EMC 2008 19 th INTERNATIONAL WROCLAW SYMPOSIUM AND EXHIBITION ON ELECTROMAGNETIC COMPATIBILITY, WROCLAW, 11 13 JUNE, 2008 www.emc.wroc.pl OVERVIEW OF SEAMCAT USAGE AND ITS OFDMA EVOLUTION Jean-Philippe Kermoal 1, Marc Le Dévendec 1, Christian Petersen 2 and Arturas Medeisis 3 1 European Radiocommunications Office (ERO), Denmark, kermoal@ero.dk, ledevendec@ero.dk 2 ipeople Aps, Denmark, Christian.petersen@ipeople.dk 3 Vilnius Gediminas Technical University, Lithuania, arturas.medeisis@el.vgtu.lt Abstract: This paper presents an overview of the usage of the tool SEAMCAT for the sharing and compatibility studies performed within the CEPT using this tool. It also presents the latest development of the tool with respect to the OFDMA module. Keywords: CEPT, Monte Carlo modelling, OFDMA, SEAMCAT, spectrum engineering. 1. INTRODUCTION SEAMCAT (Spectrum Engineering Advanced Monte Carlo Analysis Tool) is a software tool that permits sharing and compatibility studies between radio systems in the same or adjacent frequency bands. Over the years, it has reached a certain maturity and its evolution has been presented on a regularly basis at the Wroclaw Symposium [1][2][5]. This tool has been used extensively within the CEPT (European Conference of Postal and Telecommunication Administrations) for answering issues on sharing and compatibility analysis. In order to consider all technology the SEAMCAT tool is kept up to date with regular new patch releases and new radio interface modules. SEAMCAT is freely available for downloads and live updates through its web interface [3]. The software is Java based and the source code of the official release is available upon request [4]. This paper presents an overview of the SEAMCAT software and its applicability in Section 2. Section 3 addresses several most recent examples of practical studies which were undergone by the CEPT. Section 4 presents the introduction of the new OFDMA component which allows considering mobile services using this new access technique. The simulation of the OFDMA systems is conceptually similar to that of CDMA systems [2]. This is followed by some concluding remarks 2. OVERVIEW OF SEAMCAT 2.1. What is SEAMCAT? SEAMCAT is a software tool based on the Monte- Carlo simulation method which permits statistical modelling of different radio interference situations. It is well suited to deal with a diversity of complex spectrum engineering and radio compatibility problems. It is developed within the frame of CEPT since 1997. A Memorandum of Understanding between the industry and the CEPT member states was signed in collaboration with the European Radiocommunications Office (ERO) to define its specifications and its management through the SEAMCAT Technical Group (STG) which is opened to national regulatory authority, industry and academia. SEAMCAT has been adopted by the CEPT since 1997 [6] and the methodology was endorsed by the ITU-R in 2000 [7]. Traditional compatibility analysis methods normally are based on the worst case examination and performed using the minimum coupling loss (MCL) method [13]. Such methods may lead to unnecessary stringent measures to ensure the compatibility between radio systems and may results in a sub-optimal use of the spectrum. SEAMCAT, on the contrary, allows establishing and analysing compatibility scenarios that reflect real compatibility situation. It permits the quantification of the interference levels between different radio systems in terms of a probability that the reception capability of the receiver under consideration is impaired by the presence of interferers for traditional system and capacity loss for CDMA systems. This software is developed as a generic compatibility analysis tool which is neither system-specific nor service-specific and therefore can address any interference scenario regardless of the type of victim and interfering radio systems. 2.2. Applicability SEAMCAT allows flexible and easy definition of various elements of interference scenario, such as interfering and victim system parameters, propagation conditions, frequency, spatial and temporal distribution of users. Its plug-in capability allows further flexibility for the user to define additional computations allowing adaptive systems to re-adjust some of their operating parameters in order to fit the environmental conditions in a dynamic fashion.
This tool is particularly suited for compatibility studies for systems that operate in shared or adjacent frequency bands (i.e. for identifying sharing issues to be investigated and for determining appropriate frequency separations or other frequency arrangement). Furthermore, SEAMCAT can evaluate radio equipment parameters (i.e. transmitter emission masks, receiver susceptibility/sensitivity and density of interfering transmitters). All interference mechanisms may be taken into account (unwanted emission, blocking, inter-modulation) which permits an evaluation of limits of parameters such as spurious emission, blocking level and inter-modulation level. 3. CEPT COMPATIBILITY STUDIES USING SEAMCAT The SEAMCAT usage in the CEPT is extensive and some recent practical studies are briefly presented in this section. The SEAMCAT scenario files (.sws) for each of the studies presented in this paper can be found at the website [8] beside the downloadable report. 3.1. GSM on board aircraft vs terrestrial networks SEAMCAT was used to evaluate the interference impact of GSM onboard (GSMOB) aircraft into the terrestrial mobile networks (GSM900, GSM1800, UMTS900, UMTS 1800, UMTS in the 2 GHz FDD core-band and CDMA-450/FLASH-OFDM) in the 1800 MHz band (1710-1785 MHz UL (uplink) / 1805-1880 MHz for DL (downlink)) [9]. For this study, the considered GSMOB system (operating at a height of at least 3000 m above ground level) is designed to ensure that ac-ms (on board aircraft - mobile station) are unable to attempt to communicate with terrestrial networks, whilst providing onboard connectivity to ac-ms in the GSM1800 frequency band. The GSMOB system consisted of a NCU (network control unit) so that signals transmitted by terrestrial BTS were not visible within the cabin, and an ac-bts that provided connectivity. SEAMCAT investigated the onboard NCU/ac-BTS link impact on the UL and DL terrestrial communication for single or multiple aircraft scenarios. The random distribution of aircraft around a terrestrial station was taken into account in the scenarios as shown in Figure 1. SEAMCAT simulations were used to illustrate the typical influence in a victim cell. For GSM, the influence of the interference is quantified by the signalto-noise criteria C/(N+I), i.e. the probability that the C/(N+I) is below a limit given by the performance specification. The victim system cell radius was calculated by using noise limited network option of SEAMCAT, using the availability target of 95%, which is typical for similar studies. Only voice-service simulation is possible in SEAMCAT, therefore the C/(N+I) was set to 9 db. Figure 1: GSMOB interfering on the terrestrial victim UL (g-ms to g-bts) from multiple aircraft (Source [9]) The studies have shown that there is no significant increase of the level of interference due to GSMOB emissions from multiple aircraft because the dominant source of interference to a terminal on the ground is the GSMOB in closest aircraft and provided that sufficient spectrum is available, the ac-btss in different aircraft can operate on different frequencies. 3.2. UMTS 900/1800 vs systems in adjacent bands In [10], SEAMCAT was used to evaluate some of the co-existence scenarios, like the one dealing with interference between UMTS900 and GSM-R. GSM-R (GSM Railway) networks offer a linear coverage of railway lines with bi-sector radio sites installed along the railway and with two major characteristics i.e. linear coverage and high quality coverage (95% space and time availability). Figure 2 presents the GSM-R MS which are uniformly distributed on the railway line (GSM-R DL channel without PC and UL with PC). The simulations on the interference from UMTS900 DL to GSM-R DL reception of train mounted MS showed a probability of C/I=9 db less than 0.02%, and a probability of C/I=12 db less than 0.1%, which is below the required 0.5% for that specific study. Additional studies were conducted in order to assess the worst situations corresponding to the GSM-R devices at the cell-edge by varying the BTS antenna height, the GSM- R cell range and the distance between the GSM-R BTS and the GSM-R MS which is randomly drawn for the different snapshots on a railway line. The interference analysis impact of DECT (Digital Enhanced Cordless Telecommunications) on UMTS FDD 1800MHz DL and vice versa was also performed in [10]. DECT uses Dynamic Channel Allocation (DCA) to combat interference. When interference occurs on one channel, DECT has the ability to select another one, without loss of communication. To illustrate DCA in SEAMCAT, the interfering frequency was set to an equal probability distribution of frequency
between 1880 MHz and 1900 MHz. When assessing the impact of DECT on UMTS CDMA DL, the results were expressed in term of capacity losses and number of dropped users. For the impact on DECT MSs, the results were expressed in term of probability of interference. d0 2 x 104 1.5 1 0.5 1 2 3 10 11 12 13 22 23 24 25 26 27 37 38 39 40 49 50 51 analysis focused on the impact of the DVB-T system using UHF Channel 21 (470-478 MHz) on PMR/PAMR systems operating in the frequencies below 470 MHz, and vice versa. SEAMCAT was used to assess the interference of DVB-T into a PMR/PAMR MS receiver and the interference of PMR/PAMR BS (Base Station) into a DVB-T receiver (see Figure 3). One DVB-T station was assumed to be transmitting at 73 dbm (20 kw) e.i.r.p. in channel 21 (8 MHz centred at 474 MHz). Multiple PMR/PAMR BSs are deployed within the whole DVB- T covered area, using the 5 highest channels, which are the closest to DVB-T channel 21. 4 14 15 28 41 42 52 0 5 6 16 29 30 43 53 54-0.5 7 17 18 31 44 45 55-1 8 9 19 32 33 46 56 57 20 21 34 47 48-1.5 35 36-2 -2-1.5-1 -0.5 0 0.5 1 1.5 2 x 10 4 Figure 2: GSM-R MS are uniformly distributed on the railway line (GSM-R DL channel without PC and UL with PC) (Source [10]). 3.3. TEDS vs PMR/PAMR in the 400 MHz band Another use of SEAMCAT was considered for the evaluation of the impact of TEDS (TETRA enhanced data services) on existing PMR (Private Mobile Radio) systems in the 380-470 MHz frequency band and vice versa [11]. Analogue FM, TETRA, TETRAPOL and CDMA-PAMR (Public Access Mobile Radio) operating in 2-MHz band in this frequency range were considered in this study. SEAMCAT was used to define either the interference probability (for non-cdma system) or the capacity loss for CDMA system, when CDMA systems were victim of interference from a TEDS radio link. As an example of results of this study, it was concluded that for realistic PMR systems interferer density and for non-cdma system, SEAMCAT shows interference probability less than 1.5%. For CDMA as interferer system and realistic interferer density, result shows an interference probability smaller than 3%. Additional compatibility studies between TETRA and 25/50 khz-channel TEDS were performed considering interleaved frequency allocation. The statistical simulations show similar results irrespective of the interleaving scheme. 3.4. PMR/PAMR operating in the range 450-470 MHz vs DVB-T system in UHF TV channel 21 In [12] adjacent band compatibility studies between PMR/PAMR systems operating at the top end of the 450-470 MHz band and DVB-T (Digital Video Broadcasting-Terrestrial) system operating in the band 470-862 MHz were investigated. In particular, the Figure 3: PMR/ PAMR BSs (interferers) inside the coverage area of a DVB-T transmitter and near DVB-T receivers (victim) (Source [12]) The study indicates a significant probability of interference in both directions. In cases where the PMR/PAMR MS is the victim, the probability of interference is significant when the PMR/PAMR MS are located in the vicinity (around 1 km) of a highpower DVB-T transmitter. The problem increases when the victim PMR/PAMR MS is at the cell edge of the PMR/PAMR service area. In the case of the DVB-T receivers are victim, the probability of interference is significant when the DVB-T receivers are close to the PMR/PAMR BS and it becomes worse if the victim DVB-T receiver is close to the edge of the DVB-T coverage area. 3.5. On going activities SEAMCAT was used within CEPT in order to assess the impact of a revised regulation for RFID (Radio Frequency Identification systems) in the band 865 868 MHz and its impact on other SRDs (Short Range Devices) operating in the same sub-band or in neighbouring bands. The revised corresponding ECC Report is under public consulation [14]. SEAMCAT is also currently used in order to investigate the possibility to identify spectrum at 1.5 GHz for Professional Wireless Microphone Systems (PWMS) and in particular to assess their impact on the existing services/systems operating in this frequency range (Broadcasting, Satellite, Fixed Service ).
Another current example of its usage is within the studies of compatibility between the GSM on board vessels (GSMOBV) and terrestrial GSM/UMTS- 900/1800 MHz systems. Both non-cdma and CDMA simulation modes are used in those studies. 4. OFDMA EVOLUTION 4.1. Justification for evolution SEAMCAT is kept up to date with the emerging radio system development so that it remains a state of the art reference in coexistence and compatibility analysis tool in Europe. Up to now, OFDM technologies (i.e. public broadcasting, LANs and WANs) have been largely used for Fixed and Broadcasting applications, working in known fixed locations. Therefore it was possible to assign sub-carriers on a static basis that would minimise the interference. Today s trend is to have the basic OFDM technology extended to mobile environment and a MS which might have varying number of subcarriers used in a dynamic way. It was then, decided to develop a full-scale OFDMA simulating module in SEAMCAT. Its implementation in SEAMCAT started in spring 2007 with its first Beta version 3.2. 4.2. OFDMA algorithm The simulation of OFDMA systems is conceptually similar to that of CDMA systems [2], except that after the overall two-tiers cellular system structure (incl. wrap-around) was built and populated with MSs, then the OFDMA MSs is assigned a specific number of traffic sub-carriers. The number of sub-carriers assigned to a specific MS is user definable through a standard SEAMCAT distribution parameter, but the number of sub-carriers assigned to a given mobile is fixed within the snapshot. The CDMA power tuning process is replaced with an OFDMA equivalent process in which the system will try to scale power so that all connected mobiles achieve a minimum bit rate found in a lookup table based on the number of assigned subcarrier. A MS which is unable to achieve this minimum bit rate after 3 consecutive power balance iterations is dropped out from the system. Upon dropping a MS, the system will try to connect any previously dropped MSs using the now available subcarriers. The lookup tables used to define the minimum bit rate for a given amount of subcarriers is OFDMA technology dependent and can be specified by the user when defining the scenario. When the power balance is complete the system will calculate overall carried traffic per BS. Similarly to the CDMA module, the following interference modelling can easily be performed for either UL or DL: a) OFDMA vs. non-ofdma: OFDMA as interferer; b) OFDMA vs. non-ofdma: OFDMA as victim; c) OFDMA vs. OFDMA Figure 4 presents the generic high level OFDMA algorithm being implemented in SEAMCAT. It is possible to see that the multiple scenario modes can be implemented and the OFDMA module can interact with all other SEAMCAT modules which have been developed so far either for mobile or non-mobile system. Figure 4: OFDMA algorithm in SEAMCAT. This algorithm is the working version of algorithm specifications applied in production of early Beta version (September 2007). The algorithm developed at the time of publication of this paper could only consider single frequency assignment. This means that only one frequency channel is used by all BS in an OFDMA system. This algorithm is still in its Beta version and it might be further updated as a result of testing and calibration process. The further step will address the introduction of MFN (Multi Frequency Network) option for OFDMA system, where different base stations forming OFDMA cellular structure are assigned
different frequencies, thus reducing the internal interference between the cells of the same system. Figure 5 shows an example of the current OFDMA interface for setting the input parameter of the OFDMA system to simulate. It uses some of the user interface developed for CDMA. Since the OFDMA implementation is based the CDMA implementation it will be possible to use the highly detailed visual point and click view of the last simulated snapshot to inspect the values of the system. [3] SEAMCAT project description and software download, ERO homepage [On-line]. Available: http://www.ero.dk/seamcat [4] Source code available upon request to kermoal@ero.dk [5] A. Medeisis, Monte-Carlo modeling of co-existence of radio systems using smart antennas or dynamic frequency selection, Proc. EMC-2006 International Wroclaw Symposium and Exhibition on EMC, Wroclaw 2006. [6] CEPT/ERC/Report 68, Monte Carlo radio simulation methodology for the use in sharing and compatibility studies between different radio services or systems, Regensburg, May 2001 [7] ITU-R Report SM.2028, Monte Carlo simulation methodology for the use in sharing and compatibility studies between different radio services. [8] http://www.erodocdb.dk/doks/doccategoryecc.aspx?doccatid=4 [9] CEPT/ECC/Report 93, Compatibility between GSM equipment on board aircraft and terrestrial networks, Lübeck, September 2006 [10] CEPT/ECC/Report 96, Compatibility between UMTS 900/1800 and systems operating in adjacent bands, Krakow, March 2007 [11] CEPT/ECC/Report 99, TETRA enhanced data services (TEDS): Compatibility studies with existing PMR/PAMR and air ground air (AGA) systems in the 400 MHz band, Bern, February 2007, Budapest, September 2007 [12] CEPT/ECC/Report 104, Compatibility between mobile radio systems operating in the range 450-470 MHz and digital video broadcasting-terrestrial (DVB-T) system in UHF TV channel 21 (470-478 MHz), Amstelveen, June 2007 [13] CEPT/ERC/Report 101, A comparison of the minimum coupling loss method, enhanced minimum coupling loss method and the Monte-Carlo simulation, Menton, May 1999 [14] Draft Revision of ECC Report 37, Compatibility of planned SRD applications with currently existing radiocomunication applications in the frequency band 863-870 MHz (www.ero.dk) Figure 5: OFDMA scenario interface. 5. CONCLUSION This paper presented a brief overview of the different compatibility studies that were performed recently within the CEPT using SEAMCAT to allow an efficient usage of the spectrum. Through the constant development over the recent years (i.e. implementation of CDMA and OFDMA modules), SEAMCAT has remained a state of the art tool to permit optimal statistical modelling of different radio interference scenarios for performing sharing and compatibility studies between radio systems in the same or adjacent frequency bands. ACKNOWLEDGMENT The authors would like to thank colleagues in SEAMCAT Technical Group, whose work and contributions formed the conceptual basis for developing this article. REFERENCES [1] M. Le Devendec and A. Refik, The spectrum engineering advanced Monte Carlo analysis tool (SEAMCAT), Proc. EMC- 2002, International Wroclaw Symposium and Exhibition on EMC, Wroclaw 2002. [2] A. Medeisis, Adding CDMA capabilities to the SEAMCAT spectrum engineering analysis tool, Proc. EMC-2004 International Wroclaw Symposium and Exhibition on EMC, Wroclaw 2004. Jean-Philippe Kermoal managed the SEAMCAT project at ERO. He received his PhD from Aalborg University in 2002. During his PhD work, he was involved in the EU project METRA. In 2002, he joined Nokia Research Centre in Finland. He has been involved in 3GPP standardisation work and he has worked on spectrum related topics for 4G. He was involved in the EU project WINNER II where he was leading the "technologies for spectrum usage" team. Since August 2007, he joined the European Radiocommunications Office as spectrum engineer. Christian Petersen is the principal SEAMCAT developer since 2004. He works as an independent consultant specializing in advanced java development and agile development methods. He is currently involved in projects ranging from java based smartcards to enterprise financial systems. Marc Le Dévendec received the Engineering Degree in Telecommunication from the Ecole Nationale Supérieure des Télécommunications de Bretagne (France) in 1998. From 1998 to end of 2003, he has been working with the Agence Nationale des Fréquences, the French regulator dealing with frequency management. Since January 2004, he joined ERO as spectrum engineer and is participating in the activities of various CEPT and ITU forums dealing with radio systems coexistence. He is following the development of SEAMCAT since the end of 1998. Arturas Medeisis worked on SEAMCAT project from 2002 to 2007, during his tenure at ERO. Since October 2007, he is Docent of Telecommunications Engineering Department at the Vilnius Gediminas Technical University, Lithuania.