Method and apparatus for coordinating multi-point-to-point communications in a multi-tone data transmission system

BACKGROUND OF THE INVENTION

The present invention relates generally to a discrete multi-tone communication system in which a central unit services a plurality of remote units. In particular, it relates to methods of coordinating uplinks from the remote units.

Examples of prior art communication systems are given in WO-A-92 10884 and EP-A-0 29S 227.

It has been shown that discrete multi-tone data transmission schemes (DMT data transmission schemes) facilitate high-performance data transmission. Among the advantages of DMT architectures is that they have high spectral efficiencies and can adaptively avoid various signal distortion and noise problems. Because of their very high data transfer capabilities, the selection of a DMT data transfer scheme in most applications offers much room for extension of the service as the demands on the data transfer system increase. Therefore, discrete multi-tone technology has applications in a variety of data communications environments. For example, the Alliance For Telecommunications Information Solutions (ATIS), which is an accredited group of the ANSI (American National Standard Institute) standard group, has created a discrete multi-tone digital standard transmission standard for asymmetric digital subscriber line (ADSL) digital data transmission. The standard is intended primarily for the transmission of video data over ordinary telephone lines, although it can also be used in a variety of other applications. The North American standard is referred to as the ANSI T1.413 ADSL standard.

The transmission rates in the ADSL standard are intended to allow transmission of information at rates of at least 6 million bits per second (ie, at least 6 Mbps) over twisted pair telephone lines. The standard discrete multitone (DMT) system uses 256 "tones" or "subchannels" that are 4.3125 kHz wide in the forward (downlink) directions. In the context of a telephone system, the downlink direction is generally considered to be the transmission from the central office (which is typically owned by the telephone company) to a remote location that may be an end user (ie, a home or a company user). In other systems, the number of tones used may vary widely. However, IFFT modulation occurs, typical values ​​for the number of available subchannels (tones) are integer patencies of two, and for example 128, 256, 512, 1024 or 2048 subchannels result.

The standard of asymmetric digital subscriber lines also contemplates the use of a return signal at a data rate in the range of 16 to 800 kbps. The return signal corresponds to the transmission in the upward direction, for example, from the remote location to the central office. Accordingly, the term "asymmetric digital subscriber line" results from the fact that the data transmission rate in the forward direction is considerably higher than in the opposite direction. This is particularly useful in systems designed to transmit video program or videoconferencing information over the telephone lines to a remote location. For example, one possible application of the systems allows home users to receive video information, such as movies, over telephone lines, or over cables instead of lending video cassettes. Another possible application is video conferencing.

The discrete multitone (DMT) transmission scheme has the potential to be used in applications that go well beyond data transmission over telephone lines. In fact, DMT can also be used in a variety of other digital subscriber access systems. It may be used, for example, in cable-based subscriber systems (which typically use coaxial cable) and wireless subscriber systems such as digital mobile television. In cable systems, typically, a single central processing unit (central modem) is used to distribute digital signals to more than one customer, meaning that more than one remote unit (remote modem) occurs. Although all remote modems can reliably receive the same digital signals, the uplink transmissions must be coordinated to avoid confusion with the central modem about the source of the uplink signals. In some existing cable systems (which do not use discrete multitone transmission schemes), each remote unit is given a dedicated frequency band over which it must communicate with the central station. However, this method inherently introduces an inefficient use of the transmission bandwidth, typically requiring the use of analog filters to isolate the transmissions from the various remote units. In other existing cable systems, a single wideband is used by all of the remote units that use Time Division Multiple Access (TDMA) to access the upstream channel. This method is ineffective because of the lower total capacity of the single channel and the time it takes to access. Fixed Digital Mobile Broadcasting Systems Have Similar Obstacles By having the ability to access the channel on a time-multiplexed or frequency-multiple basis, the transmission channel would be used more effectively. The inherent multiplexing nature of DMT transmission has earlier limited its application to point-to-point transmission because the transmissions from different sources must be synchronized for full digital multiplexing to function properly.

In ADSL applications, there is the possibility that a similar problem will occur, although its nature is typically more limited. In particular, a single line may service multiple end points at a particular billing address (which may typically be a home or office). That is, there may be several "telephone jacks" over which the user may wish to receive signals. To enable multiple sites (jacks) to be operated over a single line, it has been suggested to use a master modem to enable synchronization. However, this is considered a relatively costly and undesirable solution. It would therefore be desirable to provide a mechanism in discrete multi-tone data transmission systems that enables the synchronization of signals from multiple remote units so that a central unit can coordinate and reliably interpret signals sent from the remote units.

Another feature of current transmission systems used for communication from a remote unit to a central unit is that they either transmit data at a fixed maximum rate (frequency division multiplexing) or transmit data in packets of a particular size (time division multiplexing). They do not allow both. This limits the effectiveness of using the transmission channels. It would accordingly be desirable to provide a mechanism by which a remote unit, if required, can specify a desire to transmit at a particular data rate and by which the remote unit, if the data rate is not a problem, can indicate that she wants to transmit a fixed amount of information.

SUMMARY OF THE INVENTION

In order to achieve the aforementioned and other objects, and in accordance with the purpose of the present invention, a number of bidirectional data transmission systems are disclosed which enable communication between multiple remote units and a central unit using a frame-based multi-carrier discrete transmission scheme. In each of the systems, frames transmitted from the plurality of remote units are synchronized at the central unit. A variety of novel fashion arrangements and methods for coordinating communication between multiple remote units and a central unit to facilitate multipoint-to-point transmission are set forth. The invention has applications in a wide variety of data transmission schemes, including ADSL systems, where signal transmission over twisted pairs, fibers and / or or hybrid telephone lines, cable systems that transmit signals over a coaxial cable, and digital mobile television systems that transmit over radio signals.

In one embodiment, a discrete multi-tone data transmission system has a plurality of discrete sub-channels including an auxiliary bus. In one method aspect, a selected remote unit, if it wishes to initiate a communication, loop synchronizes its own clock to the clock of the central unit and then transmits a remotely initiated synchronization signal over a dedicated overhead subchannel or set of overhead subchannels in the overhead bus central processing unit. The central unit responds with a centrally initiated synchronization signal containing information indicating a frame boundary phase shift required to better synchronize the selected first remote unit with other remote units currently communicating with the central unit. The remote unit responds by shifting the phase of the frames it outputs, as indicated by the centrally initiated synchronization signal. Synchronization can be done either iteratively or in a single step. This synchronizes the boundaries of the frames output by the selected remote unit with the boundaries of frames issued by the other remote units currently communicating with the central unit. The synchronization should be such that the frame boundaries of the various remote units substantially coincide when received at the central unit.

According to one embodiment of the invention, the additional bus contains two dedicated additional sub-channels, and the remote-initiated synchronization signal and the centrally initiated synchronization signal are transmitted via various additional sub-channels. In other embodiments, a single dedicated auxiliary sub-channel may be used, or multiple dedicated sub-sub-channels may be used. In some embodiments, the number of subchannels available for the selected remote unit for transferring data to the central unit is dynamically allocated. Also, specific constructions of central and remote modems suitable for implementing such a system are described.

According to another aspect of the invention, periodically synchronized idle times are provided in the uplink communication stream. The synchronized silence times are used to handle a variety of additional functions, such as initializing new remote units, checking the quality of transmission channels, and handling data transfer requests.

According to one embodiment, a method for synchronizing frames transmitted from an initiating remote unit to the central unit is described, wherein frames are transmitted from other remote units to the central unit. In this embodiment, synchronized silence times are provided periodically on the plurality of discrete subchannels provided for uplink communication. When a remote unit is initialized, it transmits a broadband initialization signal to the central unit during a synchronized idle time. The broadband initialization signal has several initialization signals transmitted via separate subchannels. In a preferred embodiment, the remote unit monitors the downlink communication if it wishes to initialize, and substantially synchronizes the frame boundary of the wideband initialization signal with a frame timing mark transmitted in the downlink received signal from the remote unit. The central unit receives the broadband initialization signal and sends back a synchronization signal to the first remote unit. The synchronization signal contains frame boundary phase shift information necessary to better synchronize the frame boundaries of signals transmitted by the remote unit with frame boundaries of signals transmitted by other remote units communicating with the central unit , The remote unit then shifts the phase of the output frames to allow synchronization.

The synchronized idle time used in this embodiment is of sufficient duration to allow a idle time tag to be transmitted from the central unit to the remote furthest from the central unit, thereby transmitting an initialization signal returned from this furthest remote unit to the central unit This can all be done within the synchronized idle time.

According to another embodiment, a method for dynamically checking the subcarrier transmission quality from the remote units to the central unit is described. This allows the allocation of bandwidth to the remote units by the central unit. In this embodiment, training signals are transmitted from one of the remote units over the multiple subchannels provided to enable uplink communication during a selected synchronized idle time. The training signals are monitored by the central processing unit which determines a set of channel properties indicating the bit capacities of the various subchannels for transmitting signals from the selected remote unit. The central processing unit may then use the set of channel properties when determining which subchannels to allocate to the selected remote unit for uplink communication.

According to a preferred embodiment, the steps of transmitting and monitoring may be repeated for several different remote units to determine the channel characteristics for each of the different remote units. It is preferably arranged so that the various remote units transmit their respective training signals during different rest periods. The set of channel properties for each remote unit may be stored within a matrix of channel properties containing information indicating the capacities of the channels from each of the remote units to the central unit. The channel property information can then be used to allow the dynamic allocation of bandwidth to different remote units. In another preferred embodiment, the remote units transmit their respective training signals only in response to receiving a retraining signal from the central processing unit. This allows control of the system.

In accordance with another embodiment of the invention, a method of informing the central processing unit of the transmission requirements of a remote unit is described. In this embodiment, a remote unit that wishes to initiate or change communication transmits a data request signal to the CPU at a time other than during a sleep interval. The central unit then sends an authorization signal to the remote unit, which allocates a certain idle time. The remote unit then transmits data request information over several discrete subchannels during the allocated idle time. If the remote unit requirements are known, the remote unit centralizes one or more subchannels in response to the data request information.

According to a preferred embodiment, the data request signal may indicate either a desire to transmit at a particular data rate or a desire to transmit a particular amount of information. In the former case, the central unit allocates sufficient subchannels to the remote unit to allow transmission at a requested data rate specified in the data request information. In the latter case, the central unit allocates one or more subchannels for a time sufficient to transmit an amount of information specified in the data request information.

According to another preferred embodiment, the remote unit may transmit a defined data packet request signal after the data request information has been defined and transmitted. When this happens, the central unit immediately allocates at least a subchannel to the selected first remote unit in direct response to the defined data packet request.

According to another preferred embodiment, the remote units monitor information provided in the downlink communication stream before a data request signal is transmitted, and transmit the data request signal only over subchannels that are reported not to be used. In another preferred embodiment, a first value of the data request signal indicates a data rate request, a second value of the data request signal indicates a data packet request, and a third value of the data request signal indicates a defined data packet request. In such a device, the data request signal may only be a two-bit signal.

In another embodiment, each frame of the discrete multi-tone signal includes a plurality of symbols, and each remote unit is assigned an associated symbol during which the data request symbol can be transmitted. In this embodiment, the central processing unit determines the identity of a particular remote unit transmitting a data request signal based, at least in part, on the symbol during which the data request signal is received.

It should be understood that the various embodiments may be used either alone or in combination with one or more of the others. The various periods of rest described need not be the same length, and the rest periods described in connection with the third embodiment occur more frequently than the other two.

According to another embodiment, a fast access transmission mode is provided. In this embodiment, a communication access request having a unique identifier for the remote unit is transmitted from the requesting remote unit to the central unit. The request is transmitted on at least one idle subchannel using a modulation scheme in which no equalization is required for decoding at the central unit. The central unit then allocates the appropriate sub-channels to the requesting remote unit.

In further aspects of the invention, discrete multipoint transmitters and receivers capable of implementing the various methods are described. It should be understood that the various embodiments may be used either alone or in combination with one or more of the others. The systems described can be used regardless of whether the downlink signals are also discrete multicarrier signals. In several preferred embodiments, the bidirectional communication system is a cable system in which signals are transmitted over a coaxial cable, although other systems are contemplated.

BRIEF DESCRIPTION OF THE DRAWING

The invention, together with further objects and advantages, may best be understood by reading the following description together with the accompanying drawings, in which:

FIG. 1 is a block diagram of a communication system with a multiple remote unit head end CPU; FIG.

FIG. 2 is a frequency diagram showing the use of multiple limited subchannels used in a DMT system having a pair of dedicated auxiliary subchannels; FIG.

FIG. 3 is a block diagram of a central office modem architecture suitable for realizing synchronization in accordance with the present invention; FIG.

FIG. 4 is a block diagram of a remote unit modem architecture suitable for realizing synchronization in accordance with the present invention; FIG.

FIG. 5 is a block diagram illustrating a remote unit synchronization device suitable for realizing synchronization and up-symbol arrangement; FIG.

FIG. 6 is a graph plotting the phase error versus frequency, the slope being proportional to the time error and the y-intercept being proportional to the phase error of the carrier; FIG.

FIG. 7 is a timing diagram of a DMT data transmission system according to an embodiment of the present invention; FIG.

FIG. 8 is a flow chart illustrating a method of initializing a remote unit according to one aspect of the present invention; FIG.

FIG. 9 is a flow chart illustrating a method of retraining a remote unit according to a second aspect of the present invention; FIG.

FIG. 10 is a flow chart illustrating the steps taken by a requesting remote unit to establish communication with a central processing unit; FIG.

FIG. 11 (a) is a flow chart illustrating a method of allocating bandwidth to a remote packet requesting remote unit; FIG.

FIG. 11 (b) is a flow chart illustrating a method for allocating bandwidth to a remote unit requesting a defined data packet request; FIG.

Fig. 11 (c) is a flow chart illustrating a method of allocating bandwidth to a data rate requesting remote unit, and Figs

Fig. 12 is a graph showing a frame transmission sequence in a time division multiple access data transmission scheme.

DETAILED DESCRIPTION OF THE INVENTION

It has been shown that discrete multi-tone (DMT) data transmission schemes facilitate high performance data transmission. One of the advantages of the DMT architectures is that they have high spectral efficiencies and can adaptively avoid various signal distortion and noise problems. Because of their very high data transfer capabilities, the selection of a DMT data transfer scheme in most applications offers much room for service expansion as the data transfer system requirements increase. Discrete multi-tone technology finds applications in a wide variety of data communication environments. For example, consideration is given to the use of a discrete multi-tone data transmission scheme by the North American ATIS Standard for Asymmetric Digital Subscriber Line (ADSL).

A detailed description of the protocols for the discrete multi-tone (DMT) transmission scheme of the North American ATIS-ADSL standard is given in the ATIS article mentioned above. The standardized system uses 256 "tones" that are 4.3125 kHz wide in the forward (downlink) direction. The frequency range of the tones ranges from zero to 1.104 MHz. The lower 32 tones can also be used for duplicated uplink data transmission. Improvements to this system, which are thought to increase the transmission bandwidth by up to an order of magnitude, have been proposed according to the present invention in other applications. In other systems, the number of subchannels and / or the used bandwidth of the subchannels can be changed within wide ranges. However, IFFT modulation occurs, typical values ​​for the number of available subchannels are integer powers of two, so that, for example, 128, 256, 512, 1024, or 2048 subchannels result.

As has been described in the Background section of this application, a limitation of discrete multi-tone transmission systems is that in order to support multiple end-connection points served by a single line, the up-signals must be synchronized when they arrive at the central unit. This synchronization problem has limited the attractiveness of discrete multi-tone (DMT) data transmission schemes in certain applications, such as cable systems and wireless mobile television broadcasting, because these systems have a single line (single medium) for servicing a relatively large number of independent remote units, typically operated by different subscribers is used.

First, a schematic transmission scheme for a typical multi-user subscriber network will be described with reference to FIG. A central unit 10 (having a central modem) communicates via a common transmission line 17, which is divided into a plurality of supply lines 18, with a plurality of remote units. Each lead 18 serves an associated remote unit which typically has a remote modem 15 receiving the signals and a remote device 22 using the data. A service provider 19 is typically arranged to supply the data to the central modem for transmission to the remote modems 15 and to handle the data received by the central modem from the remote modems. The service provider 19 may take any suitable form. For example, the service provider may take the form of a network server. The network server may take the form of a dedicated computer or a distributed system. A variety of transmission media can be used for the transmission line. For example, twisted pair telephone lines, coaxial cables, fiber lines, and hybrids that use two or more different media all work well. This approach also works well with wireless systems.

As those skilled in the art will appreciate, a requirement of discrete multi-tone data transmission systems such as that contemplated herein is that when two or more units (typically two remote units) attempt to independently send information to a third unit (ie, central processing unit 10), the signals from the remote units must be synchronized, because otherwise at least some of the signals for the central unit 10 are incomprehensible. The problem with using discrete multitone transmissions in such a system is that the length of the leads 18 typically varies from one remote unit to another. Therefore, even if the remote units are synchronized with the clock of the central unit 10, their transmissions are phase shifted back to the central unit 10 by an amount that depends, at least in part, on the length of the associated feeder. In practice, these types of phase shifts may obscure remote initiated transmissions for the central modem.

An exemplary DMT transmission band is shown in FIG. As can be seen there, the transmission band has a plurality of sub-channels 23 via which independent carrier signals (referred to as sub-carrier 27) can be transmitted. The DMT transmission already divides a transmission medium itself into a plurality of sub-channels 23, which each transmit data independently. The data on each sub-channel 23 may correspond to another signal, or they may be grouped at higher data rates representing a single or a fewer number of larger bandwidth transmissions. These sub-channels 23 are fully implemented in the DMT with digital signal processing, thus rendering unnecessary analog separation filters and maximizing spectral efficiency. The number of subchannels used can be varied widely according to the requirements of a particular system. However, when a modulation is performed using an Inverse Fast Fourier Transform (IFFT), typical values ​​for the number of available subchannels 23 are integer powers of two, so that, for example, 128, 256, 512, 1024 or 2048 subchannels 23 result. For example, in an embodiment designed for use in a cable-based subscriber system, 1024 sub-channels 27 may be used, each carrier being limited to a 32-kHz sub-channel 23. This provides a frequency bandwidth of about 32 MHz, in which the remote units can communicate with the central unit 10.

The number of remote units that can be used in a given system can vary greatly according to the needs of a particular system. For example, in one embodiment of the described cable-based subscriber system, it may be desirable to allow up to 500 remote units to communicate with a single central processing unit. In systems that consider such a large number of remote units, it may be desirable to arrange the remote units in groups. Of course, the groups do not all need to contain the same number of units. For example, a system that allows up to 500 remote units can divide the remote units into eight groups, each group allowing up to 90 remote units, and each group of remote units being assigned a fixed frequency band. For example, the frequency spectrum can be divided into several fixed frequency bands of the same size. In the particular embodiment described, each group is assigned an eighth of the 32 MHz or about four megahertz. Therefore, each group has about 4 MHz and corresponding to 128 sub-channels 23 usable for transmission to the CPU 10. Grouping allows CPU 10 to keep track of the remote units as they go online and offline.

The groupings may be formed using a number of methods. For example, a first group could consist of consecutive subchannels 0-127, a second group of successive subchannels 128-255, etc. Alternatively, the mapping of subchannels 23 to the respective groups may be interleaved throughout the spectrum. For example, the subgroups 0, 8, 16, 24, 32 can be assigned to the first group. may be assigned, the second group may comprise the sub-channels 1, 9, 17, 25, 33 ..., the third group may have 2, 10, 18, 26, 34 ..., etc. Interleaving the sub-channels 23 assigned to the groups helps reduce the likelihood that noise present in a particular area of ​​the frequency spectrum will affect a significant portion of the transmissions in a single group. Instead, the noise affects only part of the spectrum of each group. As will be understood by those skilled in the art, the frequency bandwidth of the uplink channel, the size of the subchannels 23 and the groupings are not limited to the numbers given in the described embodiment but may be chosen to suit the requirements of the particular use of the transmission system.

One method of addressing the aforementioned synchronization problems is to use dedicated dedicated subchannels 28 and 29 (Figure 2) to enable synchronization. In this embodiment, the upstream overhead sub-channel 28 transmits synchronization signals from the various remote units to the central modem. The downstream overhead sub-channel 29 transmits synchronization signals from the central modem to the various remote units. The additional subchannels 28 and 29 may be located at any suitable frequency position within the transmission band. In many embodiments, such as the asymmetric digital subscriber loop system discussed above, it may be desirable to place the additional subchannels near the upper or lower frequency edge of the downlink signal to minimize their interference with adjacent subchannels. Further, if the constraints of the system allow it, it may be desirable to separate the additional subchannels by at least one or two subchannels from other subchannels used for data transmission to minimize the possible interference caused by the synchronization signals. This is desirable because the synchronization signals are often unsynchronized with other transmissions. Therefore, they cause more distortion than other signals because they are out of sync. Accordingly, a small buffer is helpful. It may also be desirable to use relatively low power signals along the same lines as supplemental subcarriers to further reduce interference problems in some cases.

As will be described in more detail below, in another aspect of the present invention, synchronized idle times in the upstream communication stream are provided periodically. The synchronized silence times may be used to perform a variety of additional functions, such as initializing new remote units, checking the quality of transmission channels, and handling data transfer requests. Next, referring to Fig. 7, an example frame-limited transmission time sequence will be explained which has a number of synchronized silence times suitable for performing the additional functions. In the illustrated embodiment, the transmissions are divided into chains of transmission frames 32. Each transmission frame has a transmission interval 33 and a first idle interval S1. Each transmission interval 33 is further divided into a plurality of symbol periods 35 as shown. Several transmission frames 32 are then grouped into a superframe 36. In addition to the transmission frames 32, each superframe 36 also has a second idle time interval 38. In the described embodiment, the second idle time interval 38 may be used either as an initialization interval (S2) or a retraining interval (S3).

The actual time periods provided for the transmission interval 33, the idle time interval S1, the initialization interval S2, and the retraining interval S3 may vary greatly according to the requirements of a particular system. Similarly, the number of transmission frames 32 in a superframe 36 can vary greatly. For example, in one suitable embodiment for use in the described cable-based subscriber system, it is contemplated to set a transmission interval 33 to a time sufficient to transmit 63 symbols and to set the S1 time interval 34 to the time period corresponding to one symbol. The initialization interval S2 may be used as an alternative arrangement for synchronizing the remote units. Accordingly, the length of the second idle time interval 38 is typically determined by the physical aspects of the communication system, as discussed in more detail below. In general, the remote units may not transmit during an S1 or S3 sleep interval unless permission has been granted by the central unit 10. In some embodiments, the remote units may not transmit even during an S2 sleep interval unless they attempt to initiate an installation, as described in more detail below.

Next, the use of assisting supplemental subchannels to facilitate the synchronization of newly added remote times will first be described in more detail with reference to Figs. 2-4. First, the remote modem 50 has a remote synchronization controller 80 which cooperates with a central controller 60 in the central modem unit. As briefly discussed above, in the described embodiment, two auxiliary supplemental subchannels are provided to facilitate communication between the controllers. When the remote modem 50 is initialized and wants to be connected to the transmission path, the remote controller 80 observes downlink signal transmissions, which inherently contain the clock information of the central modem. This is sometimes done using pilot signals, although other schemes may be used. The remote modem is then "loop synchronized". That is, it couples its own clock to the clock of the central modem. The remote control device then sends via the additional sub-channel 28 a synchronization signal to the central unit 30. The synchronization signal passes through the transmission medium and enters the receiver section of the central modem unit 30. When the central modem 30 receives a remote initiated synchronization signal (uplink synchronization signal) while it is communicating with other remote units, it compares the frame boundaries of the remote initiated synchronization signal with the frame boundaries of other remote units of received signals. Typically, a phase shift occurs between the frame boundaries, which is detected by the controller 60. The controller 60 then generates a downlink synchronization signal which is sent back to the remote units via the overhead subchannel 29.

In the embodiment described and illustrated, the controller 80 is responsible for generating the uplink synchronization signal if the remote modem wishes to initiate communication with the central modem. The up-synchronization signal is transmitted from the controller 80 to the multiplexer and encoder 143 and directed specifically to the up-link sub-channel 28. It should be noted that because its nature is known, the synchronization signal could also be input to the transmitter at other locations, or even applied directly to the analog interface 148. Typically, the synchronization signals and / or the sequence would be the only signals sent by the remote unit until synchronization is complete. The uplink synchronization signal is then transmitted via the auxiliary subchannel 28 to the central modem where it is received by the receiver 70. The demodulator 76 of the receiver then supplies the demodulated synchronization signal to the controller 60 of the central modem. The central controller 60 detects the remote initiated synchronization signal and compares its frame boundary with the frame boundaries of all signals simultaneously received from other remote units. When the central modem 30 communicates with other remote units, it is likely that the frame boundaries of the remote unit requesting access will out of phase with the frame boundaries of those remote units that are already communicating with the central modem because lead length deviations occur. In these cases, the central controller 60 initiates a return sync signal (down sync signal) indicating the phase shift required to arrange the frame boundaries (which takes the form of a time delay). The return synchronization signal is then transmitted via the second overhead subchannel 29 to the remote units. As well as the uplink synchronization signal, the downlink synchronization signal can be inserted into the downlink data stream at the encoder.

The nature of the downlink synchronization signal may vary, however, for example, the synchronization signal may simply indicate that the remote unit should advance or delay the frame boundary by one sample. In a somewhat more complicated system, the controller may attempt to calculate the number of samples by which to advance or delay the frame boundary, and may send a signal prescribing the number of samples by which the frame boundary should be shifted. Other signal interpretations may be used. As will be discussed in greater detail below, in many embodiments the sampling rate in uplink communication is an integer factor of the downlink sampling rate. The delay described is based on the sampling rate of the central modem, not that of the remote modems.

Because a plurality of remote modems are all connected to the same transmission line 17, the synchronization signal is received by all the operating remote modems. The signal is then transmitted from the decoder of each remote modem to the associated controller 80. However, the remote controllers 80 are arranged to ignore synchronization signals on the overhead sub-channel unless they are attempting to initiate communication with the central modem. This can be achieved in many ways. For example, the downlink synchronization signals may include an address directed to a specific remote modem. Alternatively, the remote modems may simply assume that the central modem's signal is destined for them when they are attempting to initiate communication. The remote unit remote controller 80 attempting to initiate communication receives and interprets the centrally initiated synchronization signal and instructs the frame synchronizer 147 to implement the requested phase shift time delay (or the requested phase shift time advance). A second remote initiated synchronization signal is then sent. If the new synchronization signal is not in synchronization, the same process is repeated. In one embodiment, the synchronization signal merely instructs the frame synchronizer to advance or delay a sample. It is believed that such an incremental system works well in most DMT applications to quickly synchronize the remote unit. For example, in a system having a symbol rate (frame rate) of 8 kHz (and thus a symbol period of 125 μs) corresponding to 64 Kbps, each frame having 128 samples and a header, distribution networks having lead lengths deviations of up to two miles would be less than about ten milliseconds is required to synchronize by a single sample using a simple method of feed and delay.

If the remote-initiated signal is determined to be synchronous, the central controller sends a return sync signal over the second overhead sub-channel 29 indicating that no further phase shift is required and that the remote unit can initiate full communication with the central modem at the desired phase shift , If the remote modem is synchronized before it is recognized by the central modem, the data tones transmitted immediately after initialization are used to identify the remote modem. It is expected that the relative phase shift of the frame boundaries will depend primarily on fixed constraints, such as the transmission length across the various feeders. Therefore, once a remote modem is synchronized, it does not need to be resynchronized unless the connection is terminated or interrupted.

It should be understood that when the central unit is not in communication with other remote units while receiving a request to initiate a communication, the central control unit 60 returns only a synchronization signal indicating that no phase shift was required and that the full Communication can begin. Of course, a similar signal is also generated in the case that the requesting remote unit is in synchronization with the other remote modems when it first tries to initiate a communication. If the remote modem receives such a signal, the same process can be performed with the required phase shift simply zero.

Typically, the central controller 60 also provides information indicating the subchannels that the remote unit should use for its transmissions, etc. As mentioned above, the subchannel allocation may be changed dynamically during use. Although this feature is important to the discrete multi-tone transmission scheme, it is not particularly significant to the present invention, and thus will be described only briefly, although described in detail in the cited references.

Synchronizing a remote modem with the central modem requires retrieving the sample clock and central modem carrier. In a preferred embodiment, these clocks are obtained by viewing the phase errors over at least two tones. The phase error for these tones can be calculated with respect to a fixed known transmitted phase on the tones (ie "pilot tones"). They may alternatively be determined on the assumption that decisions about the transmitted phases are correct and determined by calculating the offset between the phases before the decision and after the decision (ie a decision-supported phase error calculation). The slope of the phase error curve shown in Fig. 6 is proportional to the time phase error, while the constant part (the y intercept) of the phase error curve is the carrier phase error. The time phase error (sampling phase error) and the carrier phase error are determined by a phase detector 181 and input to the phase locked loops 182, 184, which synthesize a sample clock and carrier frequency at the acquired central modem frequencies, as shown in FIG. The carrier is used to demodulate the downlink signal to baseband, and the sampling clock is used (after splitting by divider 189) to clock the analog-to-digital converter (s) (ADC). If the data tones and beeps occupy separate tones, more than one analog-to-digital converter may be used at lower sample clocks rather than a single higher-speed ADC clock. In embodiments incorporating the notch filter 185, voltage controlled oscillators 183, 186 are provided for controlling the location of the notch filter.

The same sample clock (after splitting by divider 189) is used for up-to-digital-to-analog converters. The up-beam may or may not be synchronized with the down-beam. If it is not synchronized, the central modem uplink receiver must acquire the uplink carrier phase, and the central modem uplink receiver may otherwise use a rational phase locked multiple of the data carrier downlink for data recovery. Remote broadband modems preferably use a sampling clock that is similar to the sampling clock in the central modem. These remote modems do not share the recovered sampling clock. Remote narrowband modems that receive only a few tones use a sample clock that is an integer divisor of the recovered sample clock. Accordingly, remote narrowband modems may be less costly to implement.

The DMT symbols transmitted up from the remote modems, even if generated by different remote modems, must arrive at the central modem simultaneously, as discussed above. Therefore, the delay synchronizer 147 inserts a delay by an integer number of sampling clocks into the up-transmitted signals. This delay is programmed under the control of the downlink synchronization signal, as discussed above. Again, note that the delay is based on the sampling rate of the central modem, not the remote modem. In particular, as shown in FIG. 5, the far modem sampling rate may be an integer factor of the central modem sampling rate. However, the signals must be synchronized with the central modem, and the synchronization settings must therefore be made based on the sampling rate of the central modem.

If two remote modems simultaneously attempt to initiate communication with the central modem, a conflict will occur and the central controller 60 is likely to be confused by the uplink synchronization signals. In this case, its down-synchronization signal indicates an inappropriate phase shift, and the confirmation-synchronization signals would then not be properly synchronized. In one embodiment, the central controller 60 could detect the problem and instruct the remote units to pause and attempt to establish the communication at a later point. In another embodiment, the central controller could simply send another downlink synchronization signal indicating the required additional phase shift. In both cases, the remote unit quickly recognizes that a problem exists and assumes that a conflict is occurring. In this situation, a suitable conflict resolution scheme can be used. A simple conflict resolution scheme is to delay each remote modem by a random amount of time and try to re-initiate communication after the random delay. As long as the delay is set in a manner in which it is not likely that the remote modems will consistently have the same delay pattern, their requirements will eventually be sufficiently separated for them to be brought online independently. A variety of wait time distributions can be used. For example, it has been found that Poisson distribution works well.

It should be noted that the IFFT modulation scheme described works very well with systems that are designed to transmit relatively large blocks of data and therefore require more than a handful of tones. However, in many situations, the remote units do not need to periodically transmit large blocks of data. It may be cost effective in these situations to use a simpler conventional modulation scheme for transferring information from the remote units to the central unit. In these cases, both the remote transmitter and the central receiver would be replaced by the appropriate components. However, it would still be necessary to synchronize the remote units, as discussed above.

In operation, the central modem transmits a composite DMT signal in which all (or usable) tones are used such that each remote modem knows the tones it is to receive and the number of bits associated with each of the received tones. The remote modems, in turn, each use only a subset of the available uplinks. The signals transmitted from the central modem to the remote modems can be used to dynamically assign the tones available to a particular receiver. Alternatively, the assignment could be made to a static system in the downlink synchronization signal. The dynamic allocation may either take place on another dedicated auxiliary channel or control channel, or it may be multiplexed with other non-control signals. In the system described, the up-signals are timed to arrive at the central modem approximately simultaneously. Precise placement is not required but the system works best when the limits on the sampling rate of the central modem are tightly spaced.

Next, with reference to FIG. 8, an alternative method of initializing a first remote unit during installation according to another aspect of the invention will be described using the second rest periods S2 described. As discussed above, when a remote unit goes online for the first time, it must be initialized so that the transmissions from the first remote unit arriving at the central modem are synchronized with the transmissions of all other currently installed remote units. That is, the frame boundaries of upstream DMT communications from the various remote units to the central unit must be substantially synchronized to the central unit for the transmissions to be understood by the central unit. The method described with reference to FIG. 8 is a method of achieving this synchronization using the rest times described.

First, the remote unit to be installed must establish a connection to the transmission network in step 302. The connection allows the remote unit to watch for the downlink transmissions from the central unit 10 and to transmit on any unused subchannel 23 of the uplink channel. There may be some frequency ranges in some systems that the system may not use. For example, some cable networks may have systems that use specific frequency bands. To avoid interference and maintain backwards compatibility, it is important that the remote unit never transmit within the forbidden frequency range even during initialization. Of course, certain frequency bands may also be prohibited for other reasons. Accordingly, in step 303, the CPU periodically transmits an identification of frequencies that must never be used. For systems using the concept of remote units, as discussed above, the central unit may also periodically transmit the number of the group that should be used by the next remote unit to be installed. Alternatively, the group assignment can be dealt with at a later point.

The newly attached remote unit pays attention in the downlink signals to information indicating that certain subchannels must not be used. The downlink signal also includes the frame timing and idle time flags required to synchronize the remote unit with the central unit. After the remote unit has synchronized with the downlink signal, in step 304 it sends an initialization signal at the beginning of an S2 idle period. This is done in a system by sending an initialization signal immediately after receiving an S2 silence time flag signal. The initialization signal indicates to the central unit 10 that a remote unit is requesting to install on the system. The remote unit may determine the onset of an S2 initialization silence period in a suitable manner. For example, a flag may be provided in the downlink communication from the central processing unit 10. The remote unit may transmit its initialization signal over all sub-channels 23, over a group of sub-channels 23, or on a single sub-channel 23, depending on the requirements of a particular system. In a preferred embodiment, the downlink signal indicates the group to be used by the next device to be set up, and the initialization signal is transmitted over all subchannels in that group.

The upstream initialization transmissions from the remote units to the central unit 10 can be done in any modulation scheme suitable for transmitting digital information. For example, amplitude, frequency and four-phase Umzastmodulationsschemata (QPSK modulation schemes) can be used. For the synchronization signal, differential QPSK modulation (DQPSK modulation) is desirable in a preferred embodiment to reduce the possibility of noise damage. Furthermore, the synchronization can be coded with a large amount of error correction and redundancy to ensure coherent communication.

The initialization signal preferably contains information about the remote unit. In a preferred embodiment, the initialization signal transmits the global address of the remote unit and the maximum required data transfer rate of the first remote unit. A global address is similar to the addresses used in Ethernet or mobile phone devices. These addresses are built into the communication device and remote from the addresses of all other communication devices. The maximum data rate required by the remote device depends on the device type of the remote device. For example, if the remote unit is a television, a minimum communication capacity to the central unit 10 would be required, with possibly only the uplink signals used to send information about movie selections or feedback from the viewer. On the other hand, if the remote unit is a teleconferencing transceiver, a large bandwidth would be required to transmit video and audio information from the remote unit to the central unit 10. Other components of relevant information about the first remote unit may also be sent together with the initialization signal in other embodiments.

Upon receipt of the initialization signal from the first remote unit, the CPU 10 determines in step 306 whether the initialization signal from the first remote unit collided with another initialization signal from another remote unit attempting to connect simultaneously. If a collision is detected, the CPU 10 returns a collision message to the remote units in step 308. The collision message tells the remote units that are trying to connect to try again. The colliding remote units then each wait a random number of S2 periods before an initialization signal is retransmitted. The probability of two remote units trying to initialize simultaneously is low. Requiring the colliding units to wait for random periods that are independent of one another reduces the probability of repeated collisions even further.

After the central unit 10 has received a valid initialization signal from the first remote unit, the central unit 10 sends back a synchronization signal 310 to the remote unit. In one embodiment, the synchronization signal includes the global address of the first remote unit, a node address associated with the first remote address, delay correction information, and information about the allocation of the subchannels 23 in the upstream channel. Both the global address and the node address may serve as a unique identifier for remote units, but with different degrees of transmission efficiency. The global address allows the first remote unit to identify that the synchronization signal is for you. The node address is assigned to the first remote unit to enable effective future communications. The global address can be quite long (e.g., 48 bits) to allow a reasonable number of global addresses for all likely established communication devices. The node address is a shorter address because only a limited number of remote units communicate with a single central unit 10. When a multi-clustered system is used, the node address also contains group identifier information, such as information about the group to which the first remote unit is assigned. In the embodiment described above, in which a total of eight groups are provided, this part of the address would have three bits to indicate which of the eight groups is the first remote unit. The remaining bits can uniquely identify the node, such as the specific remote unit, within its group.

It should be understood by those skilled in the art that the portion of the node address that specifies the group, that is, the group identity information, may be omitted entirely when a remote unit must uniquely identify with the central unit. This is because, by inspecting the frequency band of the unique identifier message, the central unit can determine the group from which the remote unit message was sent. In this way, a remote unit only needs the bit pattern in the node address, thereby identifying it in the group, that is to send the unique inner group identifier information, to uniquely identify itself to the central unit. This received intra-group identifier bit pattern, in combination with the acknowledged group identifier information, provides the central unit with the complete node address of the requesting remote unit. In the preferred embodiment having 128 subchannels per group, the unique identifier information may further comprise uplink units of only 7 bits.

The delay correction information tells the first remote unit how much the frames transmitted by the first remote unit need to be delayed to synchronize with signals from the other remote units being connected. The delay correction is determined from the amount of delay that the central unit detects between the time it sends an idle time flag (S2) and its receipt of the initialization signal. For example, if the maximum delay in the channel is TRT (Max), for example, the maximum round trip delay, and the delay associated with a given remote unit is TRT (i), the delay correction for that remote unit is TRT (Max) - TRT (i) , The round trip delay for a remote unit is defined as the time required for a signal to pass from the central unit to that remote unit and an immediate response due to the central unit, including a minimum random delay attributable to the processing. Using this information, the first remote unit can adapt its transmissions and be synchronized with the other remote units connected so that the frames of the remote units arrive at the central unit 10 simultaneously. The first remote unit can also know which sub-channels 23 are currently being used by the other remote units connected. In another embodiment, information about characteristics of subchannels 23 is routinely transmitted over the downstream channel to all remote units. In these systems, it is not necessary to send channel allocation information together with the synchronization signal.

An advantage of transmitting the initialization signals over a wide portion of the available spectrum is that the delays vary somewhat depending on the frequency at which the signal is transmitted. Therefore, when the initialization signals are transmitted through a plurality of subchannels 23, the required phase shift can be calculated from an average of the individual delays.

The length of the S2 time interval, as previously discussed, depends on the physical nature of the communication network. In a preferred embodiment, the S2 time interval need only be longer than the duration of the initialization signal plus the difference between the maximum and minimum round trip delays for the network. For example, in a typical system using a fiber optic link as transmission line 17 and coaxial cables as leads 18, the fiber link is common to all paths between the central unit and the remote units and the difference between the maximum and minimum round trip delays for the network only from the cable part of the network. When using a 2 mile coaxial cable length and for a given run time of about 7.5 microseconds per mile, the maximum round trip delay is about 32 microseconds and the minimum round trip delay is about 2 microseconds. In a preferred embodiment, a symbol is about 30 microseconds long, and an initialization signal has two symbols, so that, for example, an S2 time interval of 4 symbols would be appropriate.

It may be desirable in certain embodiments to repeat steps 304-310 to verify the received information and / or to ensure that the remote unit is properly synchronized.

Upon achieving synchronization, the first remote unit responds by transmitting a set of synchronized broadband training signals over all subchannels 23 during the next available S2 or S3 time interval in step 312. The specifics of the training step will be discussed below with reference to FIG Fig. 9 described in more detail. In some embodiments, the central unit 10 instructs the first remote unit to use a specified S3 time interval (eg, wait for the third S3). After receiving the training signals, the central processing unit 10 determines the capacities of the various sub-channels 23 to handle the transmission between the first remote unit and the central processing unit 10 (step 314). The central unit 10 preferably has a prior knowledge of the content of the training signals. This will allow the CPU 10 to find out the optimum balance of the sub-channels 23 as well as the maximum bit rates that a sub-carrier 27 may protrude on the sub-channels 23 between the first remote unit and the central unit 10. The CPU 10 stores the characteristics of the subchannels 23 with respect to the first remote unit 316. In a preferred embodiment, the central processing unit 10 stores the information in a bit / carrier matrix containing an indication of the number of bits that each of the sub-channels 23 can transmit from each of the remote units. Such a matrix allows CPU 10 to track the capacity of each of the various subchannels 23, and is available when bandwidth is allocated to the remote units. This also enables the dynamic allocation of subchannels based on the current characteristics of the transmission environment.

Next, referring to Fig. 9, a method for periodically checking the capacity of the various subchannels from a selected remote unit to the central unit will be described. As will be understood by those skilled in the art, the capacity of the transmission line may vary somewhat with time at different frequencies. It is therefore desirable to periodically update the information of the central processing unit regarding the characteristics of the subchannels 23 with respect to each of the remote units that it serves. In the described embodiment, this updating occurs during the S3 rest periods. In the illustrated embodiment, the S3 rest periods have the same length as the S2 rest periods. It should be understood that a single transmission line testing operation may be used for both the initial training test and the periodic test.

In the described embodiment, the CPU 10 initiates a retraining event in step 330 by transmitting a retraining command to a first remote unit (the remote unit x) that is currently communicating with the central unit 10. The first remote unit waits for the next available S3 retraining idle time interval to transmit a set of training signals over the available sub-channels 23 (step 332). In an alternative embodiment, CPU 10 may assign a specific S3 silence interval for use in transmitting the training signals, rather than the next available S3 time interval. The set of training signals is typically limited to the subchannels associated with the group, and is typically further limited to a subset of the total available subchannels of the group to achieve a cost effective design. Therefore, the number of actually used training signals can be changed widely according to the requirements of a particular system. As with the initialization process, CPU 10 analyzes the signals it receives and updates the bit rates in the channel property matrix according to the associated remote unit (step 334). The central processing unit 10 then determines whether a change in subchannel allocation is required for the remote unit. It can thus determine whether additional or fewer sub-channels 23 should be assigned to the first remote unit to meet the throughput and error likelihood requirements of the first remote unit. If a change is required, the CPU 10 reassigns the subchannels 23 to the first remote unit in step 338.

If it is determined in step 336 that no correction is required, or after the required changes have been made in step 338, the CPU 10 checks in step 340 whether immediate remote training requests have been issued by other remote units. If it is determined in step 340 that there are no immediate retraining requests, CPU 10 checks to see if retraining of the first remote unit was a result of an immediate retrain request by checking in step 347 if a valid old address (oldx ) is present. If there is no valid old address, the CPU 10 increments the counter (x) in step 349 and returns to step 330 where it transmits a retraining signal to the next remote unit. On the other hand, if it determines in step 340 that there was a valid old address, the CPU 10 sets the counter to have a value one greater than the old address, which corresponds to the address of the remote unit that belongs to the time at which a request for immediate retraining was received would have been the next (step 350). That is, x = oldx + 1.

If an immediate retrain request has been detected in step 340, the CPU 10 stores the address of the first remote unit in step 342 as an old address (oldx). The CPU 10 then sets the counter (x) to the address of the requesting remote unit and uses it as the address of the next remote unit currently being retrained 344. The logic then returns to step 330. The retraining process may then be continuously repeated for all remote units currently communicating with the central unit 10. Of course, the algorithm used to select the remote units for the retraining can be changed widely to meet the requirements of a particular system.

In one embodiment, the remote units are also retrained that have been initialized but are not currently communicating with the central unit 10. In this case, the central unit 10 does not need to determine whether the assignment of the sub-channels 23 for the retrained remote unit must be changed because it is not actively communicating with the central unit 10. The central processing unit 10 can only store the updated channel characteristics to be used when the remote unit requests communication with the central processing unit 10.

The central processing unit 10 is preferably adapted to receive a retraining request on unused sub-channels 23 during a transmission time interval 32. In a preferred embodiment, the transmission time interval 32 corresponding to the maximum number of possible remote units within a group is 64 symbols long. A remote unit that requires immediate retraining sends a flag during one of the symbol times assigned to the requesting remote unit in transmission time interval 32. In this way, the central unit 10 can immediately determine from the location of the flag, which remote unit has sent the request. For example, the remote units 0-63 in group eight may be assigned the respective symbols 0-63 in the transmission time interval. If an indicator on an unused sub-channel 23 arrives during the ninth symbol position in the frequency band of group eight, the central unit 10 knows that the ninth remote unit in group eight has sent a retraining request. As those skilled in the art will appreciate, assigning the remote units to symbols can be done in many different ways.

As discussed above, to enable a dynamically-assigned discrete multi-tone transmission scheme, there must be a mechanism by which the remote units can communicate a data transfer request to the central unit. In one embodiment, the S1 idle times are used in conjunction with a data transfer request to allow the initiation of a transfer. In the described embodiment, a remote unit may send three types of data requests to the central unit. These include a data packet request (DPR), a defined data packet request (DDPR) and a data rate request (DRR). In this embodiment, a data packet request indicates the desire of the remote unit to transmit a certain amount of information (typically defined in terms of a number of bytes of data). A defined data packet request indicates the desire of the remote unit to transmit a packet or group of packets whose properties are already known to the central unit. For example, the information about the remote unit to which data packets should be sent from the requesting remote unit may already be stored in the memory of the central unit. Other information known to the central processing unit may include, for example, the required transmission rate for the data packets, the number of sub-channels required by the requesting remote unit, and the like. A data rate request indicates the desire of the remote unit to transmit data at a particular rate.

The described data transfer requests, in one embodiment, may be coupled to the above-described immediate retrain request in a simple two-bit signal having four states. For example, a state (1, 1) may correspond to a data rate request, a second state (1, 0) may correspond to a data packet request, a third state (0, 1) may correspond to an immediate retraining request, and a fourth state (0, 0) may correspond to one defined data packet request. Of course, the same information may be included as part of a larger signal, and / or the meaning of the various states may be changed. As described above, the two-bit data transfer request signal may be transmitted from a remote unit via sub-channels that are not in use. By assigning a particular symbol period to each remote unit, the central unit can easily identify the requesting remote unit without requiring independent identification information in the data transfer request signal. This transmission mode, which assigns a particular symbol period to each remote unit, is called a polled transmission mode.

As will be understood by those skilled in the art, in addition to merely identifying the type of information the remote unit desires to transmit when both the data rate request and the data packet request occur, the remote unit must normally transmit significantly more information to the central unit for the central unit to properly handle the request can. To provide short access times, the additional information is communicated to the central units during the next available S1 sleep interval. In particular, when the central unit 10 receives a valid data packet request or a valid data rate request, the central unit 10 instructs the requesting remote unit to transmit any additional information about the requesting remote unit request during the next available S1 idle period 34. During the S1 rest period, the requesting remote unit has access to as many subchannels as it needs to transmit the header information. Because both the data rate request and the data packet request effectively only require the assignment of an S1 idle period, they could simply share a single state in the two-bit data transfer request signal. Accordingly, in alternative embodiments, a single state could be provided to indicate the desire to associate an S1 rest period, and the nature of the request could be transmitted along with the other information during the S1 period.

If the system is not heavily used, there may be a relatively large number of sub-channels available to the remote unit when sending its data transfer request. During these periods, it may be possible to send all required header information along with the transmission of the data request in the same symbol period. Accordingly, the idle state in the data transfer request may be used in an alternative embodiment to alert the central unit that the remote unit is required Transmit header information concurrently with the data transfer request on unused subchannels. In the polled transfer mode, the time of the data transfer request identifies the remote unit sending the request. Accordingly, the advantage of this method is that access times for data rate and data packet requests can be reduced even further during periods of relatively low occupancy. Conflicts between two remote units do not occur because each remote unit sends during its assigned symbol period. If the remote unit determines that there is not enough bandwidth to accept all required header information in the assigned symbol period, it simply requests the assignment of an S1 idle period, as described above.

In a further embodiment, the central unit 10 may allocate a specific S1 interval 34 to be used by the requesting remote unit. This is especially useful when two or more remote units between two S1 intervals execute data packet or data rate requests.

As previously mentioned, a relatively large number of subchannels may occur which are unused and available to a remote unit for requesting access if the system is not heavily populated. For example, if the central unit determines that the occupancy of the system is low, for example, if the occupancy drops below a predefined occupancy threshold, the central unit may issue a command to all remote units to allow the remote units to communicate their communication access requirements using a transfer mode fast access to the central unit. The fast access transmission mode differs from the above-described interrogated transmission mode in which each remote unit is assigned a symbol period in which its data transmission request signal is to be transmitted. As the name implies, the fast access transmission mode significantly improves the access speed of a requesting remote unit by allowing the requesting remote unit to transmit a communication access request on one of the idle or unassigned sub-channels during a symbol period, independently whether he has been assigned this symbol period. The remote units know which subchannels are idle because, for example, the central unit monitors subchannel allocation and subchannel allocation information from time to time for all remote units.

Because a remote unit no longer has to wait for the symbol period associated with it to acknowledge a communication access request, it can enforce its communication access request as a need arises. On the other hand, the time of request in the fast access transmission mode does not provide information regarding the identity of the requesting remote unit. Therefore, in order to identify which remote unit intercepts a received communication access request signal, the fast access transmission mode requests that each requesting remote unit send a unique identifier of the remote unit upon requesting the access. As mentioned previously, the unique unit identifier may have only 7 bits for 128 sub-channel systems per group.

In one embodiment, the communication access request signal includes a data transfer request. As mentioned above, the data transfer request identifies the type of data request desired by the remote unit, such as DPR, DDPR or DRR. If two bits are used to identify a data transfer request, the last state may be used to indicate whether the header data is being sent simultaneously in the same symbol period or during the following S1 period. Obviously, if the data request is DDPR, header information may not be available because the central processing unit may already know the transfer requests, such as the destination of the data packet, the packet size, the priority rating and the like associated with a particular remote unit. If the data request is DPR or DRR, the last state defined by the two-bit data transfer request is examined by the central processing unit to determine when header information is being sent.

In another embodiment, the communication access request further includes the header information for DRR and DPR data requests. The inclusion of the header information increases the number of bits transmitted in fast transfer access mode. As the number of bits increases, the probability of a collision increases. Collisions occur when two remote units simultaneously enforce their communications access requests on the same idle subchannel. Thus, the preferred embodiment preferably keeps the number of bits transmitted in the fast access transmission mode as low as possible to minimize collisions. It will be understood that the fastest access transfer mode is best suited for DDPR data requests because it does not require header information to be sent from the remote unit to the central unit.

Therefore, a communication access request preferably includes only the unique identifier associated with the remote unit and the two-bit data transfer request. However, in one embodiment, if a communication access request does not include the two-bit data transfer request, the central processing unit may assume that a DDPR data request is desired and continue allocating subchannels to the requesting remote unit based on the stored data packet specification information associated with that remote unit ,

In the fast access transmission mode, it is preferably required that the communication access request be made from the remote one using a modulation method in which no matching is required during decoding. Unit is transmitted to the central unit. Alignment is necessary in certain modulation schemes where the central processing unit must know the characteristics of the sub-channel and the remote unit, such as the absolute amplitude of the received signal and the phase, to decode incoming data. Obviously, if a communication access request arrives at the central processing unit during a fast access transmission mode, the central processing unit need not know the identity of the requesting remote unit prior to decoding. This is because in the fast access transmission mode, a remote unit can enforce its communication access request during any symbol period and the timing of the request provides no information regarding the identity of the requesting remote unit.

Because the identity of the requesting remote unit is not known prior to decoding, the communication access request can not be decoded by modulation schemes that require prior knowledge of the subchannel and the identity of the remote unit, as is the case with QAM. In one embodiment, the present invention preferably encodes a remote unit communication access request using differential four-phase shift keying (DQPSK). When the DQPSK is used, the information regarding a communication access request is stored in the phase differences instead of in the absolute phase. It is also possible to choose an appropriate constellation so that the amplitude is irrelevant. In this way, a communication access request may be received and decoded by the central unit without requiring prior knowledge of the identity of the requesting remote unit.

As previously mentioned, in the fast access transmission mode, the requesting remote unit is not required to wait for the symbol period assigned to it to request access. Thus, the access time may only be given by the time required to send the communication access request plus the time it takes for the central unit to send to the requesting remote unit information associating subchannels for use by the requesting remote unit ,

In one embodiment, the fast access access mode is enabled by the central processing unit when the system usage is low and, for example, below a predefined occupancy threshold. Enabling the transmission mode with quick access during these times reduces the likelihood of collisions because there are more idle subchannels on which one or more remote units can enforce communication access requests. If a collision occurs, the CPU receives failed data, such as data that can not be decoded. Therefore, without knowing which remote unit is requesting access, the central unit can not allocate any subchannels to the appropriate requesting remote unit. In this case, a requesting remote unit may wait a predefined period of time after asserting its communication access request, and then, if no assignment occurs, retransmits the communication access request, preferably after waiting a random period of time, to reduce the likelihood of another collision , If, in one embodiment, the central unit receives a malfunctioning data transmission on an unassigned or idle subchannel, it assumes that a collision has occurred between two or more communications access requests and transmits a "collision" to all remote units, preferably after waiting for a random period of time detects "message to urge the remote units to resend their communications access requests.

It will be understood that the subchannel occupancy may increase because of the re-transmission activities by the remote units and in one embodiment of the central unit's transmission activity when there are a large number of collisions. If too many collisions occur, the system occupancy may exceed the predefined occupancy threshold, causing the central processing unit, in one embodiment, to issue a control command to all remote units to quickly suspend data transmission in the transmission mode and resume data transmission in the polled transmission mode, where each remote unit transmits its data requests only during the symbol period assigned to it.

Fig. 10 is a flowchart illustrating the steps taken by a requesting remote unit to establish communication with a central processing unit. As illustrated in FIG. 10, after the start in step 360, the method proceeds to step 362, where the requesting remote unit determines whether the transmission mode has a fast access or is polled. If the requesting remote unit determines that the polled transmission mode is currently busy, for example, in response to a control signal from the central processing unit, if system usage is strong, the method moves to step 366 to transmit data in the polled transmission mode. In the polled transmission mode, the requesting remote unit transmits its data request only on its one or more idle subchannels during the symbol period assigned to it.

On the other hand, if the requesting remote unit determines that the fast access transfer mode is currently busy, for example, in response to a control signal from the central processing unit, if system usage is low, the method proceeds from step 362 to step 364 for its communication access request during a symbol period on one or more unused subchannels. As previously explained, the requesting remote unit does not have to wait for the symbol period assigned to it to transmit its communication access request in the fast access transmission mode.

From step 364 or 366, the method moves to step 368 to determine if the data request is a data packet request (DPR). If so, the process goes to step 370, where the steps of Fig. 11 (a) are performed. On the other hand, if the data request is not DPR (determined in step 368), the method moves to step 372 to determine if the data request is a defined data packet request (DDPR). If the data request is a DDPR, the process goes to step 374, where the. Steps of Fig. 11 (b) are performed. On the other hand, if the data request is not DDPR (determined in step 372), the method moves to step 376 to determine if the data request is a data rate request (DRR). If the data request is a DRR, the process goes to step 378, where the steps of Fig. 11 (c) are performed. If the data request is none of the above, the process goes to step 380, where the steps of Fig. 10 end. It should be understood that certain embodiments may include additional data request types and that the method may be adapted to handle these additional data requests as needed. Adapting the disclosed method to handling specific additional data request types is within the skill of the art with the present disclosure.

With reference to Fig. 11 (a), a method of handling a data packet request will be described in more detail. First, central processing unit 10 allocates the next available S1 time interval 34 to the requesting remote unit and forwards a message confirming the allocation with the downlink signal (step 204). Then, in step 206, the requesting remote unit transmits the additional information during the associated S1 time interval 34. The additional transmission requests may include, for example, the address to which the data is being sent, the packet size, and a priority rating. As previously discussed, the remote unit may alternatively transmit the additional transfer requests in the same symbol period as the transfer request.

The CPU 10 then stores the additional received data packet information in step 208. CPU 10 then determines the number of subchannels that should be allocated for the requests from the remote units and returns instructions regarding the subchannels to be used along with the allowed bit rates per channel to the requesting remote unit. It should be noted that the central unit 10 allocates subchannels 23 based on the stored set of channel properties corresponding to the requesting remote unit 210. In this way, the central processing unit 10 can dynamically allocate the most effective number of subchannels 23 to handle the request of the remote unit. It should be noted that the receiver of the central unit knows the amount of data to be transmitted (based on the information received during the S1 rest period) as well as the data transmission rates (which has specified the remote unit). Therefore, the CPU knows the time required to complete the transfer. Accordingly, CPU 10 allocates the specified number of sub-channels 23 only during the requesting remote unit's time required for the requesting remote unit to transmit its packet (s). After elapse of the specified time period (with a required buffer), the CPU 10 notifies that the sub-channels 23 assigned to the first remote unit are now unused and ready to be reassigned to another remote unit (step 212).

With reference to Fig. 11 (b), a method for handling a defined data packet request (DDPR) will be described. For a defined data packet request, the CPU must trust that the additional data packet defines information stored in step 208. These may in turn include such things as the address to which the packet (s) were sent, and the packet size. Accordingly, a defined data packet request in the described embodiment can only be handled if it is transmitted by a remote unit that has previously sent a DPR. In alternative embodiments, appropriate setpoints could be provided to allow the use of defined data packets even if no data packet request has been sent.

As shown in Figure 11 (b), in step 223, the central processing unit accesses the stored defined data packet transmission requests and uses this information in routing and / or handling the received data packet (s). It should be noted that in the same symbol period or during a S1 time interval 34, the CPU 10 need not receive any additional information, and therefore may immediately associate one or more subchannels 23 with the requesting remote unit at step 225. Again, because the amount of information to be transmitted and the data transfer rates are known, the central processing unit allocates the subchannels only for the period of time required to transmit the packet. After elapse of the appropriate transmission time, the CPU 10 notifies that the sub-channels 23 are free to be reassigned at 227.

While many communication devices can effectively communicate through packetized communications, others require a constant transmission rate, which is sometimes difficult to obtain using packetized transmission systems. These remote units may be enabled by allocating a number of subchannels 23 sufficient to handle the required data transfer rate over an indefinite period of time. This is true until the remote unit indicates that the bandwidth is no longer required or until an error is detected. For example, it is likely that such demands will occur in videoconferencing. In the described embodiment, this type of data transfer request is handled through the use of a data rate request.

Next, with reference to Fig. 11 (c), a method suitable for handling data rate requests will be described. The central unit 10 typically requests additional transmission information, such as the address and the requested data rates, upon receipt of a DRR request. Accordingly, in step 252, the CPU allocates the next available S1 idle period to the requesting remote unit to send the required information. The requesting remote unit then transmits the additional transmission information during the associated S1 time interval in step 254. As previously discussed, the remote unit may alternatively transmit the additional transfer requests in the same symbol period as the transfer request.

If it knows the data rate requirements as well as the allowed data rates for each subcarrier, the CPU 10 allocates an appropriate number of subchannels 23 in step 256 to handle the requested throughput. If the requesting remote unit no longer needs to transmit, it sends in step 258 a new data rate request indicating that a capacity of zero is required. The central processing unit 10 understands this as a termination request and marks the appropriate sub-channels in step 260 as idle.

There is no fixed period of time, which is ideal for repeating the S1 rest periods. On the other hand, the access times that can be achieved for the requested transmission mode or for DPR and DRR requirements are shorter, the more often the S1 rest periods occur. The system responds accordingly the better. On the other hand, more frequent S1 rest periods require more overhead, thereby reducing overall system capacity. Accordingly, the appropriate frequency of the S1 periods will vary somewhat depending on the needs of a particular system. In the illustrated embodiment, the S1 rest periods are used to limit the frames, although it should be understood that this is not a requirement. In general, the use of the S1 rest periods reduces the access time required to initiate a communication. If appropriate, the access time of the requesting remote unit can be further reduced by using DDPRs.

As described above, the initialization time intervals S2 and the retraining time intervals S3 are not as numerous as the S1 rest periods because initialization and retraining usually do not require as fast a response as a request for immediate transmissions. In one embodiment, the S2s and S3s alternate in every other superframe 36. In another embodiment, the S2s and S3s may be dynamically assigned by the central processing unit 10 to accommodate changing circumstances. For example, at times when remote units are more likely to be installed and require initialization, for example during the day, more of the reserved time intervals 38 may be allocated as initialization time intervals. During the evening, when installations are less likely, more of the reserved intervals 38 may be allocated as retraining time intervals.

Next, with reference to Fig. 3, a central office architecture suitable for realizing the described synchronization and coordination will be described. In the illustrated embodiment, the central unit has a central modem 30, a network server 19 and a network interface 41. The central modem has a transmitter 40, a receiver 70 and a control device 60. The controller 60 is used to synchronize the far modem clocks with the clock in the central modem, as well as to synchronize the frames transmitted by the remote modems. The network server 19 provides digital data to the transmitter 40 via an asynchronous transmission modem switch 41 (referred to in the drawing as a network interface). The network server 19 may provide data at any data rate up to the maximum data rate taking into account the capacity of the transmitter, the transmission link, the transmission line quality and the type of transmission line used. The transmitter 40 has a plurality of components including an encoder 43, a discrete multi-tone modulator 45, and a window filter 46. The encoder 43 is for multiplexing, synchronizing and encoding the data to be transmitted (in the manner of video data). In particular, it translates incoming bitstreams into phase and quadrature components for each of a plurality of subchannels. The coding may be done using forward error correction and / or trellis coding. The encoder is typically arranged to output a number of sub-symbol sequences, the number of which equals the number of sub-channels available to the system. For example, in a 256 sub-channel system, the encoder outputs 256 sub-symbol sequences. In the ATIS standard mentioned above, the sub symbol sequences each represent 4 Kbps. These inputs are complex inputs supplied to a discrete multi-tone modulator 45. A suitable coder is described in detail, for example, in the mentioned ATIS standard.

The modulator 45 is an IFFT modulator that calculates the inverse Fourier transform by a suitable algorithm. A suitable IFFT coder is described in the article "Multicarrier Modulation: An Idea Whose Time Has Come" by J. Bingham, IEEE Communication Magazine, May 1990. Because the coder outputs are complex numbers, the IFFT modulator receives twice as many inputs as subchannels are available. The bit distribution is adaptively set in discrete multi-tone systems. To facilitate this, the transmitter 40 also has a line monitor which monitors the communication line to determine the line quality of each of the available subchannels. In one embodiment, the line monitor (which may be part of the controller 60) determines the noise level, the single gain, and the phase shift on each of the subchannels. Typically, this line monitor is used to detect the quality of the described S3 retraining signals. The task is to estimate the signal-to-noise ratio for each of the subchannels. Therefore, other parameters could be monitored in addition to or instead of the parameters described. It is dynamically determined based on several factors over which subchannels the coded data is to be transmitted and how much data is to be transmitted over each subchannel. The factors include the detected line quality parameters, the sub-channel gain parameters, an allowable power mask, and the desired maximum sub-carrier bit error rates. It should be noted that the various factors between sub-channels need not be constant and may actually change during use. In particular, the line quality parameters may be repeatedly tested, and adjustments made to the modulation scheme in real time to dynamically adjust the modulation as the line quality on multiple subchannels changes during use. For example, a suitable discrete multi-tone modulator is generally described in the same ATIS standard document.

After the coded signal has been modulated to form a discrete multi-tone signal, a cyclic prefix is ​​appended to the discrete multi-tone coded signal. The cyclic prefix is ​​used primarily to simplify the demodulation of the discrete multi-tone signals and, strictly speaking, is not required. The ATIS standard uses a 32-bit cyclic prefix. However, in systems where larger bandwidths are used, it would be preferable to increase the length of the cyclic prefix as well. For example, it has been found that with a 512 sample signal, a 40 sample cyclic prefix works well.

The modulated signal is then passed through a window filter 46 and / or other filters to minimize out-of-band energy. This is desirable to help prevent the analog interfaces in the remote receivers from saturating. The windowing can be achieved by a wide variety of conventional windowing protocols. The transmitter also has an analog interface 48 which supplies the discrete multi-tone signal to the transmission media. For hardwired systems, such as twisted pair telephone lines and coaxial cables, the analog interface may take the form of a line driver.

The central modem 30 also has a receiver 70 for receiving multi-tone signals from the remote units. The receiver 70 has an analog interface 72, a window filter 74, a demodulator 76 and a decoder 78. Signals received by the central modem 30 are first received via the analog filter 72. The window filter 74 is adapted to perform substantially window and / or filter functions on the received signal. A suitable filter arrangement is a time domain equalizer 74. Again, windowing can be accomplished by a wide variety of conventional window protocols. The demodulator 76 demodulates the balanced discrete multi-tone signal and extracts the cyclic prefix. The decoder 78 decodes the demodulated signal. The demodulator 76 and the decoder 78 essentially perform the inverse functions of the modulator 45 and the encoder 43, respectively. The decoded signal is then transmitted via the interface 41 from the decoder 78 to the network server 19 or other suitable user of the information. The functions of the time domain equalizer 74, demodulator 76 and decoder 78, as well as algorithms suitable for achieving the desired functions, are all described in more detail in US-A-5,285,474 to Chow et al.

Next, referring to Fig. 4, a remote unit architecture suitable for realizing synchronization according to the present invention will be described. In many respects, the remote modem is similar to the central modem, although the respective uplink and downlink communication capacities may be somewhat different. A signal transmitted from the central modem 30 is received by a remote unit 50 via an analog filter 172. The remote unit 50 includes the analog interface 172, a time domain equalizer (TEQ) 174, a demodulator 176 that demodulates the equalized discrete multi-tone signal and extracts the cyclic prefix, and a decoder 178 that decodes the demodulated signal. The time domain equalizer 174 essentially performs filter functions on the received signal. It is also possible to use a window filter. The demodulator 176 and the decoder 178 perform inverse functions of the modulator 45 and the encoder 43, respectively. The decoded signal is then transmitted from the decoder 178 to a remote device 22 such as a television, computer, or other suitable receiving device. The functions of the time domain equalizer 174, the demodulator 176 and the decoder 178 are similar to the functions of the corresponding components in the central modem. A notch filter 185 may optionally be provided above the analog filter 172 of the receiver to block energy in frequency bands outside the subchannels of interest to the remote unit. This can help prevent the saturation of the analog filter. By providing an analog of a blocking filter or other suitable filtering mechanism to filter out the band energy, more cost-effective receiver components can be used because it does not require the receiver itself to handle so much energy.

Up-coding and modulation can be done in exactly the same way as the down-data transmission described above in the discussion of the central modem unit. Accordingly, the remote modem 50 also includes an encoder 143, a multi-tone modulator 145, a window or filter 146, and an analog interface 148. It also requires a frame synchronizer 147 for delaying the multi-tone signals by an amount suitable for synchronizing the remote modem 50 with other remote modems currently communicating with the central modem, as described above. In subscriber applications, typically a smaller number of subchannels are made available to enable uplink communication. It should be noted, however, that any number of subchannels could be made available for this uplink communication.

For example, if a polled transmission mode is effective, the encoder 143 may be a QAM encoder. For example, in many systems, a QAM encoder with a 16-point constellation works well. For example, if the transmission is over a high-speed transfer mode, the encoder 143 may be a four-phase differential-rate (DQPSK) coder having a four-point constellation. A suitable DQPSK coder is described, for example, in the text by J. Bingham entitled "Theory and Practice of Modem Design" published by J. Wiley & Sons (1988). In the mode switching scheme described, the control signal for effecting the switching between the requested transmission mode and the fast access transmission mode is also input to the encoder, although it should be understood that it could alternatively be added elsewhere. Similarly, if the polled transmission mode is in effect, the decoder 78 at the central processing unit may be, for example, a QAM decoder. For example, if the transmission is through the fast access transmission mode, the central unit decoder 78 may be a differential four-phase shift decoder (DQPSK decoder).

Most of the embodiments described above were directed primarily to the manipulation of uplink communication from the remote units to the central unit 10. Accordingly, no restrictions are imposed on the type of downlink communication usable for such a system. The downlink channel may use discrete multitone modulation, similar to the modulation used for uplink communication, or it may use other suitable techniques such as vestigial sideband modulation (VSB modulation) or QAM. Further, the downstream channel may consist of dedicated additional channels for transmitting the relevant formatting signals, including, but not limited to, the S1, S2, and S3 flags, synchronization signals, and subchannel 23 assignment information. As will be understood by those skilled in the art, many other methods of transmission schemes can be applied to the downstream channel in accordance with the present invention.

When the discrete multitone transmission is used in both the uplink and downlink data directions and the desired data transmission rates are relatively high, it may be desirable to include a data transmission scheme based on time division multiple access (ie, a "ping-pong data transmission scheme") , That is, the downlink communication is given a fixed number of frames or superframes for transmission over the entire bandwidth. Thereafter, uplink communication is given a fixed number of frames or superframes for transmission over the entire bandwidth. For many high-data-rate applications, such as 25.6-bit and 51.2-million-bit-per-second applications, using the ping-pong transmission scheme in the transmitter and receiver designs offers significant cost savings because it eliminates the need for costly ones Provide filters to isolate simultaneous uplink and downlink transmissions. The ping-pong method is particularly advantageous for data rates above 10 million bits per second.

Next, a ping-pong transmission scheme for an asymmetric application will be described with reference to FIG. In this embodiment, eight consecutive down-superframes (DSFs) 885 are downlinked from data and then an uplink superframe (USF) 886 is uplinked from data. In other embodiments, the actual number of frames used for transmission in each direction may be changed according to the requirements of a particular system. For example, the asymmetry ratio could be widely varied in favor of the downlink communication, the transmission periods could be symmetrical, or the uplink communication could be given better access. In systems that authorize dynamic allocation of bandwidth between the uplink and downlink communications, a controller may be provided to dynamically allocate the distribution of the frames between the uplink and downlink communications. In systems where the signals between the central processing unit and the remote units are over relatively long distances, it may be desirable to provide a fade-down period 887 after the end of data transmission in one direction to facilitate transient decay. In the illustrated embodiment, a fade period is provided after the uplink transmission but not after the downlink transmission. In practice, the settling period 887 may be appropriate after one-way or both-way transfer.

It should be understood that the initialization and / or synchronization of the remote units, the uplink subchannel access requests, and / or the training intervals may be accomplished using any of the techniques discussed above. The main advantage of the described time division multiple access method is that it eliminates the need for costly filters to isolate simultaneous uplink and downlink communications in systems where the transmission bandwidth is likely to be high. Another advantage is that the Ping-Pong transmission method can improve the asymmetric transmission as compared to the standard frequency multiplexing in which the uplink transmission is in a first frequency range and the downlink transmission is in a second frequency range. In fact, transfer rates can be increased to almost the level of a system that uses echo cancellation. However, the ping-pong method may achieve these transmission rates at much lower cost of the analog components (using current technology) than would be required to use a frequency division multiplexing system or an echo cancellation system.

Although only a few embodiments of the present invention have been described in detail, it should be understood that the present invention may be embodied in many other specific forms without departing from the spirit or scope of the invention. For example, the invention has been described primarily in the context of a discrete multi-tone transmission system. However, it should be understood that the same techniques can be applied to other discrete multicarrier systems, such as discrete wave multi-tone, vector-coded, and other multi-carrier modulation schemes. It is also to be understood that in embodiments having supplemental subchannels, these subchannels may be shared or separate for each direction. The use of two subchannels in the auxiliary bus has been described in great detail. However, it should be understood that a single subchannel could also be provided for uplink and downlink communications (particularly when echo cancellation is used). Alternatively, more than two additional subchannels may be provided if the constraints of a particular system dictate that more than one subchannel should be used to communicate in one direction (or both directions). For example, in a system with a relatively small number of remote units, each remote unit (or subgroup of remote units) could be assigned a dedicated subchannel. Alternatively, redundancy could be provided to reduce the risk of noise interference. The same could apply to the downside additional communication. Of course, the disadvantage of using dedicated subchannels for each remote unit is that bandwidth is wasted. Furthermore, dedicated additional sub-channels are described. However, in some circumstances, it would be possible to multiplex other overhead information (e.g., control information) onto the same subchannel. It should be understood in light of the foregoing that the present examples should be taken as illustrative and not restrictive and that the invention is not limited to the details given herein but may be modified within the scope of the appended claims.