Planning Factors for Digital Local Broadcasting in the 26 MHz Band

BTS-10-162 1 Planning Factors for Digital Local Broadcasting in the 26 MHz Band I. Peña(1), Member, IEEE, T. Lauterbach(2), P. Angueira(1), Member, IEEE, A. Arrinda(1), Member, IEEE, J. M. Matías(3), D. de la Vega(1), Member, IEEE, M. M . Vélez(1), Member, IEEE, C. Re(4), F. Maier(5) Abstract—The 26 MHz band is currently allocated for international broadcasting by means of the ionospheric propagation. However, sky waves are strongly dependent on the ionosphere conditions and signal reception is not always possible. As a consequence, this HF frequency range is scarcely utilized and an optimization of its use is required. The DRM (Digital Radio Mondiale) standard provides enough robustness to enable reliable audio reception taking advantage of the ground wave propagation at these frequencies. According to this, local coverage is viable in this part of the spectrum, as proved the field trials and studies carried out in the last years. Nevertheless, it is still necessary to establish the general guidelines for the set up and consolidation of the corresponding digital networks. Thus, the planning factors and procedures for digital local broadcasting services in the 26 MHz band are defined in this paper, along with reference parameter values and recommended system configurations for this new application. Index Terms—26 MHz, digital broadcasting, Digital Radio Mondiale, propagation, ground wave, sky wave, planning factors. I. INTRODUCTION T HE digitization of audio broadcasting systems is underway all over the world. The emerging digital radio technologies provide higher sound quality and value added services if compared with the analogue ones, optimizing this way the use of the radio frequency bands. Such improvement is especially meaningful in the case of the 26 MHz band, a part of the spectrum that has been allocated for international broadcasting by means of the sky wave propagation. However, the reception of this type of services is only reliable when the ionization levels of the ionospheric layers are very high. This phenomenon depends Manuscript received August 20, 2010; revised November 24, 2010. This work was partially supported by the Basque Government and the University of the Basque Country (UPV-EHU). (1) I. Peña, P. Angueira, A. Arrinda, D. de la Vega and M. M. Vélez are with the Dept. of Electronics and Telecommunications, UPV-EHU, Bilbao, Spain {ivan.pena, pablo.angueira, amaia.arrinda, david.delavega, manuel.velez}@ehu.es. (2) T.Lauterbach is with the Dept. of General Sciences, Georg-Simon-Ohm Fachhochschule - University of Applied Sciences, Nürnberg, Germany (Thomas.Lauterbach@ohm-hochschule.de). (3) J. M. Matías is with the Dept. of Telecommunications, UNAM, Mexico D. F. (JMatiasM@hertz.fi-b.unam.mx). (4) C. Re is with Radio Maria, Turin, Italy (reclaudio@alma.it). (5) F. Maier is with the Institute of Communications Technology of the Leibniz University of Hanover, Germany (friederike.maier@ikt.unihannover.de). on the solar activity, which varies over a sunspot cycle of 11 years. Consequently, the major part of this time, long distance large area services are provided at lower HF frequencies and therefore, there are long periods where the 26 MHz band is not utilized. In order to exploit this frequency range more efficiently, in the last years several broadcasters have promoted its use for the transmission of local broadcasting services by means of the ground wave propagation. This new application is based on the DRM (Digital Radio Mondiale) technology, a standard accepted for the LF, MF, HF and FM frequencies [1]-[2]. Taking into account the bandwidth and robustness options of the system, this is the most appropriate solution to increase the activity in the 26 MHz band. Therefore, it represents a unique opportunity to encourage this short-wave part of the spectrum, currently a concern for most international frequency regulatory bodies. DRM 26 MHz services were first put in practice within the RADIATE European research project [3]. Later on, a set of empirical experiments have been carried out in Switzerland [4], England [5], Germany [6], Australia [7], Mexico [8] and Brazil [9]. Besides, some studies focused on the feasibility, performance and operation thresholds of the system such as [10]-[11] have been published in different conferences and journal papers. However, as a final step, it is still necessary to establish the guidelines for the introduction and roll-out of digital local broadcasting services in this frequency range. Thus, the corresponding planning methodology and factors as well as the associated parameters and values are presented in this paper. Also, additional information will be given on key issues to be considered for the set-up of real networks. This paper is organized as follows. Section II provides a practical description concerning the propagation modes, noise values and radio regulation aspects for the frequencies of interest. Section III deals with different application scenarios and transmission features of the system. In Section IV, the minimum requirements for correct reception, the time and location signal variability, the flat fading effect and other aspects influencing the service area for both static and mobile reception are defined. On the other hand, the interference conditions due to both, the regular and sporadic ionospheric propagation of DRM local services are stated in Section V. Finally, a prediction technique, based on empirical studies, for obtaining DRM local coverage estimations in the 26 MHz band is suggested in Section VI. BTS-10-162 2 II. THE 26 MHZ BROADCASTING BAND A. Propagation Modes The 26 MHz or 11 meter band is the uppermost HF broadcasting frequency range and comprises 430 kHz distributed between 25.670 MHz and 26.100 MHz. Local broadcasting in this part of the spectrum is not trivial since depending on the case, ground waves as well as sky waves can be received within a specific coverage area. A ground wave radio signal can travel to a particular reception location by means of two different propagation mechanisms, the space wave and the surface wave. Until the development of digital radio standards, space waves had been typically considered for the transmission of audio broadcasting services in the VHF band (FM services), but not in the 11 meter band. Experiments in the past have demonstrated that it is possible to provide DRM 26 MHz local coverage this way. However, although similar effects to the ones existing at higher frequencies, such as multipath and signal diffraction, exist in this HF frequency range [12], the corresponding space wave propagation properties can be different as the Fresnel ellipsoid is wider in this part of the spectrum. On the other hand, surface waves propagate along the earthatmosphere boundary following the contour of the smooth, non-irregular, spherical soil [13]. This type of signals can bend around obstacles with dimensions lower than one wavelength by means of diffraction and suffers higher or lower attenuation depending on the frequency, polarization and terrain properties. As a consequence, they can contribute to the field strength in the vicinity of a transmitter operating in the 26 MHz band. Nevertheless, the higher the frequency, the shorter the skin depth of the wave and thus, the soil absorption increases. For this reason, in most cases, surface waves in this part of the spectrum are highly attenuated within a short distance from the transmitting location. Apart from this, the propagation of sky waves is also possible in the 26 MHz band. This phenomenon is influenced by the ionization of the ionosphere, a region of the atmosphere usually modeled as a set of layers with different electron density levels [14]. In this case, the higher the frequency, the more likely the wave will pass through this means, since the amount of free electrons necessary for signal refraction back to the earth must be greater. According to this, the reception of ionospheric signals in the 26 MHz band is only possible when the electron density levels are very high. This is likely to occur in the F layer during periods of maximum solar activity (Regular Ionospheric Propagation) or sporadically in some areas of the E layer (Sporadic Ionospheric Propagation). For both situations, long distance propagation is expected and consequently, the band is used for the transmission of international broadcasting services. Taking this phenomenon into consideration, sky wave propagation is a potential source of unwanted co-channel interferences between local audio broadcasting signals in the 26 MHz band. Therefore, it should be also considered for planning DRM services in this frequency range, as explained in Section V. Fig. 1. Sky wave interference between digital broadcasting networks providing local services in the 26 MHz band Fig. 1 depicts how the transmitter of the first local network can interfere in the coverage area of the second network because of the ionospheric signal propagation. Different factors such as service availability, ionization of the ionosphere, terrain irregularity and reception environment have influence on the ratio between the wanted signal due to ground wave propagation and the harmful sky wave interference. Hence, the relationship between the main service scenarios and the network characteristics will be described in Section III. B. Noise The reliability of digital audio services depends on the noise levels at the corresponding working frequencies, and the impact of this electromagnetic phenomenon is usually characterized by means of the SNR (Signal to Noise Ratio) [15]. As a first approximation, this parameter is influenced by the intrinsic noise level of the receiving system. For an AM reference receiver working in the HF broadcasting bands, the intrinsic noise is defined as the difference between a sensitivity of 40 dBuV/m and a required C/N (Carrier to Noise Ratio) for correct signal reception of 36.5 dB. The result is therefore a reference noise value of 3.5 dBuV/m. According to this, and taking into account that the IF bandwidths for AM and DRM signals are different (typically 8 kHz and 10 kHz respectively), the noise floor of the digital reference receiver Ni can be estimated as [16] : N i (10kHz ) = 3.5dBµV / m + 10 log 10kHz ≈ 4.5BµV / m 8kHz (1) In some cases, this value could be lower than the external noise existing in certain frequency bands and reception environments and thus, radio noise can also set a limit to the performance of audio broadcasting systems. The International Telecommunication Union states that external noise in the 26 MHz band is mainly due to human activities. Following, the definitions given in the ITU-R Recommendation P.372, MMN (man-made noise) values can be obtained for different types of environments considering a noise factor Fa or field strength units [17]. According to this, MMN data were calculated for a frequency f = 25.885 MHz and a bandwidth b = 10 kHz. BTS-10-162 3 From the previous calculations, it was deduced that the DRM signal reception in the 26 MHz band is limited by the man-made noise in “City” and “Residential” environments but not in “Rural” and “Quiet Rural” environments, where the highest mean noise level corresponds with the 4.5 dBuV/m calculated for the digital reference receiver. Nevertheless, if the time and location variability is considered, MMN levels above this value are also possible even in “Rural” locations. On the other hand, data of the ITU-R Recommendation P.372 are based on measurements carried out in the 70’s and an update could be necessary as the man-made noise sources have increased in the last years [18]. For this reason, a measurement campaign to get new empirical MMN values was carried out in Nuremberg during spring of 2009. Samples were taken at different frequencies within or close to the 26 MHz band and the corresponding mean noise figures (Fam) for both metropolitan and rural environments were calculated. In addition, upper decile deviations that account for the time and spatial variability (Dut and Dus) were studied. For comparison purposes, the results are summarized in Table I, where also statistical values previously obtained for urban areas of Hanover [6], Mexico [19] and Brazil [19] by the authors of this paper have been included. Apart from this, Fam and Dus values of 42.9 dB and 9.4 dB respectively were also obtained for the total set of urban locations where measurements took place. This means an increase of 5.2 dB in the mean MMN level and a very similar upper decile deviation when compared to the estimations derived from the ITU-R Recommendation P.372 for “City” environments. Regarding the rural locations, empirical noise levels were mainly due to the unwanted energy generated by the measurement system. Thus, it was definitively concluded that the most limiting factor for correct demodulation of DRM 26 MHz services in this type of areas can be due to the electromagnetic impairments caused by the receiver. TABLE I EMPIRICAL NOISE VALUES Measurement Area Freq. (MHz) Fam(*) (dB) Time Variation Dut (dB) C. Radio Regulation The 26 MHz band is allocated by the ITU-R (International Telecommunication Union - Radiocommunication Study Groups) to international long-distance large-areas coverage by means of ionospheric propagation [20]. Consequently, it is subject twice a year to seasonal planning procedures that involve administrations, broadcasters, FMOs (Frequency Management Organizations) and 4 regional HF coordination groups: • ASBU (Arab States Broadcasting Union) • ABU-HFC (Asia-Pacific Broadcasting Union - High Frequency Conference) • HFCC (High Frequency Coordination Conference) • URTNA (African Regional Coordination Group) Like in other HF broadcasting bands, the planning process requires multilateral coordination between FMOs. These organizations are free to select the frequencies that satisfy their broadcasting requirements so that interference problems could exist in the band. Thus, the ITU encourages the regional coordination groups to manage informal face-to-face meetings between FMOs in order to solve or minimize potential incompatibilities. After these meetings and once the compatibility between services has been analyzed, the corresponding results as well as the seasonal HF broadcasting schedule are published by the Radiocommunication Bureau [21]. Frequency allotments in the 26 MHz band are internationally coordinated due to the nature of the longdistance broadcasting services. However, a local use of this frequency range requires a different regulatory scenario. Fig. 2 shows a simplified frequency allocation based on coverage cells. Regular hexagons of radius R represents the area around a transmitting point where sky wave signals cannot propagate. On the other hand, the transmitter locations separated a distance D define a rhomboid region where, in order to avert ionospheric interferences, the frequency reuse would not be possible Spatial Variation Dus (dB) CITY ENVIRONMENT Mexico 25.620 41.0 N/A 7.3 Brasilia 25.885 43.5 2.0 5.0 Hanover Nuremberg 26.045 41.3 N/A 0.6 26.000 49.5 1.8 5.4 26.020 52.5 2.1 3.4 26.300 49.0 5.1 2.0 RURAL ENVIRONMENT Nuremberg (*) 26.000 < 33.4 (**) > 1.6 (**) > 3.0 (**) 26.020 < 37.0 (**) > 2.9 (**) > 2.3 (**) 26.300 < 34.5 (**) > 3.8 (**) > 5.0 (**) Values calculated for a signal bandwidth of 10 kHz Upper limit of the external noise close to the receiver internal noise (**) Fig. 2. Generic cellular system over Europe. R: Cell Radius, D: Co-channel distance BTS-10-162 4 According to this, the number of channels N that are needed to avoid sky wave interferences can be obtained from Equation 2, where SR and SH are the areas of the rhombus and the hexagon respectively [22]. N= SR 3D 2 / 2 D2 = = 2 S H 3 3R / 2 3 ⋅ R 2 (2) Taking into account that ionospheric signals in the 26 MHz band can be received at distances between 800 km and 4000 km [23] from the transmitter, at least 34 different channels would be necessary to establish an international regulatory approach free of interferences. This is viable, since this part of the spectrum can accommodate up to 43 channels for DRM signals of 10 kHz bandwidth. However, in that case, only one or two frequencies could be utilized within areas of 400 km radius and consequently, this solution would prevent a comprehensive use of the band. A more practical approach based on local regulations is proposed in this paper. The frequency planning would be carried out on a national basis adopting similar procedures to the ones defined for other broadcasting services in the VHF band [24]. According to this: • Each country will apply its own planning procedures. • Frequency coordination will be required in neighbouring regions. • A maximum ERP limit will be set for any station in any region. • The service areas will depend on the available channel to interference ratios (C/I). Local radio regulation is not a new concept and for example, it has been previously considered for MW band channels around 1500 kHz where signals can cause and suffer the same interference problem [25]. However, by regulating the 26 MHz band in this way, broadcasters should assume and accept sky wave interferences limiting the local coverage in certain periods and under specific conditions; the same as happens with TV services in bands I and II [26]. Even so, there would still be an area where a digital grade quality can be ensured. Its size will depend on factors such as the reception environment, transmitter antenna height and radiation pattern, maximum ERP limits, season of the year, region of the world and number of stations per country. III. SERVICE SCENARIOS AND TRANSMISSION FEATURES The first step for planning local broadcasting services in the 26 MHz band is to identify the application scenario in terms of ionospheric signal propagation and local reception conditions. This way, it will be possible to determine the optimum transmission characteristics that minimize the potential sky wave interferences and maximize the coverage area. The probability of ionospheric interference will depend on the service availability and the ionization level of the ionosphere. The first one differentiates between permanent and temporary services. That is, services limited in time, ranging from a whole season (spring, summer, autumn and winter) to a limited number of days (festival, fairs, conferences, and similar events). While permanent services are likely to cause sky wave interferences during specific time periods throughout a solar cycle, disturbances due to temporary services in the 26 MHz band are less probable if the electron density levels are low enough. This is possible for specific SSN (Sun Spot Number) values, months of the year and hours of the day and thus, the ionospheric ionization is also a parameter that should be taken into account. Regarding the local reception conditions, the orography and environment within the coverage area have to be considered. Irregular terrain profiles can produce LoS (Line of Sight) obstruction and diffraction attenuation. Urban environments imply higher man-made noise values, shadowing effects and multipath propagation. Therefore, lower ERP (Effective Radiated Power) levels would be required in flat and rural areas as they favor the service reliability. Combining the previous factors, four different service scenarios can be defined. As depicted in Fig. 3, each one of them entails low or high probability of interference propagation and favorable or unfavorable local reception conditions. A detailed study of the coverage area and the impact of the potential sky wave interferences will be necessary before setting-up a DRM 26 MHz broadcasting network. To do this, estimations of the received field strength levels with suitable prediction algorithms should be obtained for each specific case. However, some planning guidelines to select the transmission features for the general power requirements of each service scenario are described below. Experiments by the authors of this paper in Mexico [8] and Brazil [9] have demonstrated that an appropriate transmitter site is essential. Taking into account that the dominant propagation mode for local broadcasting in the 26 MHz band is the LoS tropospheric component, the coverage can be improved increasing the visibility over the target region. For this reason, high transmitter locations are especially recommended when the reception conditions are not favorable. On the other hand, the radiation pattern of the transmitting antenna can favor the local coverage and/or the existence of ionospheric interferences. Omnidirectional aerials could be an optimal solution for temporary services during low ionospheric electron density periods and transmitters placed inside the service area. However, they would not be appropriate when the probability of sky wave propagation is high. In that case, directive models that minimize this effect as much as possible are suggested to reduce this type of disturbances. However, this type of antennas represents a challenging design problem due to the electric length associated to the frequencies of the 26 MHz band. Finally, SFNs (Single Frequency Networks) allow higher service reliability and larger coverage areas when compared to the stand-alone transmitter operation. If the reception conditions are unfavorable, a single transmitter may not suffice to cover a whole target region keeping the power level lower than a specific threshold. BTS-10-162 5 Fig. 3. Service scenarios for local broadcasting in the 26 MHz band Two or more transmitters operating from different sites at the same frequency can fill possible coverage gaps and ensure high quality services without increasing the transmitted power level. Also, this network architecture might cause lower ionospheric interference levels if the transmitting sites and antennas are specifically chosen to avoid the radiation on the same ionospheric regions, so that only the signals that find high ionization levels are refracted towards the earth. Thus, high transmitting sites, directive antennas and SFN networks are the best choice to maximize the local coverage area and minimize the sky wave interferences. Nevertheless, these conditions might be too restrictive when the probability of interference is low and/or the local reception conditions are favorable. In those cases, the features could be relaxed in order to reduce the cost of the network. In this regard, a proposal based on the authors’ experience is included in Table II. Apart from this, the DRM standard specifies a set of technical parameters to define an optimal system configuration for the reception conditions of the coverage area. The source coding, MSC (Main Service Channel) modulation, protection level, OFDM (Orthogonal Frequency Division Multiplexing) mode and cell interleaving have to be appropriately selected in order to reach a compromise between robustness and available useful bit rate [1]. Several empirical tests have proved that modes A an B are the most suitable options to overcome noise, multipath, interferences and other potential impairments in the 26 MHz band [3]-[9]. System threshold estimations for DRM signals this way configured and different propagation channels are provided in the ITU-R Recommendation BS.1615 [16]. However, the channel models concerning the HF broadcasting frequencies account for the sky wave but not for the space wave propagation effects and consequently, appropriate values for planning DRM 26 MHz services are not provided. A more detailed explanation can be found in [10] and [19]. Taking into account this shortcoming, several measurement campaigns have been carried out in the past in order to determine the minimum SNR and field strength requirements for digital local broadcasting in this part of the spectrum. A summary of the corresponding results are included in next section. IV. FACTORS INFLUENCING THE SERVICE AREA TABLE II TRANSMISSION FEATURES Service Scenario Transmitter Site Transmitting Antenna Network Architecture S1 Optional Optional Single Tx High Optional Single Tx Optional Optional SFN High(*) Directive Single Tx Optional Directive SFN High Directive SFN S2 S3 S4 (*) High transmitter location to reduce diffraction attenuation and shadowing effects and transmit low power levels for interference minimization A. Minimum Requirements for Correct Reception One of the most important goals when planning a sound broadcasting network is to achieve an excellent audio signal quality. In the case of DRM services, the signals are structured in super frames of 400 millisecond duration and each super frame is typically divided in 10 audio frames of 40 milliseconds [1]. Thus, the reception quality can be evaluated by calculating the percentage of audio frames correctly received, a parameter commonly known as AudioQ or AQ [27]. Empirical tests during the developing stages and trials of the DRM standard have demonstrated that a listener perceives good quality if at least the 98% of the total received audio frames are correct. Under such assumption, minimum SNR and field strength threshold values for correct demodulation of DRM services in the MW bands have been obtained in the past BTS-10-162 6 [28]-[29]. However, lower thresholds as for example the 60% or 75% have been considered in other studies [11], [30]. Taking into account that the AudioQ is an objective parameter which does not distinguish whether an audio frame has been corrupted in a “high sensitive part” or in a “less sensitive part”, results of some subjective quality tests were presented in [31]. The main conclusion derived from the analysis was that an AudioQ threshold calculated by removing the dropouts accepted by the listeners does not differ significantly from the 98% and thus, this value can be accepted as the minimum quality required for a DRM service. Following this criterion, empirical SNR and field strength thresholds for a good quality reception of DRM 26 MHz services were computed from data collected during experiments carried out in Mexico and Brasilia. Table III shows the results obtained from [10], [19], [32] and [33] for both static and mobile reception with different system configurations. For the sake of the briefness, a reference code indicating the signal bandwidth (10 or 18 kHz), robustness mode (A or B), MSC modulation (16 or 64 QAM), protection level (0 or 1) and cell interleaving (Short or Long) has been considered to indicate the different combination of parameters considered during these trials. The static measurements that took place in Mexico revealed that 18k/B/16/0/L mode was the optimal configuration to be used in the 26 MHz band. Whereas values lower than the 98% of correct audio frames were calculated for the rest of combinations, the mean AudioQ associated to this mode for the total set of measurement points was 99.1%. Therefore, this was the only configuration tested in Brasilia. On the other hand, if the results given for both cities are compared, it can be concluded that the system threshold values depend on the reception environment. Brasilia presents more favorable reception conditions than Mexico and consequently, lower SNR and field strength values were required for obtaining a good quality service with the same DRM configuration. Besides, in this case, the mean AudioQ for the total set of measurement locations was a little bit higher. That is 99.73%. As explained in next subsection, an average window of 200 m is needed to separate the long and short term variations of a ground wave propagated signal in the 26 MHz band. Taking into account this, the study for mobile reception was done by dividing the measurement routes in different sections of this length and calculating the corresponding AudioQ values. This way, the reception thresholds that ensure the correct demodulation of a DRM service in the 90% of these sections were determined. For the 18k/B/16/0/L mode, minimum SNR values higher than the ones related to static reception were determined. The reason was that measurements with this configuration were affected by dropouts due to Doppler spread since the speed of the van where the receiving system was installed was above 100 km/h in some stretches of the measurement routes [19], [34]. Apart from this, other system threshold values have been determined from DRM mobile measurements in the 26 MHz band carried out in Hanover, Dorset, Nuremberg and Dillberg. In the case of Hanover [35], data were obtained along two different routes with 3 different configurations and UEP (Unequal Error Protection) for the more and the less sensitive part of the bit stream: 10k/A/64/0-1/L, 10k/B/64/0-1/L and 10k/B/64/1-2/L. SNR and field strength thresholds of 16 dB and 40 dBuV/m were calculated respectively for all the modes. The minimum requirements for correct reception derived from the mobile measurements that took place in rural environments of Dorset (United Kingdom) with modes A and B, signal bandwidth of 20 kHz, long cell interleaving and all the possible MSC modulations and protection levels [36], led to SNR thresholds ranging from 12 to 19 dB and minimum field strength levels from 31 to 39 dBuV/m. However these results might be a little bit pessimistic since they were calculated considering only the values associated to dropouts instead of applying the criterion of AudioQ ≥ 98%. Finally, from the trials carried out in Nuremberg and Dillberg [11], a field strength threshold of 23 dBuV/m to 26 dBuV/m was defined for the 10k/A/16/1/L mode but, on the contrary to the previous case, this result is optimistic since the criterion for correct mobile reception was fixed in an AudioQ of 60%. TABLE III SNR AND FIELD STRENGTH THRESHOLDS MEASURED IN MEXICO AND BRASILIA B. Signal Variability and Flat Fading While a 50% time coverage can be considered a reasonable target for analog broadcasting networks, digital systems require to fulfill higher percentages due to their brick-wall behavior. Therefore, planning factors based on empirical studies of the signal variations in time are necessary to satisfy this constrain [29], [37]. In the case of DRM 26 MHz services, an analysis based on static measurements gave as result correction factors to ensure a 99% time coverage for the 90% of locations [19], [38]. Global values were calculated for cities with local reception conditions similar to the ones existing in Mexico and Brasilia but also, corrections for different traffic conditions as well as the presence or absence of aircrafts in the case of Mexico were proposed. These factors are very useful to obtain more appropriate field strength predictions for planning the services once the median level has been determined. In this regard, a mathematical algorithm is suggested in Section VI. Apart from this, probability distributions of the field strength variability Static Reception Ref. Code Min SNR (dB) Min E (dBuV/m) Mobile Reception(*) Min SNR (dB) Min E (dBuV/m) MEXICO 10k/B/16/0/L 23 39 17 38 18k/A/64/1/L 25 45 21 38 18k/B/64/1/L 20 45 20 40 18k/B/16/0/S N/A N/A 20 38 18k/B/16/0/L 17 38 18 35 15 35 BRASILIA 18k/B/16/0/L (*) 13 37 Results from data taken at vehicle speeds rarely higher than 80 km/h BTS-10-162 around the median level were obtained for all the previous cases, so that the corresponding curves allow to determine correction factors for other percentages of time and 90% of locations. Outcomes of this research work are summarized in Table IV. One of the most interesting conclusions from these results is that the more favorable the reception characteristics are, the lower variability of the field strength level is. As seen in the table, a power increase of only 2.8 dB over the median value is required in cities like Brasilia to ensure coverage during the 99% of the time and for the 90% locations. However, 7 dB are necessary for the environmental features of Mexico. Regarding the traffic, although higher values were obtained in Mexico, a similar tendency is observed in both metropolitan regions. The values show that higher field strength levels are needed to achieve the same coverage area if the traffic conditions are critical. Even more, the presence of aircrafts will worsen the situation. Besides, it is deduced that a Nakagami distribution is the best option to model the time variability under isolated and non-dense traffic conditions. Nevertheless, for dense traffic cases or general coverage purposes, it is better to assume a Normal distribution. On the other hand, a mobile scenario in a DRM network implies a radio path where the transmitting terminal is stationary and the receiving unit is moving. In this case, the received field strength level will vary continuously with the location and the resultant signal r(x) will be composed of two different components: the slow fading m(x) due to local ground cover and path variations and the fast fading ro(x) associated to multipath effects [39]-[40]. These components can be separated by means of the generalized Lee’s method: an algorithm which obtains m(x) by applying a running mean to smooth the received signal and calculates ro(x) as the difference between r(x) and m(x). To do this, the size (2L) and number of samples (N) of the averaging window, as well as the distance between samples (d) have to be appropriately chosen for the frequency band under study [41]-[42]. These parameters and also the generalized Lee’s Method were analyzed in [19] and [43] obtaining that the most suitable values for a DRM 26 MHz signal are 2L=16.92λ, N=45 and d=0.38λ, being λ the wavelength associated to the operation frequency. After this, the corresponding variability of the slow and fast fading was determined by using the field strength data collected along the mobile measurement routes defined in Mexico. Taking into account that the variability of m(x) increases with the traveled distance, it was necessary to fix a reference path length of 700 m. This is the minimum value to be considered for the slow fading characterization, as lower distances would be very close to the size of the averaging window. The analysis revealed that a Log-Normal distribution is the best option to model this phenomenon so that correction factors to ensure a 90%, 95% and 99% coverage for the 90% of routes were calculated from the resultant curve, as observed in Table V. At first sight, these values might be considered very high. However, they are similar to the estimations derived from the ITU-R Rec. P.1546 [44] and to the data provided for other frequencies [45]. 7 The fast fading variability was characterized following a similar methodology so that the associated probability distribution and correction factors have been also included in Table V. However, it has to be emphasized that prior to obtain these results, the field strength values were normalized with regard to the slow fading in order to remove the dependency with the traveled distance. TABLE IV TIME VARIABILITY: PROBABILITY DISTRIBUTIONS AND CORRECTION FACTORS Correction Factor (dB) Type of Result Probability Distribution Type Parameters MEXICO 7.0 Normal m=1 σ = 0.25 Absence of Aircrafts 2.9 Nakagami Ω=1 m = 13.12 Presence of Aircrafts 5.0 Nakagami Ω=1 m = 9.68 Absence of Aircrafts 4.4 Nakagami Ω=1 m = 8.91 Presence of Aircrafts N/A N/A N/A Absence of Aircrafts 7.2 Normal m=1 σ = 0.29 Presence of Aircrafts 12.5 Normal m=1 σ = 0.27 Global Isolated Traffic Non Dense Traffic Dense Traffic BRASILIA Global 2.8 Normal m=1 σ = 0.11 Isolated Traffic 1.2 Nakagami Ω=1 m = 57.93 Non Dense Traffic 2.0 Nakagami Ω=1 m = 30.73 Dense Traffic 3.6 Normal m=1 σ = 0.13 TABLE V LOCATION VARIABILITY: PROBABILITY DISTRIBUTIONS AND CORRECTION FACTORS Type of Result Slow Fading Fast Fading Log-Normal Nakagami m = -0.06 σ = 0.35 Ω = 1.35 m = 0.92 Correction Factor for 90% of Coverage (dB) 3.9 8.7 Correction Factor for 95% of Coverage (dB) 5.0 12.1 Correction Factor for 99% of Coverage (dB) 7.0 19.8 Type Probability Distribution Parameters BTS-10-162 Apart from the slow and fast fading, flat fading is another phenomenon to be considered in the planning of DRM 26 MHz networks. The properties of the radio channel in local broadcasting are quite different from those of the ionospheric channel for which this technology was initially designed. In particular, the reflected signals that contribute to the received field strength level of the digital service come from the vicinity of the receiver and thus, present a delay spread in the order of a few microseconds instead of the several milliseconds associated to the temporal dispersion of sky wave propagation. This leads to a channel coherence bandwidth of about 50 kHz to 100 kHz in urban environments [19], [30], [34], which is a value significantly higher than the bandwidth of a DRM signal (4.5 to 20 kHz) [1]. As a consequence, mobile reception at low speeds or at stops (e.g. traffic lights and traffic jams) [36] and indoor reception without external antennas [11] can be affected by dropouts due to flat fading caused by the space wave multipath effects. This problem cannot be avoided by increasing the power [30] and therefore, a common solution to overcome it is the use of diversity techniques. Antenna diversity at the receiver site seems impractical because the separation between aerials has to be λ/4, that is, almost 3 m at the frequency of operation. However, Transmit Delay Diversity (TDD) does not imply changes on the receiver and can be implemented at a later stage of the network development [23]. This consists of transmitting the signal with different antennas and a certain time delay, large enough to increase the frequency selectivity of the composed channel but much less than the guard interval duration in order to avoid inter-symbol-interference [1]. Simulations and field tests to determine the improvement on the reception by generating the signal this way were presented in [23] for the 10k/A/16/1/L mode. The results of this work showed that with a SNR value of 21 dB, a TDD configuration reaches the bit error rate threshold required for guaranteeing the audio integrity (BER= 1×10–4) [46] whereas a single antenna implementation presents a BER of 6×10-3 for the same signal to noise ratio. Also, the trials carried out with the TDD scheme depicted in Fig. 4 demonstrated that correct reception was possible in spots affected by dropouts before using this technique, since this system turns the flat fading into selective fading and the DRM receivers can solve this problem. C. Coverage Aspects It is widely known that the coverage grade of a broadcasting network depends on the EIRP, since higher transmission power levels imply field strength and SNR values above the system thresholds at further distances from the transmitter. However, there are also other aspects that influence on the extension of the service area. The most relevant ones are: noise, reception environment, type of reception, DRM configuration and network architecture. Along with the transmitted power, noise is the other factor directly related with the SNR so that updated data of the field strength levels associated to this electromagnetic phenomenon were included in Section II.B. 8 Fig. 4. Block diagram of signal generation for TDD [23] The environment comprises the terrain profile between transmitter and receiver, the urbanization degree, and the outdoor/indoor condition in the reception point. Experiments carried out in Dillberg (Germany) demonstrated that shadowing from trees impair reception [30], field tests in Brasilia and Mexico concluded that the service quality is better in open environments than in dense urban environments [9]-[10], [19], and trials in Hanover and Australia suggested that the indoor field strength levels are between 6 and 12 dB below the outdoor values, [6]-[7]. Thus, LoS obstruction, multipath due to the presence of buildings and indoor attenuation reduce the coverage. The correct demodulation of DRM services is usually easier when the receiver is still [47]. Table VI includes some examples of the coverage associated to both static and mobile reception derived from measurements carried out in Nuremberg / Dillberg [11], [30] and Brasilia [33]. These values prove that in the second case, the service area can be significantly smaller due to the Doppler spread and the flat fading effect explained in the previous section [23], [34], [36]. DRM local broadcasting in the 26 MHz band can be significantly influenced by the transmission configuration. According to a set of experiments that took place in Dorset, the technical parameters that have more impact in the coverage degree are the MSC modulation and the protection level. Not so the OFDM mode, although slightly bigger service areas can be achieved with modes A despite being less robust than modes B. The reason is that the first ones have a smaller proportion of boosted pilot cells and thus, the signal to noise ratio on data cells is higher for a given overall SNR [36]. TABLE VI EMPIRICAL COVERAGE VALUES FOR STATIC AND MOBILE RECEPTION (*) Coverage (km) Static Mobile (*) EIRP (W) DRM Mode Dillberg 100 10k/A/16/1/L 40 20 Nuremberg 10 10k/A/16/1/L 15 3-5 Brasilia 650 18k/B/16/0/L 15 12 Measurement Area Results from data taken at vehicle speeds rarely higher than 80 km/h BTS-10-162 9 TABLE VII COVERAGE PERCENTAGES FOR DIFFERENT DRM CONFIGURATIONS MSC Protection Modulation Level Mode A Percentage of Coverage Mode B 0 92% 92% 1 90% 0 TABLE IX RECEPTION RANGES OF REGULAR IONOSPHERIC INTERFERENCES Study Case Transmitting Antenna Omnidirectional Directive SSN Level ≥ 25 ≥ 125 88% Month All year round Oct., Nov., Feb. & Mar. 82% 79% UT Hour 06:00 - 21:00 08:00 - 16:00 1 70% 64% 2 62% 60% 3 54% 54% 16 QAM 64 QAM TABLE VIII COVERAGE PERCENTAGES FOR SINGLE TRANSMITTER ARCHITECTURES AND SFNS Network Architecture Percentage of Coverage 10k/A/64/0/L 10k/B/64/0/L One Single Transmitter - Tx1 79% 78% One Single Transmitter - Tx2 66% 84% Without Delay 59% 81% Delay of 0.6 ms 92% 94% SFN Tx1 and Tx2 Nevertheless, the empirical values included in Table VII [36], show that there is not much difference between the coverage obtained with these modes when the rest of parameters remain invariant. This similarity was also observed during other measurement campaigns such as the carried out in Hanover [35], Nuremberg and Dillberg [11]. Besides, in this last case, it was determined that for mobile reception the long interleaving performs better than the short one. Finally, the network architecture is another relevant aspect that has a significant effect on the coverage. From data collected again in Dorset, it was concluded that a SFN network can give much better results than a single transmitter. However, it is necessary to transmit the signals with a certain delay [36]. As seen in Table VIII, a two transmitter SFN network without delay provides less coverage than each transmitter individually. The reason is that the signals interfere each other causing flat fading. However, in the previous section it was explained that a TDD configuration turns this effect into frequency selective so that the receiver can recover the signal and the coverage increases. V. SKY WAVE INTERFERENCE ISSUES A. Regular Ionospheric Interferences Sky wave interferences between DRM 26 MHz networks can be due to both, regular and sporadic ionospheric propagation. At the frequencies of interest, the first phenomenon mainly happens when the SSN (Sun Spot Number) is very high [30]. However, as the solar activity level was very low at the time of writing this paper, it has not been possible to analyze empirically this effect. In order to obtain some preliminary results, studies based on monthly median interference simulations that account for the reception of this type of disturbances during 50% of the time were carried out in the past by the authors of this paper [30], [48]. Nevertheless, as higher coverage percentages are recommended for the DRM services, a characterization based on upper deciles was also done to achieve more appropriate values for planning purposes [49]. Field strength estimations were calculated by using two prediction tools referenced by the ITU: REC533 and VOACAP [50]. Also a wide range of ionospheric paths in Europe and power levels of 200 W were considered for calculations with two types of antennas: one omnidirectional, so as not to restrict any of the possible regular ionospheric modes, and the other directive in order to quantify the reduction of sky wave emissions. The results were classified as a function of the SSN, month of the year and hour of the day in order to determine the reception periods associated to disturbances above a nuisance threshold. This limit was calculated from the protection ratio of the mode 10k/B/16/0/L [16] and assuming a situation where 60 DRM transmitters create long distance interference as a consequence of individual ionospheric contributions of 12 dBuV/m. The corresponding values are included in Table IX and suggest that the use of directive radiating systems can reduce significantly the sky wave interference problem. However, as previously pointed out, the design of directive antennas for the 26 MHz band is complicated due to the electric length associated to this frequency range. Thus, although this is not discarded, other solutions as the ones suggested in next section are required to stimulate the development of the corresponding DRM services B. Sporadic E interferences Since the sporadic ionospheric propagation is not directly related to the solar activity level, empirical studies of this phenomenon can be accomplished by setting up measurement systems of the DRM 26 MHz services existing on air. Nevertheless, for obtaining a set of representative values that account for this effect, an analysis of the interference probability as a function of the time and distance is previously required. This can be done by using the set of statistical data concerning the frequency of occurrence of sporadic E layers in different zones of the world and the field strength prediction method defined by the ITU-R Recommendation P.534 [51]. Their validity to predict potential mutual co-channel interference between DRM networks operating at the frequencies of interest is confirmed in [23] and values of the BTS-10-162 10 interference levels caused by 1 kW EIRP transmissions are also included. As seen in Fig. 5, these results are given for different transmitter-receiver distances. The arrows indicate the time percentages relating to the period of higher interference probability for which the critical frequency of sporadic E propagation (foEs) is equaled or exceeded in Europe and North Africa. It can be observed that transmitters at distances between 800 km and 2400 km imply significant interference levels during more than 1% of the time even for low values of foEs and thus, they could reduce the quality of a DRM 26 MHz service to AudioQ values below the 98% threshold. To prevent this, the authors of the study propose a EIRP between 10 W and 100 W for local broadcasting in this part of the spectrum. A power of 10 W was used during a test transmission at Nuremberg extending for several years and no long distance transmission for this transmission was reported, in contrast to the test transmission from nearby Dillberg using 100 W on which several reception reports from all over Europe were received [23]. Bearing in mind the previous results, a measurement campaign of the potential Es layer ionospheric interferences associated to digital local broadcasting networks operating in the 26 MHz band, was carried out in Bilbao (Spain) during late spring and early summer 2009. Samples of the electromagnetic disturbances due to the reception of DRM services coming from Hanover, Dillberg and Andrate respectively were taken from 4 May, 10:00 UT to 15 July, 23:00 UT following a twenty-four-seven recording schedule. In all the cases the transmission was done with a relative low EIRP level and from distances within the before mentioned range, that is 800 km to 2400 km. Sporadic E s signals were received during approximately the 25% of the total monitoring time. The probability of occurrence was higher in June and July and the most critical period ranged from 7:00 to 20:00 UT (Universal Time) hours. These results agree with the data derived from the ITU-R Recommendation P.534, as the predictions indicated that the situation gets worse in summer and during day time. On the other hand, it was observed that the field strength associated to this type of disturbances fits the normal distributions defined by the means and deviations included in Table X. Such values were calculated for the upper deciles of the interference levels received in each 1 hour measurement interval. As seen in the table, the nuisance threshold of 12 dBuV/m was exceed in all the cases, concluding that the DRM 26 MHz networks can interfere each other if the transmission power level is equal or higher than 130 W. This reinforces the idea of recommending low EIRP transmitters to minimize the problem. Again, installing SFNs would be a solution if one single transmitter is not enough to cover a specific target area. Even so, a scenario free of ionospheric interferences would not be possible and thus, a local regulatory approach, as the one described in Section II.C is recommended for digital broadcasting services in the 26 MHz band. VI. PREDICTION ALGORITHM FOR LOCAL COVERAGE Fig. 5. Prediction curves of the Es layer interference level TABLE X ES-LAYER INTERFERENCE VALUES DERIVED FROM MEASUREMENTS IN BILBAO Parameter Freq. (MHz) Transmission EIRP (W) Characteristics Distance (km) Received Field Mean (dBuV/m) Strength Deviation (dB) Hanover Dillberg (Germany) (Germany) Andrate (Italy) 26.045 26.000 26.010 130 202 500 1383 1289 895 24.64 27.65 29.55 3.27 4.24 5.57 A good planning of DRM local services in the 26 MHz band requires suitable field strength prediction techniques in order to determine the impact of the sky wave interferences as well as the coverage grade in a specific target area. As explained in Section II.C, this HF part of the spectrum has been allocated for long distance broadcasting so that well known algorithms can be used to estimate the ionospheric behavior at these frequencies. The most widely accepted ones are VOACAP and ITU-R P.533 for the regular propagation [50] and ITU-R P.534 for the sporadic propagation [51]. On the contrary, there are not predictions methods for estimating the local coverage of a DRM network working in this part of the spectrum. However, as the models typically used for calculating the space wave contribution in the VHF and UHF bands could be a solution to this problem, a study of their accuracy for the 26 MHz band was done by comparing the empirical field strength values obtained in Brasilia, Mexico Hanover and Nuremberg with simulations derived from 6 different techniques: Deygout [52], Giovaneli [53] Longley & Rice [54], Okumura-Hata [55], COST-WI [56] and the ITU-R Recommendation P.1546 [44]. Also, in order to characterize the possible influence of the surface wave propagation, the ITU-R Recommendation P.368 and the associated GRWAVE software tool [57] were evaluated. Results of this research work have been presented in [35], [58]-[60], concluding finally that the most accurate method for DRM local coverage estimation in the 26 MHz band is the one defined in the ITU-R Recommendation P.1546. BTS-10-162 11 TABLE XI STATISTICS OF THE DRM LOCAL COVERAGE PREDICTION ERROR IN THE 26 MHZ BAND Measurement Area Hanover Prediction Method Mean Error (dB) Deviation (dB) ITU-R P.1546 -2.13 4.54 Method for 26 MHz -1.73 4.09 ITU-R P.1546 0.48 6.27 Method for 26 MHz -0.04 6.55 All of this had led to obtain the planning guidelines, factors and methods for this new application but also, to conclude that single frequency networks of multiple low-power transmitters placed at high locations and transmit delay diversity techniques to prevent flat fading effects are the most appropriate solutions to maximize the coverage and minimize the sky wave interference problem in a local regulated scenario that allows a global use of these services. Brasilia ITU-R P.1546 16.33 7.93 Method for 26 MHz 5.22 7.00 ITU-R P.1546 11.62 6.02 Method for 26 MHz 2.91 6.11 ITU-R P.1546 6.22 10.18 Method for 26 MHz 1.55 6.53 REFERENCES [1] Mexico [2] Nuremberg [3] TOTAL Nevertheless, one must consider a new correction factor C∆h that accounts for the attenuation due to the terrain irregularities along the complete transmission-reception paths [54], [60] instead of the terrain clearance angle correction CTCA defined by the ITU model [44]. C ∆h = h1 − hTx 10 h1 ∆h h1 - Effective transmitting antenna height hTx - Transmitting antenna height a.g.l. ∆h - Terrain irregularity (3) [4] [5] [6] [7] [8] On other hand, the GRWAVE software tool should be used for those cases in which the surface wave is the dominant propagation mode. That is, rural locations at close distances from transmitting antennas with low heights above ground level (d ≤ 5 km and hTX ≤ λ/2) [60]. As seen in Table XI, better results were obtained for each one of the before-mentioned areas after applying this prediction procedure. However, the most relevant conclusion is that, considering the 4 regions as a whole, the mean value and standard deviation of the error decrease 4.67 dB and 3.65 dB respectively. VII. CONCLUSION The 26 MHz band is currently allocated for long-distance broadcasting services. However, the unstable ionospheric conditions in this HF frequency range make difficult to achieve a reliable reception in an 11 year solar cycle. On the contrary, there is an increasing demand of audio services covering small regions limited to a few kilometers around a transmission center so that this need can be satisfied by using the DRM technology in this part of the spectrum. In order to determine the suitability of this digital audio standard for local broadcasting in this particular band, several tests and studies have been carried out in the previous years. The main objective of this work has been to organize the corresponding partial results available up to date and provide new values following a coherent scheme. The paper analyzes and elaborates on the main outcomes of recent experiments and extends the sky wave and noise reference data. [9] [10] [11] [12] [13] [14] [15] [16] [17] European Telecommunications Standards Institute, “Digital Radio Mondiale (DRM). System specification", ETSI - ES 201 980 V3.1.1, August 2009. International Telecommunications Union - Radiocommunication Study Groups, “System for digital sound broadcasting in the broadcasting bands below 30 MHz”, ITU-R Recommendation BS.1514 -1, October 2002. International Telecommunications Union - Radiocommunication Study Groups, “Digital broadcasting at frequencies below 30 MHz MF and HF field- tests report summary final system performance report for ISTRADIATE 1999-20113 for ITU”, ITU-R Document 6E/54-E, February 2004. N. Shall and F. Held, “Mobile DRM measurements. Geneva, Mont Saleve (France) Jülich (Germany)”, DRM Document TC_SE223, Private Communication, June 2003. J. Briggs, “London 26 MHz measurements” DRM Document TC_SE360, Private Communication, May 2006. A. Waal, D. Pagel, “Local radio in the 11 m-band using DRM. Results of the field strength measurements”, DRM Document TC_SE368, Private Communication, November 2006. G. English, T. Harwood, “Observations and thoughts on26 MHz test transmissions”, DRM Document TC_209, Private Communication, November 2007. International Telecommunications Union - Radiocommunication Study Groups, “Digital Radio Mondiale DRM local coverage using the 26 MHz broadcasting Band”, ITU-R Document 6E/274E, September 2005. U. Gil, I. Peña, J. M. Matías, I. Losada, D. Guerra,, J. L. Ordiales, G. Prieto, “DRM field trials in ITU region 2: Local coverage using 26 MHz and MW simulcast”, in Proc. of the 7th Workshop Digital Broadcasting Fraunhofer IIS, Erlangen, Germany, September 2006, pp. 61-66. J. M. Matías, I. Losada, P. Angueira, U. Gil, J. L. Ordiales, A. Arrinda, “DRM (Digital Radio Mondiale) local coverage tests using the 26 MHz broadcasting band”, IEEE Transactions on Broadcasting, Vol. 53, No.1, pp. 59-68, March 2007. T. Lauterbach, B. Kreuzer, R. Zitzmann, and K. Blomeier, “Is DRM on 26 MHz an option for local digital broadcasting? Results from a field trial in Nuremberg, Germany,” in Proc. of the 5th Workshop Digital Broadcasting, Fraunhofer Institute for Integrated Circuits IIS, (Germany), September 2004, pp. 23-24. M. P. M. Hall, L. W. Barclay, “Radiowave propagation”, IEE Electromagnetic Waves Series 30, Ed. Peter Peregrinus, 1989, ISBN 0-86341-156-8. K.A. Norton,“ The propagation of radio waves over the surface of the earth and in the upper atmosphere”, Part I, Proc.IRE, Vol. 24, Oct.1936, pp. 1367-1387; Part II, Proc. IRE, Vol. 25, pp 1203-1236, Sept. 1937. K. Davies, “Ionospheric radio”, IEE Electromagnetic Waves Series 31, Ed. Peter Peregrinus, 1990, ISBN 0-86341-186-X. G. Prieto, D. Guerra, J. M. Matías, M. M. Vélez, A. Arrinda, “Digital Radio Mondiale measurement system design and measurement methodology for fixed and mobile reception” , IEEE Transactions on Instrumentation and Measurement, Vol. 57, No.3, pp. 565-570, March 2008. International Telecommunications Union - Radiocommunication Study Groups, “Planning parameters for digital sound broadcasting at frequencies below 30 MHz”, ITU-R Recommendation BS.1615, June 2003. International Telecommunications Union - Radiocommunication Study Groups, “Radio noise”, ITU-R Recommendation P.372 -10, October 2009. BTS-10-162 [18] G. Prieto et al., “External noise measurements in the medium wave band”, IEEE Transactions on Broadcasting Vol. 53, No.2, pp. 553-559, June 2007. [19] J. M. Matías, “Study of the local broadcasting in the 26 MHz band using the DRM standard (Digital Radio Mondiale)”, Ph.D. Dissertation, Dept. of Electronics and Telecommunications, University of the Basque Country - UPV/EHU, Spain, November 2008. [20] International Telecommunication Union, “Table of frequency allocations”, ITU Radio Regulations, Article 5-Section IV, 2008. [21] International Telecommunication Union, “Seasonal planning of the HF bands allocated to the broadcasting service between 5900 kHz and 26100 kHz”, ITU Radio Regulations, Article 12, 2008. [22] L. Gorada,” Handbook of antennas in wireless communications”, The Electrical Engineering and Applied Signal Processing Series, Ed. CRC Press, 2002, ISBN 0-8493-0124-6. [23] T. Lauterbach, F. Hofmann “Local radio using DRM in the 26 MHz Band - A Résumé at the end of the Nuremberg/Dillberg Field Trial”, in Proc. of the 9th Workshop Digital Broadcasting, Fraunhofer IIS, Erlangen (Germany), September 2008, pp. 69-76. [24] International Telecommunication Union, “Regional agreement for the European broadcasting area concerning the use of frequencies by broadcasting services in the VHF and UHF bands”, Final Acts of the European Broadcasting Conference in the VHF and UHF Bands (Stockholm, 1961) as Revised Regional Administrative Conference of the Members of the Union in the European Broadcasting Area, Geneva, 1985. [25] International Telecommunication Union, “Regional agreement concerning the use by the broadcasting service of frequencies in the medium frequency bands in regions 1 and 3 and in the low frequency bands in region 1”, Final Acts of the Regional Administrative LF/MF Broadcasting Conference (Regions 1 and 3), Geneva, 1975. [26] H. Arata, S. Yamazaki, “A rejection system for interference in TV reception caused by sporadic-E ionospheric propagation”, IEEE Transactions on Broadcasting, Vol. 35, No. 1, pp. 63-69, March 1989. [27] International Telecommunications Union - Radiocommunication Study Groups, “Digital Radio Mondiale DRM daytime MW tests”, ITU-R Document 6E/175E, March 2005. [28] D. Guerra et al, “Medium Wave DRM field test results in urban and rural environments” IEEE Transactions on Broadcasting, Vol. 51, No.4, pp. 431-438, December 2005. [29] U. Gil et al., ”DRM 20 kHz simulcast field trials in the medium wave band in Mexico D.F.”, IEEE Transactions on Broadcasting, Vol. 54, No.1, pp. 53-61, March 2008. [30] T. Lauterbach, “Local Radio in the 26 MHz band using DRM - Results of the Nuremberg field trial and general considerations”, Fachartikel 2005, Schriftenreihe der Georg-Simon-Ohm-Fachhochschule Nürnberg, Heft 13, S. 37, 2006. [31] I. Fernández et al., “Subjective evaluation of the reception quality thresholds for the Digital Radio Mondiale broadcast standard”, in Proc. of the 56th Annual Broadcasting Symposium - IEEE BTS, Washington D.C (USA), September 2006. [32] J. M. Matías et al., “DRM (Digital Radio Mondiale) tests results for mobile reception in 26 MHz band for local coverage”, in Proc. of the IEEE International Symposium on Broadcast Multimedia Systems and Broadcasting - BMSB, Bilbao (Spain), May 2009. [33] R.G. Neves et al., “Local digital radio in the 26 MHz band using DRM Results of the Brasilia field trials”, in Proc. of the International Microwave and Optoelectronics Conference SMBO/IEEE MTT-S, November 2007. [34] W. C. Y. Lee. “Mobile communications engineering. Theory and applications”, 2nd Edition, McGraw-Hill Professional, October 1997, ISBN 0-07-037103-2. [35] F. Maier, I. Peña, P. Angueira, A. Tissen, A. Waal, “DRM for local coverage in the 11 meter band - Predictions and measurements”, in Proc. of the 10th Workshop of Digital Broadcasting- Fraunhofer Institut for Digital Media Technology IDMT, Ilmenau (Germany), September 2009. [36] A. Giefer, O. Haffenden, J. Briggs G. Dobson , “Report on 26 MHz small scale single frequency network trials for DRM”, DRM Document TC_SE203, Private Communication, 2003 12 [37] D. Guerra et al., “Medium wave Digital Radio Mondiale (DRM) field strength time variation in different reception environments”, IEEE Transactions on Broadcasting, Vol. 52, No.4, pp. 483-491, Dec. 2006. [38] J. M. Matías et al., “Local radio coverage using the Digital Radio Mondiale international standard: Time variability characterization”, in Proc. of the IEEE International Symposium on Broadcast Multimedia Systems and Broadcasting - BMSB, Las Vegas (USA), April 2006. [39] W. C. Y. Lee, Y.S. Yeh, “On the estimation of the second-order statistics of log-normal fading in mobile radio environment”, IEEE Transactions on Communications, Vol. 22, No. 6, pp. 869 – 873, June 1974. [40] W. C. Y. Lee, “Estimate of local average power of a mobile radio signal”, IEEE Transactions on Vehicular Technology, Vol.34, No. 1, pp. 22-27, February 1985. [41] International Telecommunications Union - Radiocommunication Study Groups, “Generalized method for separating the long and short term variation components of a received signal. Application to ground-wave propagation in the MF band”, ITU-R Documents 3J-26E and 3K-22E, May 2008. [42] D. de la Vega et al., “Generalization of the Lee method for the analysis of the signal variability”, IEEE Transactions on Vehicular Technology, Vol. 58, No. 2, pp. 506-516, February 2009. [43] J. M. Matías et al., “Location correction factor for coverage planning tools for DRM (Digital Radio Mondiale) in the 26 MHz band” in Proc. of the IEEE International Symposium on Broadcast Multimedia Systems and Broadcasting - BMSB, Las Vegas (USA), April 2008. [44] International Telecommunications Union - Radiocommunication Study Groups, “Method for point-to-area predictions for terrestrial services in the frequency range 30 MHz to 3000 MHz”, ITU-R Recommendation P.1546-4, October 2009. [45] Y. Okumura, E. Ohmori, T. Kawano, K. Fukuda, “Field strength and its variability in VHF and UHF land-mobile radio service”, Review of the Electrical Communication Laboratory, Vol. 16, No. 9-10, pp. 825-873, October 1968. [46] F. Hofmann, C. Hansen, W. Schäfer, “Digital Radio Mondiale (DRM) digital sound broadcasting in the AM bands”, IEEE Transactions on Broadcasting, Vol. 49,No. 3, pp 319-328, September 2003. [47] T. Lauterbach, M. Unbehaun “Mobile reception of digital AM audio and data services using long-range skywave propagation” IEEE Transactions on Broadcasting, Vol. 53,No. 3, pp. 597-608, September 2007. [48] I. Peña et al., “Study of the ionospheric interference for planning DRM local services using the 26 MHz band”, in Proc. of the 11th International Conference on Ionospheric Radio Systems and Techniques - IET/ IRST, Edinburgh (UK), April 2009. [49] I. Peña et al., “Sky wave interfering signals evaluation for digital local broadcasting services in the 11 meter band”, in Proc. of the 4th European Conference on Antennas and Propagation - EuCAP , Barcelona (Spain), April 2010. [50] International Telecommunications Union - Radiocommunication Study Groups, “Prediction methods for adaptive HF system planning and operation”, ITU-R Recommendation F.1611, February 2003. [51] International Telecommunications Union - Radiocommunication Study Groups, “Method for calculating sporadic E field strength”, ITU-R Recommendation P.534-4, October 1999. [52] J. Deygout, “Multiple knife-edge diffraction of microwaves”, IEEE Transactions on Antennas and Propagation, Vol. 14, No. 4, pp. 480-489, July 1966. [53] C. L. Giovaneli, “An analysis of simplified solutions for multiple knifeedge diffraction”, IEEE Transactions on Antennas and Propagation, Vol. 32, No. 3, pp. 297-301, March 1984. [54] A. G. Longley, P. L. Rice, “Prediction of tropospheric radio transmission loss over irregular terrain”, Institute for Telecommunication Sciences, ESSA Technical Report ERL 79-ITS 67, July 1968. [55] M. Hata, “Empirical formula for the propagation loss in land mobile radio services”, IEEE Transaction on Vehicular Technology, Vol. 29, No. 3, pp. 317-325, August 1980. [56] European Cooperation in the Field of Scientific and Technical Research EURO-COST 231, “Urban transmission loss models for mobile radio in the 900 and 1800 MHz bands”, COST 231 TD(90) 119 Rev. 2, September 1991. BTS-10-162 [57] International Telecommunications Union - Radiocommunication Study Groups, “Ground wave propagation curves for frequencies between 10 kHz and 30 MHz”, ITU-R Recommendation P.368-9, February 2007. [58] I. Peña et al., “Digital Radio Mondiale field trials in Brasilia for local radio coverage using the 26 MHz band”, in Proc. of the IEEE International Symposium on Broadcast Multimedia Systems and Broadcasting - BMSB, Orlando (USA), March 2007. [59] I. Peña et al., “Analysis of the ITU-R P.1546-2 prediction method accuracy for DRM local coverage using the 26 MHz band”, in Proc. of the 57th Annual Broadcasting Symposium - IEEE BTS, Washington D.C. (USA), November 2007. [60] I. Peña et al., “Ground wave propagation analysis for DRM local coverage in the 26 MHz band”, in Proc. of the 4th European Conference on Antennas and Propagation - EuCAP , Barcelona (Spain), April 2010. View publication stats 13