DE19911438A1 - Modulation method for digital transmission in narrow band transmission channels by transferring test sequences as single carrier and selecting a number of channels so that length of symbols in it equal to length of test sequences - Google Patents

Abstract

A data stream is transferred using a small number of carriers. Periodic test sequences are embedded in the data stream. The test sequences are transferred as a single carrier method and that a number of the channels is selected in such a way that the length of the symbols in the channels equal to the length of the test sequences. The selected channel should be that, in which a spectral blackout arises.

Description

The invention relates to the field of digital signal transmission in narrowband Transmission channels.

State of the art

Digital transmission methods in narrow-band transmission channels find their way Use in military modems, especially on shortwave (see Kotolowski, Brakemeier et al .: High Data Rate Modem: Comparison of Serial and Parallel Waveforms for Shortwave Communications, Nordic HF 1998), as well as in civil applications at the Digital transmission in the long, medium and short wave ranges. This transmission channels are often subject to selective fading. Can also pass through in cable ducts Echoes of selective fading occur.

The digital transmission in radio channels takes place e.g. Currently worldwide interest. This is documented e.g. B. in the establishment of the international consortium "Digital Radio Mondiale "(DRM, http://www.drm.org) and in parallel activities in the ITU (International Telecommunication Union).

Two modulation methods are common in military applications. These are one Single-carrier process and a multi-carrier process. In the broadcasting application, a globally uniform procedure can be determined by DRM or ITU. this could e.g. B. be a single-carrier or a multi-carrier method.

The single carrier method uses a data stream (user data, for example audio data and fast information data, FIC), in which "test sequences" are periodically embedded between the data symbols ( FIG. 1), with the aid of which, among other things, the receiver is rapidly synchronized and the transmission channel can be measured (channel estimation), from which the information on adaptive channel equalization (for the receiver) is obtained (see Rudolph et al .: patent application PCT / EP97 / 06042: method for interference signal suppression when transmitting digital signals).

The test sequences are for digital transmission in a radio application from elementary importance, because especially due to the mobile operation of the receiver in Auto a fast (re) synchronization of the receiver becomes necessary.

The test sequences are used in the receiver for the following purposes:

• - Synchronization of frequency, phase, timing
• - Frequency correction and frequency control
• - level control
• - channel estimation
• - Adaptive equalization
• - suppression of interference
• - Setting the channel bandwidth (9 KHz / 10 KHz)
• - error measurements, bit error measurements
• - Parameter setting for hierarchical modulation
• - Switching between analog (AM) and digital broadcasts

While the single-carrier method uses short-term symbols that fill the entire channel bandwidth in terms of frequency, the multi-carrier method uses long-time symbols that only fill in part of the channel bandwidth in terms of frequency ( FIG. 2). This part is called the "sub-channel". Since only a correspondingly smaller amount of data can be transmitted with slower data symbols (compared to the single-carrier method), the same amount of data must be transmitted in several parallel sub-channels at the same time, which explains the name "multi-carrier". To simplify matters, many very narrow subchannels are often used. This makes the data symbols very long; in particular, they can be selected many times longer than the longest occurring echo in the transmission channel.

In such a case (normal case for OFDM, "Orthogonal Frequency Division Multiplex") becomes the overlap area of the successive symbols in the receiver disappears, causing all effects caused by echoes to disappear.

This fade-out time is called the "guard interval" (referred to in FIG. 2 as "GAP" = "gap"). The signal energy of the received signals is reduced by the portion from the gap. This technique eliminates the need for equalization in the subchannels. Each of these subchannels only suffers from flat fading. On the other hand, the effort is now in the necessary FFT (Fast Fourier Transformation) and in the indispensable coding. The synchronization and the channel estimation, which is also necessary for higher-level data symbols, are more complex than with the single-carrier method, in which a particularly simple and robust type is available by means of the test sequences.

While with analog modulation methods it is only possible to intervene in the Amplitude (→ amplitude modulation) or frequency (→ frequency modulation) or the phase (→ phase modulation) of the carrier, there is a digital modulation almost unlimited variety of intervention options or parameters open. Here it turns out that a minimum of effort is required in the receiver when all available parameter values are used to a reasonable extent.

So in the case of the one-carrier method, echoes mean very long in relation to the Symbol duration is a disproportionately increasing effort in the length adaptive equalizer. On the other hand, the size of the necessary FFT is the same Multi-carrier process directly dependent on the number of subchannels. Furthermore, the Synchronization and channel estimation are very complex here. Through the guard period you also lose part of the transmission capacity.

The invention has for itself the task of digital transmission in narrowband To offer a modulation method for the transmission channels that reflects the properties of Single and multi-carrier processes advantageously linked.  

A modified multi-carrier process with only a few subchannels is used. The data stream is structured as in the single-carrier process. The subchannels are like this kept wide that "flat fading" can no longer be assumed. It is per subchannel Adaptive equalization is required, but with much less effort manages than with the one-carrier process.

Because the effort in the equalizer is related to the ratio of the length of the echoes to the length of the Symbols increase disproportionately (square to cubic, depending on the type) a reduction in effort.

For example, a length of 100 is assumed for an equalizer. The (computational) effort is thus assumed to be 100,100 = 10,000. With 10 subchannels, the symbols are each a factor 10 longer and the equalizer effort only has to be set to 10, which means an effort of 10.10 = 100 per equalizer results. Since an equalizer is required for 10 subchannels, the total effort is 1000, so here only 10% compared to the single-carrier method. In the case of a cubic dependency, the numerical values are 10 ^∧ 6 to 10 ^∧ 4, i.e. only 1%. Such optimistic numerical values, however, only arise with very "bad" shortwave channels with long echo durations.

The modulation method according to the invention is characterized as follows:

• 1. The data stream (or symbol stream) is structured according to Single-carrier procedure, d. H. Periodic test sequences are included in the user data stream embedded.
• 2. The user data are transmitted as multi-carrier modulation with a few subchannels, which makes the symbols in the subchannels shorter. An adaptive equalization becomes necessary in each case, a guard period (gap) is not required.
• 3. The test sequences are transmitted as single-carrier modulation. This leaves everyone Get (above) advantages of the test sequences, especially the fast ones Synchronization of the receiver and the channel estimation.  
• 4. The overall channel is divided into subchannels, using window functions (Windowing) can be done. This gives you the information for the adaptive Equalization of the individual subchannels.
• 5. The test sequence can also be in turn in subchannels as multi-carrier modulation the user data are transmitted.
• 6. The number of subchannels is chosen so that the length of the symbols in the Subchannels corresponds to the length of the test sequence. The FFT definitely has that same length.
• 7. It is advantageous to use an even number of subchannels because of this there is a gap in the middle of the band. Thus, carriers of AM-Sen act that can occur in the same channel during a transition period, not from.
• 8. The Fast Information Channel (FIC) remains intact. The following properties can advantageously be achieved in this way:
  • - Channel for the first access and for quick information (such as station name, program name, etc.)
  • - Monitoring of alternative reception frequencies and measurements of their frequency, phase, timing, level
  • - Suppression of spectral interference components.

Claims (4)

1. Modulation method for digital transmission in narrow-band transmission channels with selective fading, characterized in that
that the data stream is transmitted as a multi-carrier method with few carriers,
that periodic test sequences are embedded in the data stream,
that the test sequences are transmitted as a single carrier method, and
that the number of subchannels is selected so that the length of the symbols in the subchannels is equal to the length of the test sequences, the number of subchannels being an even number, so that a spectral gap arises precisely in the center of the channel. 2. Modulation method according to claim 1, characterized in that the data stream in addition to the audio data also contains fast information data and the symbols in the Fast information channel can be selected differently from the data symbols can. 3. Modulation method according to claim 1, characterized in that the Test sequences especially for the channel estimation for the subchannels and Calculation of the adaptive equalization can be used for the subchannels. 4. Modulation method according to claim 1, characterized in that the Test sequences can also be transmitted as multi-carrier methods.