Ma Aste Er T Thes
MASTER THESIS
TITLE: Planning optimization software tool for DVB-T and DVB-T2
MASTER DEGREE: Master in Science in Telecommunication Engineering & Management
AUTHORS: María Lema Rosas, Evelyn Torras López
DIRECTORS: Silvia Ruiz Boqué, Mario García Lozano
DATE: June 25 th 2010
TIitle: Planning optimization software tool for DVB-T and DVB-T2
Authors: María Lema Rosas, Evelyn Torras López
Directors: Silvia Ruiz Boqué, Mario García Lozano
Date: June 25 th 2010
Overview
Nowadays the implementation of the Digital Terrestrial Television network is an actual fact in the Spanish territory. Its development is crucial for the digital transition in those countries which mainly depend on terrestrial networks for the reception of multimedia contents. With the aim of giving to the users a high quality signal but with the minimum cost for the operator, it arises the necessity of an optimization of the transmission network. In this way, there are variables that can be modified to obtain this goal, some of them are uncontrollable by the operators and the others are susceptible of optimization. In this last group, it can be found the static internal delays of the transmitters which are of special interest because changes can be done with cost zero. The main objective of this project is the design and implementation of a planning optimization software tool that adjusts the internal static delay in transmitters of a digital broadcasting network. The final objective is to minimize the self-interfered areas and obtain the correspondent increase in coverage. The planning optimization software tool makes use of a metaheuristic algorithm and can obtain different layers of study with its corresponding results. This report describes the problems raised by the current network, the algorithm adapted to the needs of the network to optimize, the implementation of the simulation platform as well as its subsequent validation stage, including several graphical results that allow the evaluation of the improvements introduced over the realistic scenario tested. Finally it has to be mentioned that this work has been performed within the context of the FURIA project, which is a strategic research project funded by the Spanish Ministry of Industry, Tourism and Commerce.
Título: Planning optimization software tool for DVB-T and DVB-T2
Autores: María Lema Rosas, Evelyn Torras López
Directores: Silvia Ruiz Boqué, Mario García Lozano
Fecha: 25 de junio de 2010
Resumen
Hoy en día la implementación de una red de Televisión Digital Terrestre es una realidad en España y su desarrollo es crucial para la transición digital para aquellos países en los que la recepción de contenido multimedia depende exclusivamente de las redes terrestres.
Con el objetivo de dar al usuario una gran calidad de señal pero con un coste económico y reducido para las operadoras es necesaria la optimización de la res de transmisión. En este sentido, existen parámetros de red configurables para conseguir este objetivo, algunos de ellos son incontrolables para las operadoras, en cambio otros son susceptibles a la optimización. En este último grupo, se encuentran los retardos estáticos internos de los transmisores que son de especial interés debido a que se pueden modificar a coste cero.
EL principal objetivo del proyecto es el diseño y la implementación de una herramienta software para la planificación óptima de redes que ajusta el parámetro de los retardos estáticos de los transmisores en redes de difusión digital de contenidos con la intención de minimizar las zonas de auto-interferencia y con el correspondiente incremento de cobertura de red. La herramienta software para la planificación hace uso de un algoritmo metaurístico y puede obtener diferentes capas de estudio con sus correspondientes resultados.
La memoria describe los problemas planteados por la red actual, el algoritmo utilizado y como se ha adaptado a las necesidades de la red a optimizar, la implementación de la plataforma de simulación así como su etapa de validación, incluyendo gran cantidad de resultados gráficos que permiten la evaluación de las mejoras introducidas en la red realística testeada.
Finalmente se ha de mencionar que el presente trabajo ha sido realizado en el contexto del proyecto FURIA, el cual es un proyecto de búsqueda estratégica fundado por el ministerio de Industria, Turismo y comercio de España.
INDEX
INTRODUCTION ...... 1 CHAPTER 1. INTRODUCTION TO DVB-T ...... 3 1.1 Changing to a digital mode ...... 3 1.2 Digital television around the world ...... 3 1.3 DVB-T2 physical layer overview ...... 4 1.3.1 Basic principle ...... 5 1.3.2 Time domain physical layer considerations ...... 6 1.4 Radio planning of DVB-T systems ...... 7 1.4.1 Coverage computation ...... 8 1.4.2 Interferences and their impact on coverage ...... 9 1.4.3 Solutions to improve coverage area ...... 10 CHAPTER 2. DESIGN OF AN ALGORITHM TO MAXIMIZE COVERAGE ...... 11 2.1 Problem description ...... 11 2.2 Metaheuristic solution ...... 12 2.3 Introduction to Simulated Annealing ...... 13 2.3.1 Description of the algorithm ...... 13 2.4 Simulated annealing parameters ...... 15 2.4.1 Temperature ...... 15 2.4.2 Number of iterations ...... 16 2.4.3 Conditions of convergence ...... 16 2.4.4 Definition of the cost function ...... 16 2.5 Implementation issues ...... 18 2.5.1 Study of the temperature reduction coefficient ...... 18 2.5.2 Maximum delay value delimitation ...... 20 2.5.3 Study of the equilibrium condition ...... 22 CHAPTER 3. SIMULATION PLATFORM DEVELOPMENT ...... 25 3.1 Required software tools ...... 25 3.1.1 Sirenet ...... 25 3.1.2 Microsoft Visual Studio 2008 ...... 25 3.1.3 MATLAB ...... 26 3.1.4 Google Earth ...... 26 3.2 Simulator structure ...... 27 3.2.1 Main program ...... 28 3.2.2 Scenario ...... 28 3.2.3 Algorithm ...... 30 3.2.4 Graphical User Interface ...... 30 3.3 Code optimization ...... 34 CHAPTER 4. RESULTS: OPTIMIZATION OF REALISTICS SCENARIOS ...... 35
4.1 Definition of the scenarios ...... 35 4.1.1 Tarragona scenario ...... 35 4.1.2 Lleida scenario ...... 36 4.1.3 Barcelona scenario ...... 37 4.2 Results for geographical optimization ...... 38 4.2.1 Tarragona ...... 40 4.2.2 Lleida ...... 41 4.2.3 Barcelona ...... 42 4.3 Optimizing population density ...... 43 4.3.1 Population distribution ...... 44 4.3.2 Discussion over Tarragona ...... 45 4.3.3 Discussion over Lleida ...... 47 4.3.4 Discussion over Barcelona ...... 49 4.3.5 Conclusions for the cost function approach ...... 50 CHAPTER 5. RESULTS: IMPACT OF OTHER VARIABLES OVER THE OPTIMIZATION PROCESS ...... 51 5.1 Receiving antennas ...... 51 5.1.1 Definition of the scenarios ...... 51 5.1.2 Yagi antennas: Fixed environment ...... 53 5.1.3 Omnidirectional antennas: Mobile environment ...... 58 5.2 Types of receivers ...... 64 5.2.1 Definition of the scenarios ...... 64 5.2.2 Yagi antenna: Fixed environment ...... 65 5.2.3 Omnidireccional antenna: Mobile environment ...... 67 CHAPTER 6. CONCLUSIONS ...... 69 CHAPTER 7. REFERENCES ...... 71 APPENDICES ...... I
FIGURE INDEX
Fig. 1.1 Different standards around the world ...... 4 Fig. 1.2 Time and frequency representation of the SC and OFDM. In OFDM, N data symbols are transmitted simultaneously on N orthogonal subcarriers [3] ...... 5 Fig. 1.3 Presentation of the OFDM subcarrier frequency ...... 6 Fig. 1.4 Cyclic Prefix added at the beginning of the OFDM symbol ...... 6 Fig. 2.1 Algorithm work flow ...... 12 Fig. 2.2 Block diagram of the search of the initial temperature ...... 14 Fig. 2.3 Block diagram of SA ...... 15 Fig. 2.4 Cost function computation ...... 17 Fig. 2.5 Cost function vs. Delta ...... 19 Fig. 2.6 Execution time vs. Delta ...... 19 Fig. 2.7 Delay evolution (4 transmitters, 200 iterations) ...... 20 Fig. 2.8 Cost function vs. Max Delay ...... 21 Fig. 2.9 Execution time vs. Max Delay ...... 21 Fig. 2.10 Cost function vs. Beta ...... 22 Fig. 2.11 Execution time vs. Beta ...... 23 Fig. 3.1 Visualization of KML file ...... 27 Fig. 3.2 Software tool class diagram ...... 27 Fig. 3.3 Initial correspondence ...... 29 Fig. 3.4 GUI interface ...... 30 Fig. 3.5 Example of CIR results ...... 31 Fig. 3.6 Example of population results ...... 32 Fig. 3.7 Example of transmitter coverage results ...... 33 Fig. 3.8 Example of CIR by population results ...... 33 Fig. 4.1 Coverage area in Tarragona ...... 36 Fig. 4.2 Coverage area in Lleida ...... 37 Fig. 4.3 Coverage area in Barcelona ...... 38 Fig. 4.4 Type one receiver schema ...... 39 Fig. 4.5 Initial CIR for Tarragona region ...... 40 Fig. 4.6 Final CIR for Tarragona region ...... 40 Fig. 4.7 Initial CIR for Lleida region ...... 41 Fig. 4.8 Final CIR for Lleida region ...... 42 Fig. 4.9 Initial CIR for Barcelona region ...... 43 Fig. 4.10 Final CIR for Barcelona region ...... 43 Fig. 4.11 Population density for Tarragona ...... 44 Fig. 4.12 Population density for Lleida ...... 45 Fig. 4.13 Population density for Barcelona ...... 45 Fig. 4.14 Initial population covered in Tarragona ...... 46 Fig. 4.15 Geographical based optimization in Tarragona ...... 46 Fig. 4.16 Population based optimization in Tarragona ...... 47 Fig. 4.17 Initial population covered in Lleida ...... 47 Fig. 4.18 Geographic based optimization in Lleida ...... 48 Fig. 4.19 Population based optimization in Lleida ...... 48 Fig. 4.20 Initial population covered in Barcelona ...... 49 Fig. 4.21 Geographic based optimization in Barcelona ...... 50 Fig. 4.22 Population based optimization in Barcelona ...... 50 Fig. 5.1 Example of a conventional Yagi horizontal radiation pattern ...... 54 Fig. 5.2 Initial coverage in Tarragona in fixed scenario with receiver type one ...... 54 Fig. 5.3 Initial coverage in Lleida in fixed scenario with receiver type one ...... 54 Fig. 5.4 Initial coverage in Barcelona in fixed scenario with receiver type one ...... 55 Fig. 5.5 Final CIR in Tarragona in fixed scenario with receiver type one ...... 55 Fig. 5.6 Final CIR in Lleida in fixed scenario with receiver type one ...... 56
Fig. 5.7 Final CIR in Barcelona in fixed scenario with receiver type one ...... 56 Fig. 5.8 Initial percentage of pixels covered ...... 57 Fig. 5.9 Final percentage of pixels covered ...... 57 Fig. 5.10 Initial coverage in Tarragona in portable scenario with receiver type one ..... 58 Fig. 5.11 Initial coverage in Lleida in portable scenario with receiver type one ...... 58 Fig. 5.12 Initial coverage in Barcelona in portable scenario with receiver type one ..... 59 Fig. 5.13 Final coverage in Tarragona in portable scenario with receiver type one ..... 59 Fig. 5.14 Final coverage in Lleida in portable scenario with receiver type one ...... 60 Fig. 5.15 Final coverage in Barcelona in portable scenario with receiver type one ...... 60 Fig. 5.16 Percentage of pixels covered with at least the minimum power ...... 61 Fig. 5.17 Initial coverage of omnidireccional TXs ...... 63 Fig. 5.18 Final coverage of omnidireccional TXs ...... 63 Fig. 5.19 Final coverage according to population ...... 63 Fig. 5.20 Type two receiver schema ...... 64 Fig. 5.21 Final coverage in Tarragona in fixed scenario with receiver type two ...... 65 Fig. 5.22 Final coverage in Lleida in fixed scenario with receiver type two ...... 65 Fig. 5.23 Final coverage in Barcelona in fixed scenario with receiver type two ...... 66 Fig. 5.24 Comparison of final percentage of pixels covered in fixed scenario ...... 66 Fig. 5.25 Final coverage in Tarragona in portable scenario with receiver type two ...... 67 Fig. 5.26 Final coverage in Lleida in portable scenario with receiver type two ...... 67 Fig. 5.27 Final coverage in Barcelona in portable scenario with receiver type two ...... 67 Fig. 5.28 Comparison of final percentage of pixels covered in portable scenario ...... 68
TABLE INDEX Table 1.1 Specified length of the guard interval [4] ...... 7 Table 1.2 Link budget results ...... 9 Table 3.1 Example of KML file ...... 27 Table 3.2 Pixel information ...... 28 Table 3.3 Legend of population density distribution ...... 32 Table 3.4 Example of code ...... 34 Table 3.5 Example of loop jamming ...... 34 Table 4.1 Analysed transmitters in Tarragona ...... 36 Table 4.2 Analysed transmitters in Lleida ...... 37 Table 4.3 Analysed transmitters in Barcelona ...... 38 Table 4.4 Radio parameters ...... 39 Table 5.1 Principal radio characteristics of portable scenario ...... 52 Table 5.2 Principal radio characteristics of fixed scenario ...... 52 Table 5.3 Main roads in Lleida region ...... 62 Table 5.4 Characteristics of the transmitters added ...... 62
INTRODUCTION 1
INTRODUCTION
In the context of broadcasting television networks, analogical technologies are being replaced by digital ones. This transition of analogical to digital by the roll out of the Digital Video Broadcasting Terrestrial (DVB-T) standard provides advantages in the exploitation of bandwidths, more robustness in front to the noise and another series of advantages that are translated in a clear improvement of the image and the sound, besides adding new applications for users.
One of the strengths of DVB-T is that it can be deployed with a single frequency network (SFN) scheme. This is possible because the physical layer is based on Orthogonal Frequency Division Multiplexing (OFDM) and the introduction of a cyclic prefix (CP) between consecutive symbols. SFNs allow a more efficient use of available bandwidth than classic Multiple Frequency Networks (MFNs). They also simplify the radio-planning process since frequency allocation strategies are not required.
Due to the multipath tolerance that the OFDM scenario has, one receiver is allowed to combine signals coming from different transmitters (TXs), as long as these signals remain inside the guard interval (GI). This interval copes with intersymbol interference (ISI) induced by the multipath channel. The fact of combining signals provides a diversity gain in reception only in the case that TXs are allocated near of the receiver area, however it can happen that a signal is delayed due to the distance between the TX and the receiver area. In that case, this signal cannot be combined because it falls outside of the GI. This phenomenon is call self-interference.
With the aim of giving users a high quality signal but at the same time of limiting expenditures for the operator, it turns necessary the optimization of the transmission network, bearing in mind that the most important factor for the optimization is the minimization of self- interference. There are variables that can be modified to obtain this objective, some of them are uncontrollable by operators, as for example: The propagation environment and the configuration of OFDM receivers. Given this, different receiver options should be considered and assessed when optimizing the network planning. On the other hand, there exists another type of variables that are susceptible of optimization, such as the geographic position of the TXs, their transmission power, the configuration of their radiant system and their static internal delays. Among these, the last one is of special interest because changes can be done with cost zero.
Given this, this project develops a planning optimization software tool that adjusts the static delays of the TXs in SFNs in order to minimize self-interfered areas and with the correspondent increase in coverage for a given broadcasting video network. To solve the optimization problem the software tool makes use of the metaheuristic Simulated Annealing (SA), previously studied in [1], but further improved in this project by adding a complimentary and complete software tool and more layers of study.
The report is organized as follows. First of all a brief introduction to the DVB-T and DVB-T2 standard is presented, consecutively, after setting the problem, the main objectives of the project are explained and the starting points of the simulator are established. Examples of this are the link budget that determines the simulation thresholds and the territory to optimize. Chapter 2 describes the SA algorithm and its parameters, followed by the characterization of the algorithm itself. Then the 2 Planning optimization software tool for DVB-T and DVB-T2 optimization software tool structure and implementation is explained in chapter 3. Results are divided in two parts, first of all chapter 4 presents the results obtained with realistic scenarios. Chapter 5 focuses in specific topics and assess the impact of other variables that affect on the received signal, as for example types of receiving antennas or the modification of the type of receiver. Finally, the report is closed with some conclusions, and a set of ideas that can give further lines of investigation on this topic.
This project has been developed under a more ambitious national project named FURIA [2]. FURIA is a SSP (Strategically Singular Project) in the field of Network Audiovisual Technologies, whose main objective is to develop and validate the integration of emergent technologies for the spreading of audiovisual contents in fixed and mobile devices. Joining forces from the different national organisations (companies, technological centres and universities) with the final purpose of increasing the national technological level.
Following the same FURIA consortium definition of its activity [2], this collection of enterprises and organisms will be able to finish the investigation and development stages in the new contents of broadcasting audiovisual technologies, and will realise valuable contributions to the main standardization bodies in an industrial forum context, collaborating with technical proposals in the definition of the new DVB-T standard in the recent years, named DVB-T2.
It is expected that the generated research results of the consortium will be immediately applied, by means of generating pre-industrial outcomes.
Other objectives of the FURIA project are:
Establishment of relationships with other national and European projects, which will allow the enrichment of Spanish technological level. Contributions to forums and European standards, to grow up Spanish technological consortium acknowledgment and to have influence in the standardisation section.
CHAPTER 1. INTRODUCTION TO DVB-T 3
CHAPTER 1. INTRODUCTION TO DVB-T
In the context of video broadcasting networks, analogue technologies have been already replaced by digital transmission technologies. This transition from analogue to digital through the implantation of the DVB-T standard provides the networks with certain advantages such as a better bandwidth exploitation, more robustness in front of noise, and more advantages that are reflected in the improvement on the image and sound, and also includes new applications to the users.
This chapter aims at explaining the main changes from the analogue technology to the digital one, as well as the theoretical background necessary to understand the main features of the DVB-T standard.
1.1 Changing to a digital mode
Digital television arises due to the fact that provides better characteristics than the analogue television. The old method had less spectral efficiency, as every single image was transmitted, in order to improve this mechanism, digital television introduces MPEG-2 compression, which sends the changes of the images and thus, much less information. Due to this fact, the required bandwidth is reduced and then on the same channel several programs can be multiplexed, or they can be transmitted with high definition, multimedia, interactivity can be included, etc. The spectral efficiency is then much higher on digital systems. Another set of common problems found in analogue television are ghost of images due to multipath in the radio channel, noise from weak signals, which degrade the quality of the signal and sound, etc. All this is efficiently solved by a digital transmission.
Moreover, the fact of changing to the DVB-T and DVB-T2 standard does not imply an increment on deployment cost, as most part of the already existing infrastructure can be re-used.
1.2 Digital television around the world
The change towards digital television is being done all around the world. Each country, or a set of countries have decided which standard are going to adapt in their territory. For instance in Europe it is used the DVB-T standard, but there are other possibilities as it is shown in Fig 1.1.
4 Planning optimization software tool for DVB-T and DVB-T2
Fig. 1.1 Different standards around the world
As it can be seen, nowadays there are four main standards around the world:
‐ ATSC in north America. ‐ ISDB-T, which is a Japanese standard also widely adopted in south America. ‐ DMB-T/H in China. ‐ DVB-T in Europe.
In March 2006 the consortium leading the DVB project decided to study options for an upgraded version of the standard. In June 2006, a formal study group named TM-T2 (Technical Module on Next Generation DVB-T) was established by the DVB Group to develop an advanced modulation scheme that could be adopted by a second generation digital terrestrial television standard, to be named DVB-T2.
According to the commercial requirements and call for technologies issued in April 2007, the first phase of DVB-T2 will be devoted to provide optimum reception for stationary (fixed) and portable receivers (i.e., units which can be nomadic, but not fully mobile) using existing aerials, whereas a second and third phase will study methods to deliver higher payloads (with new aerials) and the mobile reception issue. The novel system should provide a minimum 30% increase in payload, under similar channel conditions already used for DVB-T.
1.3 DVB-T2 physical layer overview
Although this project is study from the system level viewpoint, it is necessary to explain some clue ideas of how the physical layer of the system works, in order to understand the main problems that can be found when deploying the network.
CHAPTER 1. INTRODUCTION TO DVB-T 5
Fig. 1.2 Time and frequency representation of the SC and OFDM. In OFDM, N data symbols are transmitted simultaneously on N orthogonal subcarriers [3]
Digital broadcast television is based in OFDM, a well-known transmission technique widely used in communications on the last years. For instance, it is the technology adopted by ADSL, some version of the IEEE 802.11 standar, IEEE 802.16, LTE, data transmission in power-lines and many other standards. OFDM is a very powerful transmission technique. It is based on the principle of transmitting simultaneously many narrow-band orthogonal frequencies, named subcarriers. The number of subcarriers is often noted as N. These frequencies are orthogonal to each other which (in theory) eliminates the interference between channels. Each frequency channel is modulated with a possibly different digital modulation. The frequency bandwidth associated with each of these channels is then much smaller than if the total bandwidth was occupied by a single modulation, which is known as the Single Carrier (SC) (see Fig. 1.2). Having a smaller frequency bandwidth for each channel is equivalent to greater symbol time (N times longer) and then better resistance to multipath propagation (with regard to the SC option). Better resistance to multipath and the fact that the carriers are orthogonal allows a very high spectral efficiency. For these reasons, OFDM is often presented as the best performing transmission technique used for wireless systems.
1.3.1 Basic principle
OFDM makes use of the properties of the Discrete Fourier Transform (DFT) to generate the final signal without the need of one oscillator per sub-carrier. In particular, to speed the calculus, the Fast Fourier (FFT) algorithm is used instead. The FFT can be applied as long as the number of points in the sampled signal is a power of 2 (e.g. N = 256). This condition is easily imposed by the DVB-T standard (or any other standard). The IFFT is the Inverse Fast Fourier Transform operator and realises the reverse operation. OFDM theory shows that the IFFT of magnitude N, applied on N symbols, realises an OFDM signal, where each symbol is transmitted on one of the N orthogonal frequencies. The symbols are the data symbols of the type QPSK, QAM-16 and QAM-64.
If the duration of one transmitted modulation data symbol is Td, then Td = 1/f, where f is the frequency bandwidth of the orthogonal frequencies. As the modulation symbols are transmitted simultaneously, 6 Planning optimization software tool for DVB-T and DVB-T2
Fig. 1.3 Presentation of the OFDM subcarrier frequency
This duration, f, is the frequency distance between the maximums of two adjacent OFDM subcarriers, as it can be seen in Fig. 1.3. This figure shows how the neighbouring OFDM subcarriers have values equal to zero at a given OFDM subcarrier maximum, which is why they are considered to be orthogonal. In fact, duration of the real OFDM symbol is a little greater due to the addition of the CP.
1.3.2 Time domain physical layer considerations
After application of the IFFT a Cyclic Prefix must be added at the beginning of the OFDM symbol as it is shown on Fig.1.4. The CP allows the receiver to absorb the delay spread due to the multipath and to avoid intersymbol interference (ISI). The CP that occupies a duration called the Guard Interval (GI) is a temporal redundancy to give continuity to the OFDM signal. On the other hand, including this prefix reduces the effective data rate because during this time no new information is transmitted.
Following the DVB organization rules that are shown in the recommendations, there are several possible GIs. The operator may choose among these options considering the radio channel features in its particular deployment. Table 1.1 remarks the recommended lengths. Note that mode 8k and 2k stand for the two different periods of symbols that are considered in the standard: 896 s and 224 s respectively. The GI is always given as a fraction of this value. This fraction represents de percentage of time in which no new information is transmitted. Consequently, for the same ‘degree of inefficiency’ the 8k mode allows deploying larger SFNs.
Fig. 1.4 Cyclic Prefix added at the beginning of the OFDM symbol CHAPTER 1. INTRODUCTION TO DVB-T 7
Table 1.1 Specified length of the guard interval [4]
For instance, in this project it has been applied the OFDM 8k mode, with 56μs of GI. The longest GIs are suitable for networks with longer distances between TX stations, as for example with national SFNs. The shortest intervals are suitable for regional or local broadcast transmissions. In summary, the longer the guard interval is, the less interference will appear, but less information is sent.
1.4 Radio planning of DVB-T systems
Radio planning DVB-T networks allows basically two types of deployment, on the one hand the classic Multi – Frequency Networks (MFN), and on the other hand Single – Frequency Networks (SFN).
Conventionally planned DVB-T networks consist of TXs with independent programme signals and with individual radio frequencies. Therefore they are also referred to as MFN. In order to cover large areas with one DVB-T signal a certain number of radio- frequency channels is needed. The number of channels depends on the robustness of the transmission, i.e. the type of modulation associated with the applied channel code rate and on the objective of planning, (full area coverage or coverage of densely populated areas only). As the robustness of a broadcasting system is generally expressed in terms of protection ratios, one might expect that the number of channels needed for DVB-T is significantly lower than for analogue broadcasting as the protection ratios are generally lower in the digital case. However, due to some other phenomena, the number of radio-frequency channels needed for conventionally planned DVB-T networks tends to be in the same order as with analogue TV systems. The frequency resource expressed as the number of channels needed to provide one signal at any location is far higher with MFN than with SFN. Nevertheless, one of the advantages that MFNs have is that it is not necessary to have synchronous emissions as one area is only served by one TX.
In a SFN, all TXs are synchronously modulated with the same signal and radiate on the same frequency. Due to the multi-path capability of OFDM with its GI, signals from several TXs arriving at a receiving antenna may contribute constructively to the total wanted signal. However, the limiting effect of the SFN technique is the so-called self- interference of the network. If signals from far distant transmitters are delayed more than allowed by the GI they behave as noise-like interfering signals rather than as wanted signals. The strength of such signals depends on the propagation conditions, which will vary with time. The self-interference of an SFN for a given transmitter spacing is reduced by selecting a large GI. In order to keep the redundancy due to the GI down to a reasonably low value (25 %), the useful symbol length has also to be large given the transmitter spacing in most European countries. Thus the 8k-mode was introduced. On the other hand a smaller GI would lead to a higher number of TXs. 8 Planning optimization software tool for DVB-T and DVB-T2
With the SFN technique large areas can be served with a common multiplex at a common radio frequency. Therefore the frequency efficiency of SFNs appears to be very high compared to MFNs. Gaps in the coverage area of an SFN are easily filled by adding a new transmitter or repeater without the need for additional frequencies.
In conventionally planned networks and particularly in single TX situations, a common way to achieve service continuity at a high percentage of locations is to include a relatively large fade margin in the link budget and thus to increase the transmitter power significantly. However with omnidirectional reception in SFNs, where the wanted signal consists of several signal components from different transmitters the variations of which are only weakly correlated, fades in the field strength of one transmitter may be filled by another transmitter. This is translated into a receiving gain, and therefore transmitters with the SFN technique should be able to transmit with lower power.
As it has been explained on the problematic, this project has adopted the SFN technique as it seems to be the most efficient when deploying a DVB-T network and in fact is the option used in Spain at several geographical levels (national, regional, local).
1.4.1 Coverage computation
In order to make possible a feasible analysis of the DVB-T/T2 coverage, it is necessary to compute a link budget to guess the amount of power required at the receiver. It is important to fix the initial level of carrier to interference ratio (CIR) that the receiver must achieve in order to have a good visualization of TV, this value was extracted from [4], and placed as an input data in the link budget.
The link budget takes into account all the losses or degradation the signal suffers since it is emitted until arrives to the receiver end. Depending on the sophistication that is desired to provide to the calculation more or less parameters can be taken into account. For instance, in this project, environments focused on fixed scenario have the same link budget than the portable ones, as it is quite difficult to combine several parameters on a single receiver in Sirenet.
The whole link budget is placed in Appendix A, however below these lines there is a summary in Table 1.2 with the most relevant aspects.
CHAPTER 1. INTRODUCTION TO DVB-T 9
Table 1.2 Link budget results
Parameter Units Result Band Band V Receiving Condition Fixed antenna (outdoor 10 m) Frequency f [MHz] 800 Boltzmann Constant k [J∙K‐1] 1,38E‐23 Bandwidth B [Hz] 7,60E+06 Temperature T0 [K] 290 Thermal noise power Pn,th [dBW] ‐135,2 Receiver noise figure F [dB] 7 Total noise power Pn [dBW] ‐128,2 Minimum carrier to noise ratio CNR [dB] 2 8 14 18 26 required by system Minimum receiver signal input Ps min ‐126,2 ‐120,2 ‐114,2 ‐110,2 ‐102,2 power [dBW]
In order to assure an 18 dB of carrier to noise ratio, the minimum power received must be almost -110 dBW, which corresponds to -80 dBm. Based on these results every pixel that receives a value of power beyond this threshold is considered to be covered by one or more transmitters.
Regarding the coverage area, the one selected is the entire region of Catalonia. The TXs placed in this area are a total of 180, which is a very high number taking into account that the scenario must be optimized, so in the case the whole Catalonia territory is set to be analysed, the simulation time would be prohibitive, in this case different pieces of terrain are selected. As for the scenario, the pixel resolution may vary depending on the exact detail that the results are desired, but it is necessary to mention that, if this resolution is set too small the execution time is going to rise. In this project it is considered that on each pixel there’s placed a receiver sharing all the same characteristics.
In DVB-T there are basically two types of receiver defined, one that catches the transmitter whose echo first arrives, and the other selects the transmitter taking into account the most powerful signal. This is more detailed when explaining all the sets of scenarios prepared for the simulations. The TX’s information is more specific, as each one has different values of altitude or position. However, the emitting power of all the TXs is considered the same, although it is known that in real conditions each TX can have its own configuration of power, tilt, delay value, etc.
1.4.2 Interferences and their impact on coverage
Several types of interferences can interact and contribute negatively on the received signal. For instance the same signal received, can have one or more echoes due to the multipath phenomenon. DVB-T standards offer protection against these, due to the orthogonal frequency multiple division, every frequency carrier is divided into a subset of subcarriers, being smaller in frequency bandwidth terms. Even more protection is added with the GI, which can protect the receiver from other types of interferences. In a SFN network all the TXs send information at the same frequency, being possible the reception of one or more signals coming from different sources. If this signal echoes are received inside the GI, then it is said that the signal contributes positively, on the 10 Planning optimization software tool for DVB-T and DVB-T2 contrary will degrade the signal quality. Moreover, other services operating at near frequencies can have an interfering contribution to the signals.
1.4.2.1 Deployment of an SFN network
As it is previously explained in sections above, the SFN network allows to complement the received signal with the ones coming from other transmitters. However, this idea can be extrapolated into the analogue, which means that one pixel that initially is covered by one or more transmitter, loses its quality due to the strong self-interference that suffers from the transmitters placed nearby. For this reason SFN networks must be optimized, in order to reduce these levels of interference and make sure that all the signals coming to a given receiver perform a good quality one, instead of destroying it.
It is clear that the longer the GI is, the easier the reduction of self-interference, as many echoes will arrive inside the CP. However this also implies a less efficient transmission since no new information is contained in the added interval and so the effective data rate is reduced. Besides, mobile television is gaining focus particularly in the context of the DVB-T2 standard, and long symbols with large GIs are much more sensitive to Doppler Effect. A good system design implies as short as possible GIs while maintaining sufficient multipath protection.
1.4.3 Solutions to improve coverage area
Several solutions arise in order to reduce the impact of the self-interferences in SFN networks. There exist basically two types of variables, the ones which are uncontrolled by the operator and those that are susceptible of optimization.
On the first classification of variables one can contemplate the propagation environment and the configuration of the OFDM receivers. The fact that one receiver is set to one type or another changes the coverage and the impact on the interferences, this is studied in further chapters in this project. However, this cannot be managed by the operator, the variables that can control are those regarding the transmitters such as, its geographic position, the configuration of the radiant system (radiation pattern, downtilt, nullfilling techniques), the transmission power or the static delays. Moreover, in a context of operative DVB-T networks and in some cases in the beginning of a transition towards DVB-T2, powers, antennas and positions (in this order) are increasingly more static and unlikely to be dramatically changed.
Given this, this project is focused on the optimization of the static delays of the transmitters in an SFN network. The final objective is then, to reduce the self-interfered areas and with its consequent increase in coverage area. This action can be performed manually, but it turns very difficult to find an optimal solution due to the interdependencies of the variables. For this reason, this project proposes a technique that optimizes a set of transmitters in a given area, and searches for the set of delays that minimize the areas affected by ISI. CHAPTER 2. DESIGN OF AN ALGORITHM TO MAXIMIZE COVERAGE 11
CHAPTER 2. DESIGN OF AN ALGORITHM TO MAXIMIZE COVERAGE
2.1 Problem description
The problematic analysed in this project, as it was presented in the section before, is to improve the coverage area by means of changing the delays of all the transmitters so the maximum number of pixels under a test area are improved. First of all, let analyse with a simple example which is the impact of changing the internal delay of a repeater.
Let consider a canonical scenario in which two TXs (L on the left and R on the right) are deployed in a flat terrain. Under these circumstances, the border between the areas with the second contribution falling inside or outside the GI is given by the locus of points where the difference of the distances to the two TX is a constant, that is a hiperbola with both TXs as foci. However, not the full area on the left of the left semi- hiperbola and on the right of the right semi-hiperbola are necessarily out of coverage. As long as the CNR is good enough, other contributions can be received out of the GI. Self-interfered areas can be modified by means of changes on static delays. Thus, for example, if the internal delay of L is increased, then R has virtually got closer and consequently the left semi-hiperbola is reduced (eventually eliminated). Conversely, this action has a negative effect on R, because now L has been virtually moved further away and so the self-interfered area on the right is increased. This simple modification could be useful for example in an environment in which R transmits with a higher power and so can cope with the signal from L causing interference. More details on this can be found in [1].
To be able to solve this problem in a case with many TXs, it is necessary to find the combination of delays such as the highest number of pixels is improved. The simplest possibility to solve the problem is just using a brute force search. Of course this cannot be done, the search requires a prohibitive computational time. In particular, assuming a finite set of m possible delays and n TXs, each value can be assigned to every TX with repetition. The assignment of the same values to different TXs also changes the solution (TXA-delay1, TXB-delay2) (TXA-delay2, TXB-delay1). Thus, the solutions space is a variation with repetition of m values taken from n in n: