsoum, Sunnyvale,
`EAWS); Hclglglnjgng?hall% H h
`upemno’
`(
`)>
`“01 ' “yn ’
`San Jose, CA (US); Fred Berkowitz,
`Palo Alto, CA (US); Bin Gu0, Fremont,
`CA (US)
`
`(73) Assignee: Legerity, Inc., Austin, TX (US)
`
`( * ) Notice:
`
`Subject to any disclaimer, the term of this
`patent is extended or adjusted under 35
`USC 154(b) by 0 days.
`
`(21) Appl NO . 09/286 997
`
`.
`
`..
`
`,
`
`5,426,643 A * 6/1995 Smolinske et a1. ....... .. 375/354
`5,477,550 A * 12/1995 Crisler et al. ............. .. 714/748
`5,666,383 A * 9/1997 Huang et al.
`375/219
`5,910,970 A * 6/1999 Lu . . . . . . . . . . . . . .
`. . . .. 375/222
`6,134,274 A * 10/2000 Sankaranarayanan et al.
`375/
`254
`6,157,612 A * 12/2000 Weerackody et al. ..... .. 370/215
`6,415,410 B1 * 7/2002 Kaneiva etal. .......... .. 714/749
`
`_
`_
`* cited by examiner
`
`Primary Examiner—Salvatore Cangialosi
`(74) Attorney, Agent, or Firm—McDermott, Will & Emery
`
`(57)
`
`ABSTRACT
`
`A network node con?gured for transmitting and receiving
`data to and from other netWork nodes is able to adapt the
`transmission rate based on the netWork conditions. The node
`initially transmits the data to a receiving node at a ?rst rate.
`If the data is not received error-free, the node is able to
`reduce the number of data bits of the current packet that are
`being transmitted and to increase the amount of redundant
`data. The node repeats the process until error-free transmis
`.
`.
`.
`sion is obtained.
`
`Apr‘ 7’ 1999
`(22) Filed:
`(51) Int. Cl.7 ................................................. .. H04J 3/06
`(52) US. Cl. ...................................... .. 370/354; 370/389
`(58) Field of Search ............................... .. 370/354, 403,
`370/543, 389, 506; 375/222, 254; 714/749
`
`(56)
`
`References Cited
`
`U.S. PATENT DOCUMENTS
`
`4,661,657 A * 4/1987 Grenzebach et al. ...... .. 375/359
`
`17 Claims, 5 Drawing Sheets
`
`5
`:
`'
`
`:
`:
`5
`i
`:
`
`—>
`_.
`DIFF.
`ENCDR+ IFFT
`
`‘9
`
`\
`14
`
`/
`12
`
`OUT
`-
`PARJ ~>AFE
`SER.
`
`/
`18
`
`1
`16
`
`DIFF.
`DECDR
`FFT —> 8‘
`SLCR
`
`I
`26
`
`\
`28
`
`v
`
`5
`:
`;
`:
`:
`E
`5
`E
`i
`
`CSCO-1013
`Cisco v. TQ Delta
`Page 1 of 12
`
`

`
`U.S. Patent
`
`Jul. 8, 2003
`
`5f01¢I.Ce.hS
`
`US 6,590,893 B1
`
`
`
`Pm?‘&OEntFGE
`
`Page 2 of 12
`
`

`
`U.S. Patent
`
`Jul. 8, 2003
`
`5f02¢I.Ce.hS
`
`US 6,590,893 B1
`
`Page 3 of 12
`
`

`
`U.S. Patent
`
`Jul. 8,2003
`
`Sheet 3 0f5
`
`US 6,590,893 B1
`
`3 .QE
`
`E58”.
`
`0 mew
`
`< O 0
`
`6.2 Q2
`
`mew
`
`Q1 .
`
`EEEEE
`
`mm .QE
`
`um .61
`
`om .GI
`
`3|
`
`CHANNEL
`
`at 2.:
`
`O O
`
`0 O
`
`O O O
`
`Page 4 of 12
`
`

`
`U.S. Patent
`
`Jul. 8,2003
`
`Sheet 4 0f 5
`
`US 6,590,893 B1
`
`TRANSMITTER
`CONTROLLER
`
`32: RECEIVER
`
`CONTROLLER
`34
`
`20\
`
`TRANSMITTER
`CONTROLLER
`
`I
`
`RECEIVER
`CONTROLLER
`44
`
`\CHANNEL 40
`
`FIG. 4
`
`Page 5 of 12
`
`

`
`U.S. Patent
`
`Jul. 8,2003
`
`Sheet 5 of5
`
`US 6,590,893 B1
`
`TRANSMITTER CONTROLLER 200
`ADDS CRC CODE TO END
`'4
`OF DATA PACKET
`I
`TRANSMITTER ENCO DES
`DATA PACKET
`
`202
`
`Y
`TRANSMIT TIME WORK,
`DATA AND CRC CODE
`
`204
`
`DATA
`RECEIVED
`AT DESTINATION
`WITHOUT ERRORS
`
`208
`DESTINATION
`NODE XMITTER
`CONTROLLER
`XMITS ACKNOW.
`SIGNAL
`
`210
`
`SIGNAL RECEIVED
`FROM DESTINATION
`NODE
`
`YES
`
`CURRENT
`PACKET TRANS.
`PREDETERM.
`NUMBER
`
`TRANSMITTER CONTROLLER 214
`TAKES PREDET. # OF BITS OF
`CURRENT PACK.
`I
`XMITTER CONT. ADDS CRC
`CODE AND REDUNDANT DATA
`
`216
`
`I
`TRANSMITTER ENCODES DATA 21a
`PACKET AND ADDS TIME MARK
`___—I
`
`FIG. 5
`
`Page 6 of 12
`
`

`
`US 6,590,893 B1
`
`1
`ADAPTIVE TRANSMISSION SYSTEM IN A
`NETWORK
`
`TECHNICAL FIELD
`
`The present invention relates to network communications
`and more particularly, to an adaptive transmission system
`used in a netWork.
`
`BACKGROUND ART
`Modern society continues to create exponentially increas
`ing demands for digital information and the communication
`of such information betWeen data devices. Local area net
`Works use a netWork, cable or other media to link stations on
`the netWork for exchange of information in the form of
`packets of digital data. A typical local area netWork archi
`tecture uses a media access control (MAC) enabling netWork
`interface cards at each station to share access to the media.
`Most conventional local area netWork architectures use
`media access controllers operating according to half-duplex
`or full-duplex Ethernet (ANSI/IEEE standard 802.3) proto
`col and a prescribed netWork medium, such as tWisted pair
`cable.
`These architectures have proven quite successful in pro
`viding data communications in commercial applications.
`HoWever, these common local area netWork architectures
`require installation of specialiZed Wiring and use of speci?c
`Wiring topologies. For example, the most popular netWork
`protocols, such as Ethernet, require special rules for the
`Wiring, for example With regard to quality of Wire, range of
`transmission and termination.
`Due to the success of the Internet and the rapid decreases
`in the prices of personal computers and associated data
`equipment, a demand has arisen for data communications
`betWeen a limited number of devices Within relatively small
`premises, typically a residence or small business. While
`existing local area netWorks can serve the purpose, in such
`installations, the cost of installing physical netWork Wiring
`satisfying the rules for the particular protocol can be pro
`hibitively expensive.
`Most existing buildings, including residences, include
`some existing Wiring, for phones, electrical poWer and the
`like. Proposals have been made to communicate data using
`such existing infrastructure. This reduces the costs of Wiring
`for the netWork, but the existing Wiring raises a variety of
`issues regarding transport of high-speed digital signals.
`For example, efforts are underWay to develop an archi
`tecture that enables computers to be linked together using
`conventional tWisted pair telephone lines. Such an
`arrangement, referred to herein as a home netWork
`environment, provides the advantage that existing telephone
`Wiring in a home may be used to implement a home netWork
`environment Without incurring costs for substantial neW
`Wiring installation. HoWever, any such netWork must deal
`With issues relating to the speci?c nature of in-home tele
`phone Wiring, such as operation over a media shared With
`other services Without interference from or interfering With
`the other services, irregular topology, and noise. With
`respect to the noise issue, every device on the telephone line
`may be a thermal noise source, and the Wiring may act much
`like an antenna to pick up disruptive radio signal noise.
`Telephone lines are inherently noisy due to spurious noise
`caused by electrical devices in the home, for example
`dimmer sWitches, transformers of home appliances, etc. In
`addition, the tWisted pair telephone lines suffer from turn-on
`transients due to on-hook and off-hook and noise pulses
`
`15
`
`25
`
`35
`
`45
`
`55
`
`65
`
`2
`from the standard telephones coupled to the lines, and
`electrical systems such as heating and air conditioning
`systems, etc.
`An additional problem in telephone Wiring netWorks is
`that the signal condition (i.e., shape) of a transmitted Wave
`form depends largely on the Wiring topology. Numerous
`branch connections in the tWisted pair telephone line
`medium, as Well as the different associated lengths of the
`branch connections, may cause multiple signal re?ections on
`a transmitted netWork signal. Telephone Wiring topology
`may cause the netWork signal from one netWork station to
`have a peak-to-peak voltage on the order of 10 to 20
`millivolts, Whereas netWork signals from another netWork
`station may have a value on the order of one to tWo volts.
`Hence, the amplitude and shape of a received pulse may be
`so distorted that recovery of a transmit clock or transmit data
`from the received pulse becomes substantially dif?cult.
`At the same time a number of XDSL technologies are
`being developed and are in early stages of deployment, for
`providing substantially higher rates of data communication
`over tWisted pair telephone Wiring of the telephone netWork.
`XDSL is used herein as a generic term for a group of
`higher-rate digital subscriber line communication schemes
`capable of utiliZing tWisted pair Wiring from an office or
`other terminal node of a telephone netWork to the subscriber
`premises. Examples under various stages of development
`include ADSL (Asymmetrical Digital Subscriber Line),
`HDSL (High data rate Digital Subscriber Line) and VDSL
`(Very high data rate Digital Subscriber Line).
`Consider ADSL as a representative example. For an
`ADSL-based service, the user’s telephone netWork carrier
`installs one ADSL modem unit at the netWork end of the
`user’s existing tWisted-pair copper telephone Wiring.
`Typically, this modem is installed in the serving central
`office or in the remote terminal of a digital loop carrier
`system. The user obtains a compatible ADSL modem and
`connects that modem to the customer premises end of the
`telephone Wiring. The user’s computer connects to the
`modem. The central office modem is sometimes referred to
`as an ADSL Terminal Unit—Central Office or ‘AT U-C’. The
`customer premises modem is sometimes referred to as an
`ADSL Terminal Unit—Remote or ‘ATU-R’. The ADSL
`user’s normal telephone equipment also connects to the line
`through a frequency combiner/splitter, Which is incorporated
`in the ATU-R. The normal telephone signals are split off at
`both ends of the line and processed in the normal manner.
`For digital data communication purposes, the ATU-C and
`ATU-R modem units create at least tWo logical channels in
`the frequency spectrum above that used for the normal
`telephone traf?c. One of these channels is a medium speed
`duplex channel and the other is a high-speed doWnstream
`only channel. TWo techniques are under development for
`dividing the usable bandWidth of the telephone line to
`provide these channels. One approach uses Echo Cancella
`tion. Currently, the most common approach is to divide the
`usable bandWidth of a tWisted Wire pair telephone line by
`frequency, that is to say by Frequency Division Multiplexing
`(FDM).
`FDM uses one frequency band for upstream data and
`another frequency band for doWnstream data. The doWn
`stream path is then divided by time division multiplexing
`into one or more high-speed channels and one or more loW
`speed channels. The upstream path also may be time
`division multiplexed into corresponding loW speed channels.
`The FDM data transport for ADSL services utiliZes dis
`crete multi-tone (DMT) technology. A DMT signal is basi
`
`Page 7 of 12
`
`

`
`US 6,590,893 B1
`
`10
`
`15
`
`4
`continue to reduce the number of bits sent and increase the
`amount of redundancy data until an error-free transmission
`occurs.
`According to one aspect of the invention a device is
`con?gured to transmit and receive data over a communica
`tions medium. The device includes a transmitter con?gured
`to transmit a ?rst packet comprising bits of data. The device
`also includes a receiver con?gured to receive an acknoWl
`edgement signal from a destination node indicating that the
`?rst packet Was received Without errors. The transmitter is
`further con?gured to transmit a second packet comprising a
`?rst plurality of portions, When the acknoWledgement signal
`is not received. The ?rst plurality of portions each include
`the same predetermined bits of the ?rst packet.
`Another aspect of the present invention provides a method
`of transmitting data from a netWork node. The method
`includes transmitting a ?rst packet comprising bits of data.
`The method also includes receiving an acknoWledgement
`signal from a destination node When the ?rst packet Was
`received Without errors. The method further includes trans
`mitting a second packet comprising a ?rst plurality of
`portions, When the acknoWledgement signal is not received
`Within a preset period of time. The plurality of portions each
`include the same predetermined bits of the ?rst packet.
`Other advantages and features of the present invention
`Will become readily apparent to those skilled in this art from
`the folloWing detailed description. The embodiments shoWn
`and described provide illustration of the best mode contem
`plated for carrying out the invention. The invention is
`capable of modi?cations in various obvious respects, all
`Without departing from the invention. Accordingly, the
`draWings are to be regarded as illustrative in nature, and not
`as restrictive.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`Reference is made to the attached draWings, Wherein
`elements having the same reference numeral designations
`represent like elements throughout.
`FIG. 1 is a block diagram of a conventional transmitter
`and receiver using discrete multi-tone technology.
`FIG. 2 is a block diagram illustrating a transmitter and
`receiver utiliZing DMT according to an embodiment of the
`present invention.
`FIGS. 3a—3a' schematically illustrate the constellation
`points associated With the transmission and reception of data
`using the transmitter/receiver of FIG. 2, according to an
`embodiment of the present invention.
`FIG. 4 is a block diagram of a pair of nodes used in a
`netWork in accordance With an embodiment of the present
`invention.
`FIG. 5 is a How diagram illustrating the method for
`transmitting data according to an embodiment of the present
`invention.
`
`3
`cally the sum of N independently quadrature amplitude
`modulated (QAM) signals, each carried over a distinct
`carrier frequency channel. The frequency separation
`betWeen consecutive carriers is 4.3125 KHZ With a total
`number of 256 carriers or tones (ANSI). An asymmetrical
`implementation of this 256 tone-carrier DMT coding
`scheme might use tones 32—255 to provide a doWnstream
`channel of approximately 1 MHZ analog bandWidth. In such
`an implementation, tones 8—31 are used as carriers to
`provide an upstream channel of approximately 100 kHZ
`analog bandWidth. Each tone is QAM to carry up to 15 bits
`of data on each cycle of the tone Waveform (symbol). An
`example of a conventional DMT-based system is illustrated
`in FIG. 1.
`The existing DSL systems provide effective high-speed
`data communications over tWisted pair Wiring betWeen
`customer premises and corresponding netWork-side units,
`for example located at a central of?ce of the telephone
`netWork. The DSL modem units overcome many of the
`problems involved in data communication over tWisted pair
`Wiring. HoWever, for a number of reasons, the existing DSL
`units are not suitable to providing local area netWork type
`communications Within a customer’s premises. For
`example, existing ADSL units are designed for point-to
`point communication. That is to say, one ATU-R at the
`residence communicates With one ATU-C unit on the net
`Work end of the customer’s line. There is no Way to use the
`units for multi-point communications. Also, the existing
`ADSL modems tend to be quite complex, and therefore are
`too expensive for in-home communications betWeen mul
`tiple data devices of one customer.
`As described above, multi-point netWorks using conven
`tional technology are not suitable for in-home use.
`Additionally, even conventional multi-point netWorks
`requiring specialiZed Wiring and having predetermined
`topologies often suffer from poor signal quality betWeen tWo
`or more nodes in the netWork.
`For example, the medium connecting tWo particular nodes
`may be of poor quality resulting in drastic signal attenuation
`and phase distortion. The attenuation and distortion often
`lead to data errors When transmitting the data over such a
`medium. Prior art systems often retransmit the data When
`errors occur. HoWever, When the errors are caused by the
`communications medium or the netWork layout, simply
`retransmitting the data often results in another erroneous
`transmission.
`
`25
`
`35
`
`45
`
`SUMMARY OF THE INVENTION
`
`There is a need for an arrangement that provides an
`adaptive data transmission system for use in a netWork.
`There is also a need for an arrangement that provides an
`adaptive data transmission system for use in a netWork
`employing discrete multi-tone technology.
`These and other needs are met by the present invention,
`Where a data transmission device used in a netWork node
`includes a transmitter portion and a receiver portion. The
`transmitter transmits a data packet to a receiving node.
`When the data is received Without errors, the receiving node
`transmits an acknowledgement signal to the transmitting
`node and the transmitting node is ready to transmit the next
`packet. HoWever, When an error in transmission occurs, the
`transmitting node is able to retransmit a portion of the data,
`along With at least one redundant copy of the portion. If at
`least one of the redundant data portions is received Without
`errors, an acknowledgement is sent back to the transmitting
`node. If errors still occur, the transmitting node is able to
`
`55
`
`BEST MODE FOR CARRYING OUT THE
`INVENTION
`The present invention Will be described With the example
`of a netWork node in a multi-point netWork using discrete
`multi-tone (DMT) technology. A description Will ?rst be
`given of an exemplary DMT-based netWork, folloWed by the
`arrangement for providing an adaptive transmission system.
`It Will become apparent, hoWever, that the present invention
`is also applicable to other types of netWorks.
`
`65
`
`NETWORK ARCHITECTURE OVERVIEW
`FIG. 2 illustrates an exemplary system in Which the
`present invention may be advantageously employed. Net
`
`Page 8 of 12
`
`

`
`US 6,590,893 B1
`
`5
`work nodes 10 and 20 are nodes, e.g., personal computers,
`in computer network 100. Each node is capable of trans
`mitting and receiving data over channel 40. Channel 40 may
`be a twisted pair telephone line or another medium used to
`transmit data. In FIG. 2, a detailed transmitter 11 is shown
`in network node 10 and a detailed receiver 21 is shown in
`network node 20. It should be recognized, however, that
`each node 10 and 20 includes both the transmitter and
`receiver circuitry. Additionally, although not shown, other
`network nodes may be connected to nodes 10 and 20 in a
`ring topology, star topology or any other network topology.
`According to the exemplary embodiment of the invention,
`network 100 in FIG. 2 utiliZes DMT-based technology to
`transmit data over channel 40. The present invention how
`ever departs from conventional DMT technology by utiliZ
`ing a differential coder 12 to encode an input bit stream into
`a predetermined number of tones. According to the exem
`plary embodiment, differential coder 12 uses 256 tones to
`encode the input bit stream. In alternative con?gurations,
`differential coder 12 may utiliZe other numbers of tones to
`encode the bit stream, based on the particular network
`requirements. Additionally, as described previously, in a
`DMT-based system utiliZing 256 tones, each tone is capable
`of transmitting up to 15 bits of data on the tone waveform.
`According to the exemplary embodiment of the present
`invention, each tone is used to transmit two bits of data,
`which corresponds to four constellation points. However, in
`alternative con?gurations the present invention may trans
`mit other numbers of bits per tone.
`The differential coder 12 overcomes the drawbacks asso
`ciated with a home network environment by utiliZing a
`reference tone encoded with a reference bit pattern, before
`transmitting the input data stream. In alternative
`con?gurations, a sequence of reference tones, e.g., up to 256
`tones, may be encoded with a reference bit pattern. Trans
`mitter 11 modulates the reference tone(s) to carry the
`predetermined bit pattern over channel 40. For example,
`assume that bit pattern “00” is the predetermined bit pattern.
`The reference tone is then quadrature amplitude modulated
`to carry bit pattern “00”. After processing by differential
`coder 12, the bit pattern “00” is represented by constellation
`point A in the complex plane shown in FIG. 3a. The
`constellation point represents the amplitude and phase of the
`signal at that particular tone.
`The Inverse Fast Fourier Transform (IFFT) block 14,
`receives the tone information and converts the frequency
`domain-based tone information into a time domain-based
`waveform and outputs the time domain waveform to
`parallel-to-serial converter 16. A guard-band cyclic pre?x
`may be applied between the IFFT block 14 and the parallel
`to-serial converter 16 to transmit a pre?x before the actual
`reference bits. The pre?x data is discarded at the receiver 21,
`thereby eliminating the effects of intersymbol interference
`(ISI) associated with the start of the transmission. Parallel
`to-serial converter 16 converts the data to a serial format for
`analog front end
`block 18. AFE block 18 then
`transmits the data over channel 40 to node 20.
`At node 20, AFE block 22 receives the line signal and
`performs ampli?cation, ?ltering and digitiZing. The signal is
`then fed into serial-to-parallel converter 24. After the data is
`converted to a parallel format, FFT block 26 computes the
`amplitude and phase information of the reference tone.
`Differential decoder and slicer 28 then decodes the ref
`erence signal, represented by constellation point B in FIG.
`3b. Next, the transmitter 11 transmits the actual input bit
`stream representing the data from network node 10 destined
`
`6
`for network node 20. For example, assume the current bit
`stream being transmitted from node 10 to node 20 is bit
`pattern “01”, which is illustrated by constellation point C in
`FIG. 3c.
`The receiver 20 receives the encoded carrier tone and
`processes the signal information using AF E block 22, serial
`to-parallel converter 24, FFT block 26 and differential
`decoder and slicer block 28. After processing, the tone
`information is represented by constellation point D in FIG.
`3d. The receiver 21 includes logic to compare the received
`point D with the previous received point B to determine the
`phase relationship between the points. Referring to FIG. 3d,
`the phase relationship between points B and D is 90 degrees,
`with point D leading point B. Using this information, the
`decoder and slicer block 28 assumes that the transmitted
`data bits encoded via the second carrier tone is 90 degrees
`out of phase with the reference data bits. In this example, the
`reference data bit pattern was “00”, and a bit pattern leading
`“00” by 90 degrees maps to bit pattern “01”, as shown in
`FIG. 3a. Therefore, the decoder and slicer block 28 in this
`example determines that the data bits transmitted with the
`second tone correspond to “01”.
`In the manner described above, the transmitting node 10
`uses a reference tone encoded with a reference bit pattern at
`the beginning of every data packet. The phase distortion
`between the reference tone and the subsequent tone is then
`used to determine the value of the data associated with the
`?rst tone. Each successive tone is processed in a similar
`manner by comparing the phase relationship between the
`constellation point associated with the tone with the previ
`ous constellation point. By assuming that every constellation
`point will have the same degree of amplitude attenuation and
`phase distortion, the present invention is then able to deter
`mine the value of the data transmitted on each tone without
`knowing the particular channel characteristics.
`Advantageously, the present invention is able to decode
`transmitted data without the use of an equaliZer to reverse
`the effect of amplitude attenuation or phase distortion asso
`ciated with poor quality wiring.
`In the exemplary embodiment, since a known reference
`bit pattern is transmitted before each packet of data as a new
`reference, any receiving node is able to determine the values
`of the subsequent received data bits in that packet.
`Advantageously, the system for transmitting/receiving data
`of the present invention is usable for different paths employ
`ing different channels in network 100.
`The present invention, as described above, may be advan
`tageously employed in a multi-point network. In such
`networks, a sending signal is transmitted only when there is
`data that needs to be transmitted. In accordance with an
`embodiment of the present invention, the transmitting node
`transmits a predetermined time mark to identify the begin
`ning of a packet. In the exemplary embodiment, the time
`mark consists of one cycle of a sinusoidal waveform. The
`receiver node then matches received patterns with the pre
`de?ned time mark pattern to identify the start of the packet,
`and thus begin decoding the received packet. In alternative
`con?gurations, the time mark may be several cycles of a
`sinusoidal waveform or any other predetermined waveform.
`Additionally, a predetermined transmitting node identi?er
`may also be transmitted after the time mark for identi?cation
`purposes. An exemplary node ID may consist of pulse
`amplitude modulated (PAM) sinusoidal waveforms unique
`to each network node.
`
`15
`
`25
`
`35
`
`45
`
`55
`
`65
`
`ADAPTIVE TRANSMISSION
`As described above, the network nodes in network 100 are
`able to transmit and receive data over channel 40 without
`
`Page 9 of 12
`
`

`
`US 6,590,893 B1
`
`7
`knowing the particular channel characteristics. The present
`invention is also able to adapt the transmission rate based on
`the particular channel characteristics. For example, if the
`channel characteristics of channel 40 are poor and the
`receiving node is unable to receive the transmitted data
`Without errors, the transmitting node is able to adapt the
`transmission rate to ensure that error-free data is received.
`FIG. 4 is a block diagram of netWork nodes 10 and 20, in
`accordance With an embodiment of the present invention.
`Network node 10 includes transmitter controller 32 and
`receiver controller 34 and netWork node 20 includes trans
`mitter controller 42 and receiver controller 44. These
`transmitter/receiver controllers of nodes 10 and 11, as
`described in detail beloW, enable the present invention to
`optimiZe the data transmission rate, based on the particular
`operating conditions of channel 40. The transmitter and
`receiver circuitry of FIG. 2 are not depicted in nodes 10 and
`20 of FIG. 4 in order not to unduly obscure the thrust of the
`present invention.
`FIG. 5 is a How diagram illustrating the method for
`providing an adaptive transmission system in accordance
`With an embodiment of the present invention. At step 200,
`transmitter controller 32 appends a cyclic redundancy check
`(CRC) code at the end of a packet of data to be transmitted
`over channel 40. For an Ethernet packet, the CRC code is
`included With the packet and therefore step 200 is skipped.
`In alternative con?gurations, other error check codes may
`also be utiliZed.
`Node 10 then encodes the data packet and CRC code at
`step 202. Next, at step 204, node 10 transmits the predeter
`mined time mark along With the data stream and CRC code
`over channel 40. According to the exemplary embodiment of
`the invention, the time mark is a PAM sinusoidal signal one
`period in duration, optionally folloWed by node ID infor
`mation. The time mark information is transmitted at the start
`of each packet of data. The time marks essentially acts as a
`signal to alert the receiving node 20 that the data stream
`folloWs.
`When node 20 receives the data step 206, the receiver
`controller 44, illustrated in FIG. 4, checks the received data
`and CRC code to determine Whether the data Was received
`Without errors. If the receiver controller 44 determines that
`the data Was received Without errors, the transmitter con
`troller 42 of node 20 transmits the predetermined time mark
`back to the transmitting node 10 as an acknoWledgement
`signal, at step 208. Alternatively, any other prede?ned
`acknoWledgement signal may be used. When the receiving
`node 20 determines that the data Was received With errors at
`step 206, the receiving node does not transmit the acknoWl
`edgement signal.
`Next, at step 210, the receiver controller 34 of transmit
`ting node 10 Waits a predetermined period of time for the
`acknoWledgment time mark from node 20. When the
`acknoWledgement time mark is received in the predeter
`mined period of time at step 210, the process returns to step
`200 for the transmission of a neW packet. If the acknoWl
`edgment time mark is not received in the predetermined
`period of time, at step 210, the transmitting node 10 assumes
`that the data packet Was received With errors or Was lost.
`Next, at step 212, the transmitter controller 32 of node 10
`determines Whether the current packet has been transmitted
`a predetermined number of times. According to the exem
`plary embodiment of the invention, the predetermined num
`ber is three. HoWever, in alternative con?gurations, the
`predetermined may be any other number including one. If
`the current packet has not been transmitted the predeter
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`8
`mined number of times, the process returns to step 204
`Where the current packet along With the time mark is
`retransmitted.
`Next, assume that the current packet has been retransmit
`ted the predetermined number of times Without an acknoWl
`edgement signal from receiving node 20. The transmitter
`controller 32 of node 10 then reduces the effective data rate
`by transmitting only a portion of the current data packet,
`along With at least one redundant copy of the portion. More
`speci?cally, at step 214, the transmitter controller 32
`retrieves a ?rst predetermined number of bits at the head of
`the current data packet. For example, suppose the original
`data packet Was 8 bits in length consisting of 11010001. At
`step 214, the transmitter controller retrieves only the prede
`termined number of bits from the beginning of this packet.
`Further assume in this example that the predetermined
`number of bits is four. In this situation, the transmitter
`controller 32 retrieves bits “1101”, i.e., the ?rst four bits of
`the current packet.
`Next at step 216, the transmitter controller 32 appends the
`appropriate CRC code to the end of the neW bit pattern. The
`transmitter controller 32 also provides at least one redundant
`copy of the neW bit pattern and CRC code, after the ?rst
`CRC code. In the example described above, and assuming
`that the number of redundant copies is one, the transmitter
`controller 32 Would provide the folloWing bit pattern for
`encoding: 1101(CRCcode)1101(CRCcode). HoWever, in
`alternative embodiments, the number of redundant copies of
`the ?rst bit pattern Would generally be limited only by the
`bandWidth of the channel.
`Next, the transmitting node 10 encodes the neW data
`packet, at step 218, and transmitter controller 32 adds the
`time mark at the beginning of the packet to indicate the start
`of the packet. The process then returns to step 204.
`The processes at steps 204—210 are the same as described
`previously With an exception at step 206. At step 206, the
`receiver controller 44 determines Whether any one of the
`redundant data patterns in the packet has been received
`Without errors. For example, assume that the receiver con
`troller 44 determines that the second group of bits, 1101 in
`the example described above, Was received Without errors,
`based on the CRC code information. The receiving node 20
`at step 208 Would then transmits an acknoWledgement time
`mark to transmitting node 10. Advantageously, a single
`reception of error-free data associated With any one of the
`redundant patterns enables the present invention to be more
`robust and to operate in conditions Where conventional
`systems are unable to operate.
`The processes at steps 212—218 continue until the receiver
`node 20 receives a group of data bits Without errors and
`sends the acknoWledgement signal to the transmitting node
`10. Each time the process reaches step 214, the transmitter
`controller 32 reduces the number of bits of the current packet
`being transmitted and increases the number of redundant bit
`patterns.
`According to the exemplary embodiment of the invention
`illustrated in connection With FIG. 2, assume that the ?rst
`packet utiliZes 256 carrier tones to transmit 512 bits of data
`as a data packet. The ?rst time the process reaches steps
`214—218, the transmitting node 10 may transmit the ?rst 256
`bits of the 512 bits. In this case, the ?rst 128 tones are QAM
`to carry the ?rst 256 bits. The next 128 adjacent to