The use of multi-position quadrature amplitude modulation (QAM) is associated with the problem of insufficient noise immunity. Therefore, in modern high-speed protocols, QAM is used in conjunction with trellis coding - a special type of convolutional coding. The result was trellis modulation.
Signal-code structures in modems.
A combination of QAM and an error-correcting code selected in a certain way refers to a signal-code structure (SCC). SCC allow increasing the noise immunity of information transmission systems along with reducing the requirements for the signal-to-noise ratio in the channel by 3-6 dB. In this case, the number of signal points is doubled by adding one redundant one formed by convolutional coding to the information bits. The block of bits expanded in this way is subjected to QAM. In the process of demodulation, the received signal is decoded using the Witsrby algorithm.
Modern communication systems are modem technologies implemented for specific speeds and volumes of information transfer. Information about the internal structure and architecture of recent models of modems is not as readily available as, for example, information about the structure of computers. One reason is that there are no industry standards for modem design. Another reason is that modems are built on specialized chipsets that provide basic functionality. This leads to the fact that in modems of different design, the same methods and protocols are implemented in different ways. Almost all modern modems have similar structural diagrams containing (Fig. 7.31):
- channel ports and DTE - DCE interfaces;
- main or general-purpose processor (PU);
- digital signal processor (DSP);
- modem processor (MP);
- read-only memory (ROM, ROM)
- reprogrammable storage device (EPROM, ERPROM);
Rice. 731.
- random access memory (RAM, RAM);
- diagrams of modem status indicators.
The abbreviation DTE (data terminal equipment - data terminal equipment) denote digital terminal devices generating or receiving data. The abbreviation DCE (data communication equipment - data transmission equipment) denote modems. Communication line in DCE - analog, between DCE and DTE - digital.
DTE interface port - DCE provides interaction with DTE. If for communication DTE and DCE using a unified digital interface, this often makes it possible to connect two adjacent DTE direct digital line - the so-called null-modem cable. In case of exploding DTE over a long distance in the gap, instead of a null modem cable, a pair of modems and an analog communication line are included, providing transparent connection and data transmission.
Universal processor performs the functions of managing interaction with DTE and modem status indication circuits. It executes the sent DTE AT- commands (A Г-commands are sequences of characters starting with Latin letters AT; they are used to configure and control the operation of the modem) and controls the modes of operation of the remaining components. The general-purpose processor is responsible for receiving and executing commands, buffering, processing some data, and also controlling the signal processor. The intellectual capabilities of the modem are determined by the type PU and control firmware stored in ROM.
The DSP is responsible for the implementation of the basic functions of modulation protocols (convolutional coding, relative coding, scrambling, decoding, compression / decompression).
MP deals with modulation / demodulation, frequency division, echo cancellation, etc. Depending on the complexity of the modem, the intellectual load is shifted towards the MP. In low-speed (300-2400 bit / s) modems, the main work is performed by the MP, in high-speed (4800 bit / s and higher) modems - by the DSP.
The ROM stores programs for the main and signal microprocessors (SMP), as well as the control microprogram - firmware, which includes sets of commands and data for controlling the modem. ROM can be one-time programmable (PROM), reprogrammable with UV erasure ( EPROM) or electrically reprogrammable ( EEPROM). The latter type of ROM allows you to quickly change firmware as errors are corrected or new features appear. By replacing or reprogramming ROM sometimes it is possible to achieve a significant improvement in the properties of the modem, i.e. make its modernization (upgrade). This type of upgrade for some modem models may provide support for new protocols or service features such as automatic caller ID (Caller ID). To facilitate such modernization, recently instead of microcircuits ROM flash memory chips ( FLASHROM). Flash memory allows you to easily update the firmware of the modem, fixing errors of the developers and expanding the capabilities of the device.
EPROM allows you to save the modem settings in the so-called profiles (profiles) at the time of shutdown. Most of the state change commands affect only the current set of parameters, which lose their values when the modem is turned off or reset. The contents of the current set can be written to one of the previously saved sets in the EPROM; in addition, a number of commands can directly change the contents of the EPROM. There are usually two settings to be saved - the main (profile 0) and additional (profile 1). By default, the main set is used for initialization, but it is possible to switch to an additional one.
RAM is used extensively for temporary data storage, compression algorithms, and intermediate computations for both general purpose and digital signal processors. RAM also stores the current set of modem parameters (active profile).
A modern telephone modem also has an analog part responsible for interfacing the modem with the telephone network - a dialer, amplifier, ADC and DAC. Almost all telephone (and other) modems process information in digital form, without any complex analog preprocessing, since this allows achieving high stability and greatly simplifying the development and analysis of algorithms. In this case, the sampling frequency is usually in the range of 7-12 kHz or more. The number of quantization levels for DAC and ADC of modern modems reaches tens of thousands. Usually, since the DAC and ADC are written or read as a number on the "digital side", they talk about the number of bits in the DAC / ADC, i.e. the number of bits of a binary number required to represent all possible levels, for example, a 16-bit ADC can recognize 65,536 levels, denoted by numbers from 32,768 to +32,767.
Classification of modems. Currently, there is no strict classification of modems. Nevertheless, a number of their classification features can be distinguished: field of application; functional purpose; channel type; support for modulation protocols, error correction and data compression. Modems are also divided by data transmission rates (14,400 bit / s, 28,800 bit / s, 33,600 bit / s, 56 Kbps).
By Areas of use modems can be divided into a number of main groups:
- for cellular communication systems;
- radio modems;
- for packet radio networks;
- for local radio networks;
- satellite;
- for digital data transmission systems (ISDN);
- cable;
- for dedicated telephone channels (?> 5X modems);
- for fiber optic lines.
Cellular modems are used for mobile radiotelephony. These modems do not contain a radio station (unlike radio modems), but only transmit their signal to it. Modems for cellular communication systems are distinguished by their compact design and support for special modulation and error correction protocols, which allow efficient data transmission in conditions of cellular channels with a high level of interference and constantly changing parameters. When crossing the border of cells (in the case of cellular communication), there is a switch to another radio station and the associated temporary loss of the signal. Most conventional modems under these conditions either try to re-establish the connection or break it, which is wrong. Due to the reflection of the signal from the buildings, several signals come and their superposition, the signal is distorted or even periodically disappears. It is clear that such work requires special protocols. Leading manufacturers supply this type of modem.
Radio modems use free space as a signaling medium. Therefore, instead of a telephone connector, the radio modem has an antenna connector where an antenna or antenna cable is inserted. In addition, the radio modem contains a transmitter / receiver. The radio modem looks like a desktop one and connects to a computer via a standard interface RS- 232C, only has an antenna output. It connects either a small whip antenna or an antenna cable, amplifier and directional antenna.
In modern radio modems, noise-like signals have begun to be used, which are sufficiently resistant to common interference and pose almost insurmountable obstacles to data interception. However, the high frequency used (about 900 MHz and higher) requires a line of sight, although it is possible to bypass this limitation by building a relay along a broken line.
Data transmission in wired subscriber access systems can be significantly improved using modem technology ("last mile" equipment), by solving the problem of increasing the speed of information transfer in the "subscriber-station" section without replacing telephone cables with fiber-optic ones.
Packet radio modems are designed to transmit data over a radio channel between mobile users. In this case, several radio modems use the same radio channel in multiple access mode. The radio channel, but its characteristics are close to the telephone one, and is organized using standard radio stations tuned to the same frequency in the meter or decimeter range.
Radio Local Area Networks are an emerging network technology that complements conventional local area networks. Their key element is specialized radio modems of local radio networks. Unlike packet radio modems, such modems provide data transmission over short distances (up to 300 m) at a high speed (2-10 Mbit / s), comparable to the transmission speed in wired local networks. In addition, LAN radio modems operate in a specific frequency range using complex waveforms such as pseudo-random frequency hopping.
Physical line modems differ from other types of modems in that physical line bandwidth is not limited to 3.4 kHz. However, the bandwidth of the physical line is also limited and depends mainly on the type of physical medium (shielded and unshielded twisted pair, coaxial cable, etc.) and its length. From the point of view of the signals used for transmission, modems for physical lines can be divided into low-level modems and baseband modems, which use modulation techniques similar to those used in modems for telephone channels. Modems of the first group usually use digital bi-pulse transmission methods, which allow the formation of pulsed signals without a DC component and often occupy a narrower frequency band than the original digital sequence. Modems of the second group often use various types of quadrature amplitude modulation, which can drastically reduce the bandwidth required for transmission.
Short distance modems are used to communicate between computers, routers and other digital communications equipment, for example, inside buildings, within city boundaries.
Satellite modems are designed to transmit information via satellite communication channels. Currently produced satellite modems operate in various frequency ranges, have the ability to tune and set basic parameters, including operating frequency, gain, output power, modulation type, coding rate, scrambling type, data buffer sizes, etc. These parameters can be changed in small steps over a wide range of values.
Modems for digital data transmission systems resemble low-level modems, but unlike them, they provide connection to standard digital channels such as ISDN, and support the functions of the corresponding channel interfaces.
Cable modems are used to exchange data over specialized cables - for example, through a collective television cable. Instead of telephone lines, cable modems use high bandwidth coaxial cables for video transmission. Up to a hundred television channels is only a small part of the information that can actually be transmitted to an apartment. If the entire cable were used to transmit information, it could be received at speeds in excess of 750 Mbps, which is thousands of times faster than a telephone connection.
DAL modems (DSL - digital subscriber line - digital subscriber line) use dedicated regular telephone lines for communication. A subscriber using a regular telephone connection at the moment has the ability to use the technology DSL significantly increase the connection speed, for example, with a network of modems for physical lines. As a result, he gets access to the Internet while maintaining the normal operation of telephone communications.
Fiber optic modems operate on both single mode, 860 nm, and multimode fiber, 1300 or 1550 nm:
- 860 nm are the most popular, but they have a significant limitation on the cable length - up to 5 km over multimode cable. Radiation source - LED;
- 1300 nm - more versatile - up to 20 km via single-mode fiber with LED, up to 50 km using a semiconductor laser;
- 1550 nm - but single-mode fiber using a semiconductor laser up to 100 km. The maximum distance also depends on the cable diameter.
By transmission method modems are divided into asynchronous and synchronous. This usually implies transmission over a communication channel between modems. Synchronization is typically accomplished in one of two ways, related to how the sender and receiver clocks operate: independently (asynchronously) or concurrently (synchronously).
Asynchronous transfer mode is used when the transmitted data is generated at random times. In this transmission, the receiver must re-sync at the beginning of each received symbol. To do this, each transmitted character is framed with an additional start and one or more stop bits. This mode is often used when transferring data to the interface DTE - DCE. When transmitting digital data over a communication link, the possibilities of using the asynchronous mode are largely limited by low efficiency and the need to use simple modulation methods, such as amplitude and frequency.
Error control. The possibility of errors is not excluded, therefore, a special bit is used in asynchronous transmission - parity bit. The applied error checking and correction scheme is called parity check. Synchronous mode is based on synchronization agreed between two devices. Its purpose is to separate bits from a group when transmitted in blocks. To establish synchronization and check the correctness of work, special characters are used. Since the information bits are in synchronous mode, the start and stop bits are unnecessary. Data transmission ends at the end of one frame and starts at the beginning of another.
Protocols, used in modems are divided into four main groups: modulation and data transmission; error correction; compression of transmitted data; connections DTE P DCE.
The first group of protocols establishes the rules for modems to enter into communication, its support and disconnection, parameters of analog signals, modulation and coding rules. Protocols directly relate to signals transmitted over an inter-modem analog communication line. Connection of two modems is possible only if they support any common or compatible protocols of this group. In a seven-level hierarchy of communication protocols OSI this group of protocols has layer 1 (physical) and forms a digital communication channel in real time, but is not protected from transmission errors.
The second group establishes the rules for detecting and correcting errors that occur during the transmission stage using the protocols of the first group. These protocols only deal with digital information; to check the integrity of the information, it is divided into packets with check redundancy codes. If the control code does not match at the receiving end, the transmitted packet is considered erroneous and its retransmission is requested. This group of protocols forms a reliable (error-free) channel of a higher layer from an unreliable physical channel, but this leads to a loss of real-time communication and comes at the cost of certain overheads. In the model OSI this group corresponds to layer 2 (channel).
The third group sets the rules for compressing transmitted data. At the same time, at the transmitting end, they are analyzed and packed, and at the receiving end, they are unpacked into their original form. Compression allows you to increase the transmission speed beyond the physical bandwidth of the channel. Compression implementation requires some overhead for data analysis and packet formation; in the case of ineffective compression, the transmission rate may be lower than the speed of the physical channel.
The fourth group of protocols sets the rules for interaction DTE and DCE.
Intellectual capabilities of modems. Now modems are intelligent devices that allow, in addition to their main task - the conversion of transmitted signals - to implement many other functions, providing additional convenience to the user. These modems are called intellectual or smart modems. The intellectual capabilities of the modems are realized due to the presence of a control circuit based on one or another microprocessor.
The use of multi-position KAM in its pure form is associated with the problem of insufficient noise immunity. Therefore, in all modern high-speed protocols, QAM is used in conjunction with trellis coding - a special type of convolutional coding. The result is a new modulation method called trellis modulation(TCM - Trellis Coded Modulation). A combination of a specific QAM of an error-correcting code selected in a certain way in the domestic technical literature is called signal-code structure (CCS). SCM allow increasing the noise immunity of information transmission along with reducing the requirements for the signal-to-noise ratio in the channel by 3-6 dB. In this case, the number of signal points is doubled by adding one redundant one formed by convolutional coding to the information bits. The block of bits expanded in this way is subjected to the same QAM. In the process of demodulation, the received signal is decoded using the Viterbi algorithm. It is this algorithm, due to the use of the introduced redundancy and knowledge of the prehistory of the reception process, which allows choosing the most reliable reference point from the signal space by the criterion of maximum likelihood.
The choice of modulation and coding methods is reduced to the search for such filling of the signal space, which provides high speed and high noise immunity. The combination of various ensembles of multi-position signals and error-correcting codes gives rise to many variants of signal structures. Options coordinated in a certain way, providing an improvement in energy and frequency efficiency, are signal-code constructs. The problem of finding the best CCM is one of the most difficult problems in communication theory. Modern high-speed modulation protocols (V.32, V.32bis, V.34, etc.) imply the mandatory use of signal-code structures.
All CCMs in use today use convolutional coding with a rate (ha-1 /P), those. when transmitting one signal element, only one redundant binary symbol is used.
A typical encoder used in conjunction with the FM-8 modulator is shown in Fig. 6.7. It is a convolutional encoder with a relative code rate of 2/3. To each two information bits at the input, the encoder compares three-symbol binary blocks at its output, which are fed to the PM-8 modulator.
Rice. 6.7.
The use of PM signals is associated with resolving the problem of phase ambiguity of the carrier recovered at the reception. This problem is solved due to relative (differential) coding, which in systems without error-correcting coding leads to multiplication of errors. In systems with error-correcting coding, relative coding is also used. In this case, the sequence of switching on of the relative and error-correcting encoder matters.
Distinguish between external and internal relative encoding. With internal coding, the relative encoder is located at the output of the error-correcting encoder, and on the receiving side, the relative decoder is turned on at the input of the error-correcting decoder (Fig. 6.8, a). In this case, the error-correcting encoder must be able to deal with grouping errors.
External relative coding in some cases is more advantageous, since the source of error propagation - the relative decoder - is included at the output of the noise-immune decoder (Fig. 6.8, b). However, this now presents decoding difficulties due to ambiguity in the phase of the reference waveform during demodulation. With FM-2, the ambiguity of the phase of the reference oscillation (0 or z) leads to the phenomenon of "reverse operation", which consists in the fact that the transmitted single bits are received as zero, and zero bits, on the contrary, are received as ones. With a larger number of phase positions, not only inversion, but also permutation of binary symbols is possible. The solution to this problem lies in the use of error-correcting codes, transparent, i.e. insensitive to the phase uncertainty of the reference oscillation. There are several types of SCCs that provide transparency to the phase uncertainty of the reconstructed carrier. They are also based on convolutional encoding at a rate (n - \ / n), those. only one redundant binary character is used.
Close-packed multi-position signals (eg, FM, AFM) provide high specific velocity y by reducing energy efficiency. On the other hand, correction codes can improve energy efficiency with a certain reduction in specific velocity. Each of these methods provides a gain in one indicator in exchange for a deterioration in the other. At the same time, in many cases, it is important to simultaneously increase both energy and frequency efficiency. The solution to this problem is possible when using ensembles of multi-position signals together with error-correcting coding. In this case, it is obviously necessary to form such signal sequences, the points of which in the multidimensional space are densely packed (to ensure high frequency efficiency) and sufficiently spaced (to ensure a sufficiently high energy efficiency). Such signal sequences, built on the basis of error-correcting codes and multi-position signals, are called signal-code structures (see Chapter 7). Convolutional and concatenated codes are commonly used in CCMs as error-correcting codes, and PM, AFM, and FMNF signals are used as multi-position signals.
The device that implements the CCM consists of a codec, a modem and matching devices. To match the codec of a binary error-correcting code and a modem-positional signals, the Gray manipulation code is often used, in which a greater Hamming distance between blocks of code symbols corresponds to a greater Euclidean distance between their corresponding signals. The Gray code, inserted between the error-correcting codec and the modem, converts the-positional channel without memory into a binary channel with memory for the length of symbols. However, Gray's code is not optimal. The binary representation of channel symbols generally requires unequal protection with a correction code. This is due to the fact that the ensembles of multi-position signals used in the channels in most cases turn out to be non-equidistant at the receiving point. The corresponding sets of binary symbols of the manipulation code are also nonequidistant. Other methods of matching message sources and channels are currently known. In particular, methods based on hierarchical
dividing the ensemble of signals into a set of nested sub-ensembles with monotonically increasing distances between them and selecting codes for each level of the hierarchy so as to equalize the resulting distances. More fruitful in this direction is the method of constructing the CCM on the basis of generalized concatenated coding. In this case, external error-correcting codes are consistent with internal codes, which are nested signal sub-assemblies. An example of building a signal-code structure (combined modulation) using lattice Angerboek codes is given in § 7.3.
It is also possible to construct an SCC on the basis of multidimensional signals, which make it possible to increase the number of signal positions without significantly reducing the distance between them. However, it should be remembered that the construction of more advanced CCMs is associated with the inevitable complication of their implementation.
The performance indicators of the CCM are determined by the following ratios:
where um - indicators of the efficiency of the modulation system (modem); DRC - energy gain of coding (codec); frequency efficiency of the codec.
The results of calculations show (Fig. 11.6) that the use of the CCM allows you to simultaneously obtain a gain in both energy and frequency efficiency and, in any case, a gain in one indicator without worsening the other. So, the FM-8-SK system, when using a perforated convolutional code with a rate, provides an energy gain without reducing the specific rate y, and the AFM-16-SK system, with a code constraint, a gain in specific rate without reducing the energy efficiency Information efficiency of these systems
Signal-code constructions based on FMNF signals and convolutional codes are of considerable interest. The phase changes of the FMNF signals have the form of a regular lattice, similar to the lattice diagram of the SC. This makes it possible to combine demodulation and decoding procedures in the ChMNF-SK system by processing signals at reception using a single signal-code lattice using the Viterbi algorithm (AB) or the Klovsky-Nikolaev algorithm (AKN).
