Lecture 4 Multiple access technologies Plan

Lecture 4


Multiple access technologies
Plan:
1. Multiple access with frequency, time and code
2. channel separation.
3. The real subscriber capacity of cellular systems of mobile
4. radio communication.
5. Organization of duplex mode in mobile systems.
The concept of multiple access means a set of measures to ensure the possibility of parallel operation of many users within the time-frequency resource allocated to a given system. By identifying each subscriber with a certain physical channel, we can say that a specific multiple access technology is a way of distributing a limited time-frequency resource between channels.
Let s (t) be the signal through which the i-th communication channel is realized, (i = 1,2, ..., k, where k is the total number of channels of the system. Adhering to the linear model of the physical propagation medium, ignoring the effects insignificant in this context multipath and considering the accompanying noise n (t) additive, the oscillation observed by the receiving side can be represented in the form, respectively, attenuation and delay
t-th signal on the propagation path. The task of the receiving side is to select a message from a specific subscriber, with third-party signals acting as interfering signals.
As is known, for linear selection of any component of the superposition of signals with the elimination of the influence of other components, linear independence of all signals is necessary and sufficient. Orthogonal signals, which are linearly independent, are separated without mutual interference by a conventional correlation receiver.
Orthogonality of channel signals can be ensured by their frequency or time spacing, or by suitable coding. In accordance with this, a classification of multiple access methods has been established.
All mobile radio communication systems fall into two general categories: analogue and digital. At the same time, they usually use three main methods of multiple access: FDMA, TDMA, and CDMA.
Frequency division multiple access (FDMA). This multiple access technology was originally used in analog cellular communication systems due to its availability and ease of implementation. FDMA systems are constructed in such a way that each subscriber is allocated a frequency channel with a band Δfp for the duration of a communication session, within the overall frequency range of the Δfp system, which does not coincide with any of the channels already provided to active subscribers. In this case, the spectra of the channel signals do not overlap (Figure 5.1a), which ensures the fulfillment of the orthogonality condition.

Figure 5.1 Spectra of channel signals (a) and distribution of the time-frequency resource between subscribers (b)

The rectangle with sides Δfр and Tr in Fig.5.1b characterizes the total time-frequency resource of the system. As you can see, with FDMA this common resource is "cut" into k (according to the number of subscribers) horizontal strips, each occupying the entire available time resource and only a part of the frequency resource. Thus, subscriber channels in the system are spaced apart in frequency, like radio stations on the air.

The FDMA method is used in all analog SSMS, i.e. systems of the first generation, and the bandwidth of the subscriber channel for them is Δfп = 10 ... 30 kHz. For example, in the AMPS standard, the frequency resource allocated to the system consists of two sections: in the range 869 ... 894 MHz for transmitting information from BTS to MS (forward channel)) and 824 ... 849 MHz for transmitting information in the opposite direction (reverse channel). Thus, the duplex frequency spacing is 45 MHz. Each frequency channel is assigned a band Δfp, = 30 kHz, so that taking into account the guard intervals, 832 channels are placed within the allocated range and numbers from 1 to 799 and from 991 to 1023 are assigned.
Forward (downlink) and reverse channels (uplink), having a spectral separation, completely eliminate the influence of channels on each other, can only theoretically. In practice, it is impossible to avoid the occurrence of inter-channel (intra-system) interference, for example, due to the imperfection of the crossover filters in the receiver, as a result of which a part of the signal energy of one channel leaks into the adjacent one. The influence of interchannel interference can be mitigated by appropriate choice of signal manipulation (reduction of "out-of-band" emissions) and filters (improvement of suppression in the adjacent channel). Another way to reduce the level of mutual interference is the introduction of guard intervals between frequency channels, which, however, leads to a decrease in the frequency band used for communication, i.e. decrease in the efficiency of spectrum use.
It should be noted that in duplex communication, strictly speaking, two Δfp bands are used, one for the forward and the other for the reverse channels.
The main advantage of FDMA is the full use of the allocated bandwidth by the subscriber. The disadvantage is obvious - the low efficiency of using the frequency and energy resources per subscriber. In addition, it should be noted that, as a rule, analog standards are accompanied by low immunity from interference, the lack of effective methods to combat signal fading under the influence of the surrounding landscape, buildings or due to the movement of subscribers, as well as from listening to subscriber channels on the air (more often its complete absence). This implies: a relatively high required signal-to-noise ratio - up to 15 dB (against 9 in GSM and even 3 ~ 5 in CDMA), which causes the use of high transmitter power, and the ability to receive any radiotelephone channel of the system on a modified VHF-FM receiver. The range of value-added services in analogue networks is also quite small in comparison with digital ones. Some of the many analog standards that use FDMA are NMT and AMPS.
In fig. 5.2 shows an example of allocation of the allocated frequency range between 4 subscribers. Each of the subscribers is provided with a bandwidth of 30 kHz.



Figure 5.2 Allocation of the allocated frequency range
Time division multiple access TDMA (TDMA) in the traditional sense is that each subscriber of the system for the duration of the communication session is allocated a time slot T0 (or time channel Tk) within the total time resource of the system Tp (cycle or frame of the system Ti) that does not coincide with any of the intervals already provided to other active subscribers. Thus, each channel signal is located in its own individual window (word) without overlapping with others (Figure 5.3a).


Fig.5.3 Location of channel signals in time (a), distribution of the time-frequency resource between subscribers (c) and the combination of frequency and time division of the FDMA / TDMA (d).

At the same time, the spectra of subscriber signals can occupy the entire frequency band Δfр allocated to the system and completely overlap. An illustration of such a resource distribution is Fig.5.3c, from which it can be seen that the total frequency-time resource is "cut" in the form of k vertical bands, each of which occupies the entire available frequency range and only the K-th (Tc) part of the allotted time.

Each subscriber connects to the path periodically with a period Ti, and sends its channel signal (CS) to the group path. The duration of the COP should be Tk After separation at the reception, for each CC, the original information is restored, i.e. interpolation is performed.
Therefore, in time division multiplexing (TDM) systems, transmission is performed in frames equal to Ti. Moreover, for synchronous switching in a cycle, a cyclic signal is transmitted. In addition, in the cycle, time is allocated for the transmission of the CW of the service communication, then the time allotted for one channel Tk is equal to

Usually Т_к=Т_цс=Т_сс, hence (Figure 5.4).

Figure 5.4 Transmission Cycle with VRK

As a rule, the COP is transmitted by a binary message and τi = τprotection, then Fig.5.5.

Figure 5.5 Transmission of channel pulses to the VRK
The shape of the COP can be different, most often bell-shaped pulses are used, because they are easier to shape. Thus, one of the main differences from FDMA arises - the need for synchronization.
Ideally, the mismatch of the channel signals in time ensures their orthogonality, and therefore excludes the influence on each other. In reality, due to the limited system bandwidth, the transient processes from the signals of the previous channels to the beginning of the appearance of the subsequent ones may not end and, adding up with the latter, create cross-channel (inter-channel) interference. Reduce the influence of adjacent channels, i.e. the level of inter-channel interference is achieved by introducing guard time intervals, which, in turn, leads to a decrease in the range during which information transmission is possible, i.e. to the actual reduction in the transmission speed.
In order to improve efficiency, narrowband TDMA (NB TDMA or MC / TDMA) is used in cellular communication systems - combined FDMA and TDMA. In this case, the allocated frequency range, as in FDMA, is divided by carriers into separate bands. And already within each frequency band, time division of subscribers is carried out. As a rule, this method is used in digital systems (GSM, D-AMPS). Figure 5.6 shows an example of dividing the allocated band into 4 subcarriers, providing communication to a maximum of 4 subscribers simultaneously within each band. Thus, using the narrowband TDMA method, 4 times more subscribers can work in the same frequency range than in FDMA. Therefore, the capacity of networks using TDMA is, on average, 3-6 times greater than that of FDMA.



Figure 5.6 Dividing the allocated frequency range
The disadvantages of TDMA include sensitivity to synchronization failure. In analog standards that do not have synchronization, the cell radius in practice is 35-40 km, while it can be increased to 60 or even 80 km (in theory) in accordance with the principle of radio communication - "shout louder, hear farther", i.e. ... increasing the output power of the transmitter. In digital standards, at least implemented on TDMA technology, absolute delay time compensation systems are capable of operating in an interval of up to 250 μs, which corresponds to a maximum cell radius of 35 km.
Multiple access with code division multiple channels CDMA (CDMA - Code Division Multiple Access). It is based on an orientation towards the spread spectrum ideology of building information transmission systems, which provides for a conscious and multiple expansion of the bandwidth of the transmitted message compared to that which is characteristic of traditional narrow-band systems
The theory of this method was developed back in 1935, but it received practical implementation much later than FDMA and TDMA. When it is implemented, all subscribers constantly use the entire available bandwidth allocated for communication to the system. The basis of this technology, as a rule, is the orthogonal division of traffic channels by means of Walsh functions. In total, 64 of them have been identified. Thus, theoretically, one base station allows organizing the work of 64 subscribers. However, due to the influence of interference and the need to provide "soft handover" - "soft" switching between cells (in general, which is one of the advantages of the CDMA method), in practice they use up to 45 for fixed, and up to 25 for mobile communications. Communication systems using CDMA, as well as TDMA, need synchronization. However, here the Global Positioning System (GPS) acts as a synchronizing link. The advantage of communication systems based on the CDMA principle is the absence of frequency planning necessary for efficient operation using other methods. In addition, the low signal-to-noise ratio (ideally 3-5 dB) and the very principle of CDMA implementation (using noise-like signals) allows the use of transmitters with a radiation power 2-3 orders of magnitude lower than other standards, while being only 10 mW and below. Unlike TDMA, the coverage area of ​​CDMA-based systems depends mainly on the power of the subscriber station. The network capacity based on this technology is on average 5 times more than TDMA, and more than 10 times more than FDMA. At the same time, its construction requires approximately 40% fewer base stations than for a network with the same coverage, but based on TDMA technology.
The implementation of the CDMA method consists in a significant artificial expansion of the base of the primary signal (subscriber's voice), which is implemented in one of two main ways:
• direct extension - direct sequence spread spectrum (DSSS);
• frequency hop spread
spectrum (FHSS).
In the first version, the information message manipulates a pseudo-random sequence (PSP), consisting of elements (chips) of duration T, and the duration of the chip is many times (N times) less than the duration T of the transmitted information bit or symbol.
The value of N directly characterizes the degree of bandwidth expansion in comparison with the bandwidth of the primary message and is therefore called the spreading factor (in English texts, the spreading factor or processing gain).
The mentioned manipulation of the PSP c (t) with the transmitted data stream D (t) is usually realized by their simple multiplication (Fig. 5.7a).
The diagrams in Figs 5.7b - d illustrate the content of the direct expansion procedure for an example of a binary transmission and binary PSP. Figure 5.7, c shows a periodic binary PSP, whose period, containing N = 8 chips, coincides with the duration of one message message (in the general case, the PSP period can be arbitrary, in particular, much longer than the duration of the information message; moreover, the PSP in general may be aperiodic). Direct expansion is clear
(Figure 5.7d), if the information message carries a zero bit (positive polarity D (t), Figure 4.7b), the original version of the PSP is present at the output of the multiplier. When the value of 1 of the current bit is transmitted by the message, the polarity of the PSP is reversed. The signal after the multiplier is fed to a standard carrier modulator (BPSK, QPSK, etc.).



Figure 5.7 Binary forward spreading procedure
transmission and binary PSP

As can be seen, the forward spectrum spreading procedure does not degrade the noise immunity of the binary transmission in the Gaussian channel, leaving opposite signals corresponding to the values ​​of 0 and 1 of the transmitted bit.

When using the second spreading method, each data message symbol must be transmitted using a set of discrete frequencies specified in a specific sequence.
In existing and future development of cellular communication systems, direct spectrum spreading is mainly used, implemented either in a synchronous or asynchronous version. The differences between these two DSSS modifications are quite significant. The first can be used when it is possible to synchronize with each other all individual address sequences (signatures) assigned to individual subscribers so that signals of different subscribers on the receiving side do not have mutual time shifts. A similar situation is typical for the SSMS downlink (from BTS to MS), since the BTS signals sent by different MSs strictly simultaneously arrive at a separate MS along the same path, i.e. without mutual delays.
In the uplink, ensuring synchronization of signals of different MSs received by the BTS, although theoretically not denied, is rather difficult and not always technologically justified due to the random location of the MS relative to the BTS within the cell and, hence, random mutual signal delays. Such situations are characterized by the use of the asynchronous version of DSSS, which does not imply the mutual timing of signatures of individual subscribers.
During transmission, as shown above, the information signal base expands; it is, as it were, "smeared" over the entire working range. Unlike narrowband TDMA, code division takes into account not only frequency and time division, but also instantaneous signal power division (Figure 5.8). So, instantaneous frequency, time and power are the building blocks of the code.
At the reception at the subscriber, using the code, the signal is “convolved” in the entire range, while the signals of the other correspondents remain “unrolled” and for the receiver they represent an ideal type of interference - white Gaussian noise (GBN).

Figure 5.8 Separation by instantaneous signal strength
This technology can be illustrated by using a room in which there are several pairs of people speaking different languages. They only communicate with each other and are not interested in others. If each couple knows only one language and uses it, and all languages ​​are different, then the air of the room can be "carrying" for their voices. The analogy is that the air in a room is a broadband channel and languages ​​are distributed as codes. Thus, those, for example, who speak Spanish, will not hear and understand anything other than Spanish.
You can increase the number of couples as long as the general background does not limit the ability to communicate. By adjusting the signal strength of all subscribers, which should not exceed a certain value, it is possible to provide communication for a large number of people while maintaining high speech quality.
The maximum number of users, or traffic channels, depends on the usage of each channel, and therefore is not specific. This is reflected in the concept of "soft overload", whereby an additional subscriber can gain access at the expense of some increase in interference to other subscribers. As a result, when implementing the CDMA standard, when designing a network, it is necessary to minimize the overall level of interference, since together with the interference created by other subscribers and base stations, they are a factor determining the upper threshold for throughput.
Cellular networks built on the CDMA principle have additional services that are traditional for digital standards implemented on TDMA.
The advantages of the CDMA in relation to the CDMA and TDMA can be roughly divided into two groups. The first of them consists of those that distinguish any spread spectrum systems - high noise immunity to concentrated and broadband interference (including deliberate), the ability to work effectively in multipath propagation conditions, a wide range of available crypto protection measures, high accuracy of measurement of frequency-time parameters, good electromagnetic compatibility with radio communication and broadcasting systems, etc. The second group is directly related to the aspects of multiple access - a large subscriber capacity per cell (sector), a "soft" nature of the decline in communication quality with increasing traffic intensity, the ease of implementation of the "soft" relay transmission.
Real subscriber capacity of cellular mobile radio communication systems. When describing various mobile radio communication systems, their subscriber capacity, i.e. the number of subscribers served is implicitly determined by the number of radio communication channels. It is quite obvious that these two concepts are not identical. It seems unreasonable to limit the number of subscribers served by the number of communication channels, since the probability of simultaneous communication of all subscribers of the system is usually small. Therefore, in the presence of communication channels, the system, in principle, is able to serve more to subscribers, although there is a possibility that in some cases the subscribers will not get access, and this probability is the greater, the greater the number of subscribers compared to the number of communication channels. This implies a two-pronged problem, the conditions of which can be formulated as follows: how many subscribers can be served by a system that has a fixed number of communication channels for a given probability of denial of access, or - how many communication channels in the system are necessary to service a given number of subscribers with a fixed probability of failure in access. The solution to this problem is based on the provisions of the queuing theory. Indeed, mobile radio communication is inherently an example of a queuing system, i.e. systems with a random flow of requests (calls), a random service duration (communication session) and a finite number of service (communication) channels.
The most common characteristics of random call flows are the average frequency of their arrival y, measured by the number of calls per unit of time (for example, the number of calls / hour), and the average duration of the communication session, T, expressed in units of time. The product A = y T determines the average traffic (traffic intensity, load flow, total load on the communication channel), measured in Erlangs (in honor of the Danish scientist A.K. Erlang, famous for his works in the field of teletraffic theory). It should be noted that the characteristics of the load on the communication channel y and T are usually estimated at the interval of the highest system load, i.e. at rush hour.
The number of calls during a fixed time t is a discrete random variable, usually described by the Poisson distribution: the probability of k calls arriving in time t. The duration of the communication session (the duration of one channel busy) t is a continuous random variable, the probability density of which is usually taken exponential (T is the average value)
Depending on how the system behaves, which does not have free channels at the time a new request arrives, the following models are distinguished:
• a system with a limited waiting time (Erlang A model), according to which a call is queued in the absence of a free channel and after a fixed time T, if during this time none of the busy channels is free, is canceled;
• a system with failures, i.e. calls that came when there are no free channels are canceled (Erlang B model);
• a system with an expectation, i.e. calls are queued and can wait indefinitely for the channel to be released (Erlang C model).
There are formulas that determine the probability of denial of access to the system. However, they are cumbersome and practically not used. In practice, to calculate the permissible Erlang load for a communication system with channels at a given blocking probability, its presentation is used either in the form of a graph or in the form of a table, which are published in the relevant technical literature.
Organization of duplex mode in mobile systems. The total time-frequency resource allocated to a specific system has to be spent not only on organizing multiple access, but also on ensuring duplex mode, i.e. parallel information exchange in both directions: from the system to the subscriber and in the opposite direction. Frequency and time duplex are used in mobile radio communication systems. In the first variant, referred to as FDD (frequency division duplex), the duplex pair occupies two frequency bands of the subscriber channel bandwidth, separated by a certain guard interval, called the duplex frequency spacing. Thus, the transmission and reception of information between subscribers is carried out at different frequencies. The FDD principle is illustrated in Figure 5.9.

Figure 5.9 Principle of organization of duplex frequency separation

Systems of standards of the first and second generations (AMPS, DAMPS, GSM, IS-95, etc.) are built on the basis of FDD.
With time division duplex (TDD), two-way communication uses the same carrier with time division of transmission and reception channels (Figure 5.10). TDD mode is not typical for existing cellular communication systems, but it is widespread in cordless telephone standards (CT2, DECT, etc.). In addition, it has a certain place in the third generation UMTS and cdma2000 standards.



Figure 5.10 Principle of duplex timing

Let's consider a typical structure of channels of a TDD system, focusing for specificity on the numbers underlying the cdma2000 project. The main element of the channel architecture of the BTS is Tk = 20 ms (Figure 5.11), which is divided into 8 pairs of intervals intended for organizing duplex.

Figure 5.11 Frame structure of CDMA2000 TDD communication channel
The first interval of the pair has a duration of TT and is reserved for transmission. In the second (duration TR), the MS signal is received. Any adjacent intervals are separated by guard intervals of duration Δt, determined by the length of the service area. A simple calculation shows that with a guard interval of 52 µs and a time slot synchronization accuracy at the base station of +3 µs, the maximum radius of the service area is 14 km.
Mobile stations have a frame structure similar to BTS, but the transmission and reception intervals are reversed.
Comparison of the two duplexing options leads to the conclusion that the FDD mode is more efficient with large cell sizes and high subscriber movement speeds, while the TDD option is more suitable for use in microcells, i.e. in small service areas of subscribers moving at low speed. At the same time, the TDD mode has a number of additional advantages that deserve special mention. Since the uplink and downlinks in TDD occupy the same frequency band, the fading characteristics are highly correlated, which can be used to simplify the radiated power and space diversity control procedures.
In addition, the flexible frame structure inherent in TDD allows for efficient reallocation of temporary resources with asymmetric traffic flows in the forward and reverse channels. Such asymmetry will turn out to be a very frequent phenomenon in third generation systems, due to the imposition of communication functions with the Internet on mobile terminals. During this contact, downlink traffic is generally much richer than the other way. In this case, you can act as shown in Figure 5.12, which schematically represents the transition from a symmetric (Figure 5.12a) distribution of the temporary resource between the “down” and “up” lines to an asymmetric (Figure 5.12b), on which the “down” arrow is responsible for receiving information MS, and "up" - for transmission.



а

б
Figure 5.12 Symmetrical (a) and asymmetric (b) location of the temporary
resource between the "down" and "up" lines in the communication channel

As an advantage of the TDD option, the possibility of a simpler implementation of a single-mode TDD subscriber terminal can also be considered, due to the absence of a duplexer in it. As for the hardware complication of a dual-mode (FDD / TDD) terminal designed for both duplexing options, compared to a conventional FDD terminal, it is not too significant and does not have a critical impact on economic performance.

From the above it follows that the recommendations of the European UMTS project, on the combination of both considered duplexing modes, seem to be quite rational. Such a solution gives the system flexibility in terms of using the allocated spectral range and allows you to adapt the bandwidth to the operating conditions and the nature of the services. According to the European concept of building a third-generation system, in two allocated WARC-92 spectral sections with a width of 230 MHz: 1885 ... 2025 and 2110 ... 2200 MHz, bands 1920 ... 1980 and 2110 ... 2170 are intended for SSMS with frequency duplex, and bands 1900 ... 1920 and 2010 ... 2025 MHz - for SSMS with time duplex.
Control questions
1.Explain frequency division multiple access.
2.Explain time division multiple access.
3.Explain code division multiple access.
4. How the number of users and the real subscriber capacity of cellular systems are estimated
mobile radio communication?
5.How is the duplex mode organized in mobile communication systems?