KR100252011B1 - Data transmission method of tdd dmt system - Google Patents

Abstract

PURPOSE: A method for transmitting data in a DMT(Discrete Multi-Tone) system with a TDD(Time Division Duplexing) type is provided to improve efficiency in data transmission, by measuring an SNR(Signal-to-Noise Ratio) of a frame by which a NEXT(Near End Cross Talk Noise) can be created separately from SNRs of other frames, and then transmitting the data properly to the tables. CONSTITUTION: Some of received super-frames by which the NEXT noise may be created are separated from other frames(400). A channel PSD(Power Spectral Density) for each frame and a noise PSD are measured(410-440). The channel PSD and the noise PSD are measured until the super-frames are repeatedly by the times of K(450). The SNRs are calculated for each sub-channel, based on the channel PSD and the noise PSD(460,480). The first bit table and the second bit table are obtained(470,490), and thereby a channel analysis process is completed.

Description

Method of data transmission in time division duplex (DTD) system discrete multi-tone (DMT) system

The present invention relates to a DMT system, and in particular, in the TDD scheme, a time division duplexing (TDD) scheme for minimizing the effects of NEXT (Near End Cross Talk) occurring when asymmetric services and symmetric services exist together in the same cable. DMT) system data transmission method.

Very high bit rate Digital Subscriber Lines (hereinafter referred to as VDSL) are approximately 10 Mbps over telephone lines with a shorter distance (300-1500m) than for asymmetric Digital Subscriber Lines (hereinafter referred to as ADSL). -This technology transmits high speed data over 50Mbps. Unlike ADSL, which uses frequency division duplexing (hereinafter referred to as FDD) when duplexing upstream and downstream, time division duplexing (hereinafter referred to as TDD) in VDSL. It is strongly suggested that the method be used. In the U.S. T1E1.4, which is working on VDSL standardization, Carrierless Amplitude and Phase Modulation (CAP) and Discrete MultiTone (DMT) methods are reviewed as line codes. In other words, those advocating the CAP method prefer the FDD method.

)을 갖는 독립된 반송파 신호들의 합으로 구성된다. 전송로의 특성이 좋지않은 전화선같은 채널로 고속의 데이타를 전송하고자 할 때에 DMT 시스템을 이용하면, ADSL응용의 경우 6Mbps급 이상의 고속 데이타 전송이 가능하며, VDSL(Very high bit rate Digital Subscriber Lines) 응용의 경우에서도 전송로의 거리 및 특성에 따라 10 - 50Mbps 이상의 고속 데이타 전송을 할 수 있다. 특히 RFI(Radio Frequency Interference) 잡음이 많은 고주파 영역에서는 그 부근의 부채널을 사용하지 않을 수 있으므로 이러한 각종 장애(impairments )에 강한 시스템을 구현할 수 있는 장점이 있다.The DMT system transmits data using multiple carriers (subchannels) to maximize channel efficiency. That is, the transmission signal using the multi-carrier has a center frequency of each subchannel f _i, i = 1,2,3, ..., N (N is the number of subchannels), and the same bandwidth (

It consists of the sum of independent carrier signals with When using DMT system to transmit high speed data through a channel such as a telephone line, which has poor transmission line characteristics, high speed data transmission of 6Mbps or more is possible for ADSL application, and VDSL (Very High Bit Rate Digital Subscriber Lines) application Even in the case of high speed data transmission of 10-50Mbps or more depending on the distance and characteristics of the transmission line. In particular, in the high frequency region where the RFI (Radio Frequency Interference) noise is high, the subchannel in the vicinity thereof may not be used, and thus, a system that is resistant to such various obstacles may be implemented.

When using TDD in a VDSL DMT system, the first thing to do is Synchronized DMT (SDMT), which synchronizes all remote transmitter clocks with ONU (Optical Network Unit) clocks and synchronizes all frames within the same cable. . 1 illustrates a superframe structure of a TDD type SDMT. As shown in FIG. 1, a superframe is composed of 20 frames, and each frame is composed of a DMT symbol which is an execution unit of fast fourier transform (FFT). Mds represents the number of downstream frames transmitted downstream from the ONU to the remote terminal, Mus represents the number of upstream frames transmitted upstream from the remote terminal to the ONU, and S represents the line-over. turnaround) is a silent period in which no signal is transmitted. This SDMT has the following advantages.

1. The analog circuit requires only minimal filtering for smoothing at the transmit end and anti-aliasing at the receive end.

2. The transceiver can perform both modulation and demodulation using one FFT processor. That is, in FDD or EC canceling (Echo canceling) method, two FFT processors are required at the transmitting and receiving end.

3. You can easily change the data rate by adjusting the number of downstream and upstream frames (Mds, Mus) with software.

4. The silent period (S) without signal transmission can be used to measure and eliminate RFI.

When all superframes are synchronized in SDMT, Mds = 16, S = 1, Mus = 2, and S = 1 enables asymmetric services up to about 52 Mbps downstream and 6.4 Mbps upstream. In this case, there is no NEXT (Near End Cross Talk) noise because downstream and upstream occur in the same time zone within the same cable. However, according to the recent opinion that symmetric transmission of 26 Mbps or 13 Mbps in both directions is required, superframes consisting of Mds = 9, S = 1, Mus = 9, and S = 1 may exist in the same cable. In this case, some of the 20 superframes will be affected by NEXT. 2 illustrates a superframe structure for asymmetric service and symmetric service. The superframe structure for the asymmetric service is 16-1-2-1, and the superframe structure for the symmetric service is 9-1-9-1.

Thinking in ONU, NEXT occurs on the line where 9-1-9-1 is sent when downstream of 16-1-2-1 asymmetric service and upstream of 9-1-9-1 symmetric service are transmitted in the same time zone. 2, it can be seen that NEXT exists in the 1, 2, 14, 15, and 16th frames. That is, the 1,2,14,15,16th frames are downstream in FIG. 2A but upstream in FIG. 2B. The reason why two upstreams are first sent in FIG. 2B is to reduce the number of frames generated by NEXT.

In general, the DMT system always undergoes an initialization process before data transmission, and mainly performs timing recovery and various equalization training. In addition, the DMT system measures the signal-to-noise ratio (SNR) of each subchannel, and allocates and transmits a number of bits suitable for each subchannel. A subchannel having a high SNR is allocated a large number of bits and a subchannel having a bad SNR. Does not allocate or allocate a small number of bits.

In the case of VDSL, which has a shorter distance and wider frequency band than ADSL, NEXT is very fatal. If there is no NEXT during the initialization process and NEXT occurs during data transmission, one bit table is used, so the efficiency of the entire frame is very low. This is because the bit table value for the entire frame should be based on the frame in which NEXT exists. In addition, if the frame does not use NEXT, symmetric service can use only four frames (17, 18, 19, 20) out of nine upstream, the data rate is only 4/9 overall.

The present invention was created to solve the above-described problems, and the technical problem to be achieved by the present invention is to measure the SNR of the frame and other frames where the NEXT can be generated separately to make each bit table and transmit accordingly The present invention provides a data transmission method of a TDD-type discrete multi-tone system for improving data transmission efficiency.

1 illustrates a superframe structure of a TDD type SDMT.

2 illustrates a superframe structure for asymmetric service and symmetric service.

3 is a flowchart illustrating an initialization and data transmission process of the DMT system for ADSL.

4 is a flowchart illustrating a more detailed operation of the channel analysis process.

5 is a flowchart illustrating the operation of bit swapping used in the present invention.

6 (a) and (b) show a simulation graph of the effect of NEXT on the data rate.

In order to solve the above technical problem, according to the present invention, a method for transmitting data through the channel in a TDD DMT system for a VDSL having a transmitter for transmitting data through a channel and a receiver for receiving the transmitted data, Checking the preparation of signal transmission between stages and preparing timing recovery and various equalizers to be applied when transmitting and receiving signals; A channel analysis step of generating bit loading information to perform bit loading for each channel of the DMT system; A channel information exchange step of transmitting information on channels to each other in order to share information on the channel analyzed in the channel analysis step at a transmitting and receiving end; And a data transmission step of transmitting data by performing bit loading for each channel according to the bit loading information shared through the channel information exchange step.

The channel analysis step may include a first step of separating a frame that may be affected by a near end cross talk (NEXT) and a frame that is not; A second step of measuring channel power spectral density (PSD) and noise PSD for each of the frames that may be affected by the NEXT separated in the first step and those that are not; A third step of measuring a signal-to-noise ratio for each subchannel from the channel PSD and the noise PSD measured in the second step; And generating bit tables (first and second bit tables) having bit loading information for each subchannel according to the signal-to-noise ratio measured in the third step, for frames that may be affected by the NEXT and frames that are not. It is provided with a fourth step.

The superframe consists of 20 frames, and the superframe of the asymmetric service has 16 downstream, two upstream, and two silence frames, and the superframe of the symmetric service has nine downstream, nine upstream, and It has two silence frames.

On the other hand, if the channel status changes while transmitting data in the data transmission step, a bit swapping step of transferring the bits of the sub-channel having a bad signal-to-noise ratio to a sub-channel having a good signal-to-noise ratio.

And after performing the bit swapping step, checking whether a sum of bits of a bit table (first bit table) for a frame that may be affected by the NEXT satisfies a desired predetermined value; And performing the data transfer step if satisfied in the checking step, and performing the data transfer step after performing bit swapping between the first bit table and the second bit table if not satisfied. do.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. 3 is a flowchart illustrating an initialization and data transmission process of a DMT system for ADSL. As shown in FIG. 3, step 300 is an activation and acknowledgement step, which is a period of checking whether a signal is ready to be transmitted between a transmitter and a receiver, and step 310 is a transceiver training. In step), timing recovery and various equalizer training are performed. Reference numeral 320 is a channel analysis step, in which a signal-to-noise ratio (SNR) of each subchannel is measured to generate a bit table having bit loading information for each subchannel. This will be described with reference to. Reference numeral 330 denotes a step of exchanging channel information, and is a period of exchanging various parameters determined in the above process between the transmitting and receiving end.

Steps 300 to 330 are referred to as an initialization process 360, and data transmission starts after the initialization process is completed (step 340). At this time, each subchannel is determined by a bit table having predetermined bit loading information. In the case of transmitting bits suitable for SNR, the characteristics of the channel may be changed by various factors during data transmission. At this time, bit swapping is performed to change the bit table according to the change. The bit swapping serves to reduce the bits of the subchannel having a bad SNR and add the bits to the subchannel having a good SNR.

In the DMT system for ADSL using the FDD method, the SNR of each subchannel is measured at the remote side based on a training signal sent from the central office, bit-loaded accordingly, and the information is transmitted during the channel information exchange. Transfer to the central office for optimized data transfer.

On the other hand, in the VDSL, which prefers the TDD scheme, five frames (1) as described above, particularly when asymmetric (16-1-2-1) transmission and symmetric (9-1-9-1) transmission are performed in the same cable. , 2,14,15,16) are affected by NEXT. Therefore, ONU makes two bit tables by measuring SNR separately for NEXT-affected frame and NEXT-affected frame, and then transmits the bit corresponding to each bit table.

After this initialization process, the actual data is transmitted in the normal state. If the initial operation of the system has only asymmetric (or symmetric) service in the same cable and bitloading is performed in the absence of NEXT, then it operates in a normal state and then starts a symmetric (or asymmetric) service in the middle. In this case, the bit table is converted by bit swapping by a bit swapping algorithm.

In ADSL (one bit table) using FDD, bit swapping is performed within the same frame. In VDSL (two bit tables) using TDD, bit swapping between two bit tables can be performed in addition to bit swapping in the same frame. .

Meanwhile, FIG. 4 is a flowchart illustrating a more detailed operation of the channel analysis process 320, and the bit loading by channel analysis will be described in detail with reference to FIG. 4. The channel analysis process 320 is performed after the transceiver training process 310 of the initialization process, during which a training sequence is transmitted from the ONU to the remote direction. In the present invention, it is also necessary to transmit the training sequence from the remote to the ONU direction. The receiver (ONU or remote receiving end) first determines which of the 20 frames constituting the superframe is the frame being transmitted, and the 1,2,14,15,16th frame that may cause a double NEXT. And the remaining frames (step 400). The 1,2,14,15,16th frames may be changed according to the structure of the superframe, and thus the present invention is not limited thereto.

Next, the channel power spectral density (hereinafter referred to as PSD) and noise PSD are measured for each of them (steps 410 to 440). In this case, a DLS method is used to measure the channel PSD. For this purpose, it is preferable to use the WALSH method.

The above steps 400 to 440 are continued until the superframe is repeated K times (step 450). The reason for repeating K times is necessary for accurate SNR calculation, and the K value is usually about 50 to 150. . Here, the larger the K value, the more accurate the SNR can be calculated, but the disadvantage is that it takes longer.

After the superframe is repeated K times, the SNR is calculated for each subchannel from the measured channel PSD and the noise PSD (steps 460 and 480), and then the first bit table (1, 2, 14, 15, 16) is used. ) And the second bit table (from the rest of the frame), the channel analysis process 330 ends (steps 470 and 490), and the channel information exchange step 330 is performed.

Here, there is another matter to consider when determining the first and second bit tables. For example, if you want to perform symmetrical transmission (9-1-9-1) of 12.96Mbps, the 1, 2, 14, 15, 16th frame out of the nine superframes as shown in FIG. There is a possibility to receive. Since the transmission rate of one frame is 25us (40KHz), the bits to be transmitted per frame are 12.96Mbps / 40KHz (per frame) * 20/9 = 720 bits / frame. Therefore, in this case, the average number of bits of the 1,2,14,15,16th frame affected by NEXT and the 17,18,19,20th frame not affected by NEXT should be 720 bits / frame. It can be seen that 720-4 * n bits / frame is allocated to the bit table and 720 + 5 * n bits / frame (where n = 1,2,3 ...) are allocated to the second bit table.

6 (a) and 6 (b) show a simulation graph for this, and simulation conditions are as follows. 26 gauge and 24 gauge lines, the maximum number of bits of the subchannel is 11 bits, the noise considered is AWGN (Additive White Gaussian Noise), 12 FEXT (Far End Cross Talk), RFI (Radio Frequency Interference), 6dB noise margin On the assumption of (noise margin), the transmittable data rate when NEXT number is 0-12 is expressed as a function of distance. As can be seen in FIG. 6, as the distance increases and the number of NEXTs increases, the data rate that can be transmitted decreases, and when there is no NEXT, data of 12.96 Mbps at a distance of 1Km or more and N700 in the presence of NEXT is about 700-800m. It can be seen that transmission is possible.

5 is a flowchart illustrating the operation of bit swapping used in the present invention. If the channel condition changes due to various factors during the data transmission after the initialization process 500, the SNR value of each subchannel may also change, so that the initially determined bit table may not be optimal. Therefore, in this case, bit-swapping is referred to as shifting bits of the subchannels with poor SNR one bit into the subchannels with improved SNR. In ADSL using FDD, there is only one bit table, so you can do bit swapping in it.

However, the case of VDSL using TDD is different. If the subchannel condition is changed due to a factor during data transmission (step 510), bit swapping is performed first within the frame (step 520). However, the data rate with and without NEXT is shown in FIG. As very large. If there is no NEXT during initialization and a NEXT occurs during data transmission, the sum of the number of bits of the first bit table may not satisfy a desired predetermined value (step 530), that is, the first bit table (1, 2, 14). No matter how bit swapping occurs within the 15,16 bit table), the average 720bits / frame cannot be matched. In this case, the bit of the first bit table may be subtracted and added to the second bit table (step 540). In the case of the symmetric transmission (9-1-9-1) of the previous example, nine upstreams (1, 2, Among 14, 15, 16, 17, 18, 19, and 20, five frames are affected by NEXT, and four frames are not affected by NEXT. Therefore, the total data rate is matched by subtracting bits in 4 * n units from the first bit table and adding 5 * n bits to the second bit table (where n = 1,2,3, ...).

According to the present invention, a bit table of a frame affected by NEXT and a frame not affected by NEXT is made for each situation.

In addition, since the bit swapping that sends a bit that cannot be sent in the NEXT-affected frame in a frame that is not affected by NEXT can be performed, there is much gain in overall efficiency.

Claims (7)

1. A method for transmitting data through a channel in a TDD DMT system for a VDSL having a transmitter for transmitting data through a channel and a receiver for receiving data, the method comprising: Checking a signal transmission preparation between the transmitting and receiving ends, and preparing timing recovery and various equalizers to be applied when transmitting and receiving signals; A channel analysis step of generating bit loading information to perform bit loading for each channel of the DMT system; A channel information exchange step of transmitting information on channels to each other in order to share information on the channel analyzed in the channel analysis step at a transmitting and receiving end; And And a data transmission step of transmitting data by performing bit loading for each channel according to the bit loading information shared through the channel information exchange step. The channel analysis step A first step of separating the frame that may be affected by the NEXT (Near End Cross Talk) of the super frame and the frame that is not; A second step of measuring channel power spectral density (PSD) and noise PSD for each of the frames that may be affected by the NEXT separated in the first step and those that are not; A third step of measuring a signal-to-noise ratio for each subchannel from the channel PSD and the noise PSD measured in the second step; Generating bit tables (first and second bit tables) having bit loading information for each subchannel according to the signal-to-noise ratio measured in the third step, for a frame that may be affected by the NEXT and a frame that is not; And a fourth step, a data transmission method of a TDD-type discrete multitone system. The method of claim 1, wherein the superframe 20 frames, Superframe of asymmetric service 16 downstream, 2 upstream and 2 silence frames, The superframe of symmetric services A data transmission method of a TDD-type discrete multi-tone system characterized by having nine downstream, nine upstream, and two silent frames. The frame of claim 2, wherein the frame may be affected by the NEXT of the second step. A data transmission method of a TDD-type discrete multitone system, characterized in that the 1,2,14,15,16th frame. The channel PSD of claim 1, wherein Measured by DLS method, The noise PSD is A data transmission method of a TDD-type discrete multitone system, characterized by a Walsh method. The method of claim 2, And repeating the first and second steps a predetermined number of times and performing a third step. The method according to claim 1 or 5, In the data transmission step, if the channel conditions change, the data of the TDD-type discrete multi-tone system further comprises a bit swapping step of transferring the bits of the subchannel having a bad signal-to-noise ratio to the subchannel having a good signal-to-noise ratio. Transmission method. The method of claim 6, After performing the bit swapping step, checking whether a sum of the number of bits of a bit table (first bit table) for a frame that may be affected by the NEXT satisfies a desired predetermined value; And Performing the data transfer step if satisfied in the checking step; and if not satisfied, performing the data transfer step after performing bit swapping between the first bit table and the second bit table. A data transmission method of a TDD-type discrete multitone system, characterized in that.

Images