Satcoms 2020 R&D challenges: Part II: mobile communications

SATCOMS 2020 R&D CHALLENGES: PART II: MOBILE COMMUNICATIONS P. Angeletti, A. Bolea Alamanac, F. Coromina, F. Deborgies, R. De Gaudenzi, A. Ginesi European Space Agency European Space Research and Technology Centre (ESTEC) Keplerlaan 1, 2200 AG Noordwijk, The Netherlands {e-mail: rdegaude@xrsun0.estec.esa.nl, Phone : +31-71-5654227} Keywords: Satellite Communication Systems, Mobile Satellite Services, Satellite Communication Payloads, OnBoard Processors, Mobile User Terminals. Abstract In this paper we intend to complement the information contained in the companion paper [1] dealing with broadband fixed communications with possible scenarios of development for future mobile satcom networks, including the underlying key challenges as well as critical techniques and technologies to be developed in the coming years. 1 Introduction Mobile terrestrial networks have seen an enormous growth in the last years and three generation of commercial networks have been deployed while the fourth one is being defined. Together with the expansion of service provision (from pure voice and small messages to full Internet access and multimedia capabilities), terrestrial networks, coverage, level of interoperability and Q.o.S. made giant steps. The dream of global satellite mobile networks based on large satellite constellation became a reality in the late ‘90s with a very modest commercial success. Nonetheless today, despite some important failure, global and regional satellite mobile networks are still representing an important commercial reality and generating good revenues in regions or applications where the terrestrial offering is weak. The last generation Inmarsat BGAN network demonstrates that satellite networks are also moving from 2G to 3G type of services at a much more affordable cost than in the past thanks to the innovation which took place both at the space and at the ground segment levels. Furthermore, another opportunity for the satellite appears in the rollout of hybrid satellite/terrestrial mobile networks in USA (ICO, Skyterra, Terrestar) and Europe (Solaris W2A, Inmarsat…) providing multimedia broadcasting as well as interactive (emergency) services. But, as for broadband satellite networks, the past success should not relax the need for a continuous yet reinforced innovation effort to guarantee a comparable satellite role in the medium/long term. This is particularly true when we look at the impressive pace at which terrestrial wireless technologies are developing. Furthermore, due to the extreme competition existing in this field, service quality is constantly growing at the same time as the access cost is decreases. The satellite component successful development in 3.5/4G type of networks requires extra effort to be dedicated in the following high-level areas: • Reduction of the performance gap between terrestrial and satellite networks; • Terminal cost, power consumption and form factor; • Service fee reduction; • Level of satellite network integration in terrestrial network; • Spectrum availability. Satellite is extremely well positioned for the provision of rich multimedia content to a large number of mobile users satellite, in particular for broadcasting/multicasting type of services. For mobile broadcasting applications terrestrial networks are intrinsically penalised by the large infrastructure cost to achieve the required coverage and for which hybrid solutions may have better chances of success. But interesting perspectives are also emerging for the interactive type of services, in particular for collecting information from vehicles (e.g. road toll, location-based automated services etc…). In line with [2], the identified key technical challenges for the satellite mobile services are: • Service enhancements (volume, peak rate, cost); • Integration with 3.5G/4G terrestrial networks; • Commercial exploitation of ATC/CGC networks; • Support of enhanced interactive services. broadcast/multi-cast and 2 Proposed Way Forward For each of the identified challenges the following key techniques/technologies have been identified to play a pivotal role in the mobile satellite networks evolution: 1. Smaller satellite beam width for higher antenna gain and frequency reuse; 2. Extended spectrum and frequency reuse; 3. Integrated RF front-ends supporting larger number of beams; 4. Enhanced on-board equipments (reduction in cost, size, mass..) 5. Enhanced on-board processors; 6. Enhanced return link capabilities; 7. Enhanced waveform and on-ground signal processing; 8. System/payload design tools; 9. Low-cost enhanced mobile terminals; 10. Enhanced terminal capabilities. The following table shows the mapping of the selected technologies onto the application scenarios and the preliminary assessment of the potential benefit of each technique/technology. Small satellite antenna beams/high accuracy pointing – System level de-pointing countermeasures performance. We have shown there that under the assumption of uniform traffic distribution, the system throughput of a point-to-point system is almost linearly growing with the number of beams covering the service area. The commercial GEO MSS satellites under constructions, like MSV-2, are already using a 22 m diameter reflector, while 12-18 m reflectors are currently flying. To accommodate the future increased performance requirements, larger antenna reflector will be required. The antenna pointing accuracy, as well the reflector surface stability, may end-up being the limiting factor. To make this technology viable, large reflector (20++ m) technology shall offer adequate thermo-elastic distortion performance combined with compensation of residual pointing errors by means of digital Beam Forming Networks (BFN), Adaptive Coding and Modulation (ACM), networkaided beam handover. The increase of reflector size/number of beams leads to large increase of number of feeds and associated input/output section hardware causing major satellite accommodation issues. The access to this key space segment technology currently dominated by US companies is subject to ITAR restrictions and the current European technology gap constitute one of the major weakness for European industry competitiveness in this sector that shall be urgently addressed. As shown in the companion paper [1] for point-to-point applications, the satellite beam size is a key to achieve higher Techno/ Application scenario Small Beams / Accurate Pointing Low-cost Enhanced Terminals Extend bandwidth and frequency reuse Advanced Payload Concepts Enhanced On-Board (or-On-Ground) Processors System/Payload Design Tools Enhanced Waveforms and OnGround Processing Enhanced RL Capabilities Interactive Service Performance Enhancement +++ +++ +++ +++ +++ ++ ++ +++ Integration with 3.5/4G Networks +++ ++ +++ +++ +++ ++ +++ +++ Enhanced Broadcast/Multicast Services ++ +++ +++ ++ ++ + ++ ++ Table 1: Technology-to-Application Mapping Enhanced Terminal Antenna The mobile terminal antenna represents a major bottleneck for the mobile satellite networks performance as its omnidirectional antenna pattern requirement, together with the compact front-end and adjacent band filtering requirements determines a low G/T figure. This is particularly true for the hand-held type of terminals. New challenges appear with the need to support antenna (polarisation) diversity which is likely to be adopted by enhanced terrestrial terminals exploiting MIMO techniques. In this respect, the use of dual linear or dual circular polarisation becomes a new requirement for satellite (hybrid) terminals. Ideally, the satellite terminal should be able to support the same type of dual polarisation technique adopted by terrestrial systems. For vehicular terminals the satellite antenna integration is easier than for hand-held since the vehicle antenna assembly already contains a GPS antenna and space is not as constrained as in the hand-held case. But again economical and aesthetical antenna implementation poses important technical challenges. Extended Spectrum and Frequency Reuse Current and planned MSS systems exploit the MSS frequency allocation in the L and S bands. These frequencies present excellent propagation properties for mobile terminals particularly when they are equipped with low or omnidirectional antennas. However, the MSS allocation on these bands is relatively scarce: the current L-band allocation is saturated; S-band is limited to 30 MHz which will rapidly saturate in particular for the expected provision of mobile broadcast/interactive applications which are particularly bandwidth hungry. To illustrate this point, Figure 1 shows the EIRP capabilities of current and next generation payloads. The current generation payloads designed for the provision of broadcast/interactive services can address up to 15 MHz of capacity (approximately 8Mbps) when handheld terminals are considered (black star in Figure 1). Next generation payload will be able to address up to 65 MHz under the same conditions (grey star in Figure 1). However, for linguistic beams systems, assuming 15 MHz allocated per S-band system and a frequency reuse factor of 3 (thanks to the use of the cross polarization for an improved isolation) only 45 MHz of capacity will be available for each system. This scarcity of spectrum will be particularly penalising for systems addressing vehicular terminals since next generation payloads are capable of addressing a much higher capacity for those terminals (white star in Figure 1). More frequency reuse can be achieved by increasing the number of satellite beams as well by using dual polarisation (see later). However, even if a spot-beam coverage can increase the frequency re-use, that kind of coverage is generally more adequate for the provision of two-way unicast services. Those services are far more data rate hungry than the broadcast/interactive based services. A similar analysis has been done for systems dedicated to the provision of twoway unicast mobile services and, although the overall capacity handled by the system is superior, the data rate requirements are such that similar conclusions in terms of spectrum scarcity can be derived. Additionally, narrow spot beam coverages in those frequency bands require very challenging antenna technologies. Payload EIRP capabilities 73.0 Dual Polar Power Limited Area EIRP/MHz (dB) 68.0 63.0 58.0 Spectrum Limited Area 53.0 48.0 195 185 175 165 155 145 135 125 115 105 95 85 75 65 55 45 35 25 15 5 43.0 Capacity (MHz) Current Generation Payload Next Generation Payload Required EIRP/MHz Handheld Rx Required EIRP/MHz Vehicular Rx Figure 1: Current and Next Generation Payload EIRP capabilities Therefore, there is a definite need for new MSS frequency bands if we want to expand the offer of services keeping the pace of terrestrial mobile networks. To cope with this spectrum scarcity problem which may severely constrain the satellite mobile sector development a stepped approach may be envisaged. In Step 1, an increase of S-band spectrum efficiency may be achieved by exploiting dual polarisation reuse within each satellite beam. This will require the development of a new space segment and related user terminals as well as the adoption of an evolved waveform capable to operate in dual polar mode. However, linguistic beam systems, dual polarization will likely require lowering the frequency reuse factor down to 2 to assure enough isolation between co-frequency beams; therefore systems will still be limited to a maximum capacity of 60 MHz. This could be enough for handheld oriented systems but it is still very constraining when only vehicular terminals are considered. Step 2 consists in extending the S-band CGC frequency allocation beyond the current 30 MHz. This is likely to require a major lobbying effort with possible opposition from terrestrial operators not eager to see more spectrum allocated to MSS in S-band. Step 3 consists in requesting a frequency allocation at C-band to support higher data rates for vehicular and nomadic terminals, or terminals on-board large platforms etc… A CGC approach similar to S-band in support of the development of hybrid networks may be envisaged in C-band. This step is also very challenging because the lower part of Cband is also currently used for FSS services and terrestrial operators are already asking for allocations to deploy wireless terrestrial services on these bands. Nevertheless, in accordance with Resolution 231 of WRC-07, the Agenda Item 1.25 is currently considering possible additional allocations to the mobile-satellite service particularly in the frequency bands 4-16 GHz to rectify the predicted shortfall of the MSS allocation. Advanced Mobile Payload Concepts The support of larger antenna reflector and higher number of beams (> 300) calls for a reduction of cable losses through closer RF front-end integration, enhanced RF power stage efficiency, more powerful scalable processors, digital BFNbased antenna errors compensation. The reduction of payload equipments mass, size and power requires the adoption of new RF and optical technologies as well the exploitation of more advanced deep sub-micron (DSM) digital technologies. Figure 2: 20 W GaN MMIC (8.5 - 11 GHz) Courtesy of IAF To stimulate and accelerate the market introduction of this innovative (disruptive) technology in Europe ESA has assembled a consortium of competent partners under the: GaN Reliability Enhancement and Technology Transfer Initiative (GREAT2). Primary objective of GREAT2 is to establish a commercial, space compatible, GaN MMIC foundry process through UMS by 2011. New technologies for on-board RF equipment Current mobile communication satellites make extensive use of solid-state power amplifiers (SSPA) and rely exclusively on gallium arsenide (GaAs) as the preferred semiconductor technology. However, a new semiconductor technology is emerging, namely gallium nitride (GaN see Figure 2) which brings significant advantages to the design of SSPAs. As a matter of fact, a GaN based SSPA has a number of attractive new features as compared to current GaAs implementation: • Higher base plate temperature operation, associated thermal mass reduction; with • Improved linearity and wider band operation; • Higher efficiency and reduced RF power combining losses; • Possibility of SSPA vertical mounting with significant improvement on output section accommodation and reduction of output losses; • Possibility of matrix less implementation: lower losses, increased numbers of SSPA/feeds. This is an aspect with a good potential to enhance hot-spot capabilities; • More compact design (both chip and PA) • Increased robustness. As mobile communication satellites are generating a large number of beams, it is essential that the numerous RF equipments such as LNAs and converters are both compact and cost effective. The key to this is generic design for those equipments which become stand-alone products that are highly integrated as well as easily customisable through either embedded flexibility or at least modular approach. This is not only valid for the RF chain but also for the TM/TC interface and the power conditioning unit. The active elements of the RF chain can easily be integrated into a chipset of microwave monolithic integrated circuits (MMIC). Although this integration requires a significant investment as compared to a COTS based solution, it does provide the most effective solution for a large number of units. However, the filters tend to be difficult to integrate and may become the bottleneck of a further size reduction of the RF chains. Luckily, innovative design ideas as well as new technologies (e.g. lossy filters and micro-machining respectively) are emerging and are expected to lead to more compact LNAs and converters. A number of ESA contracts are currently supporting those new developments which are expected to appear in enhanced onboard equipments in the coming years. As the number of equipment gets larger, the RF harness required to interconnect all the units becomes both tricky and “massive”. Fibre optics is expected to be a quite effective solution to this issue. As a matter of fact, the fibre has many advantages over an RF cable, namely its small weight and size, small bend radius, very low transmission loss and total immunity to EMC. Indeed this approach implies the qualification of new components such as lasers, photodiodes, optical connectors which currently have little or no space heritage at all. But this is not considered as a showstopper as exemplified by the ESA Earth Observation SMOS Mission which is relying to a large optical harness. A number of ESA contracts addressing the different building blocks as well as the whole concept of microwave photonics distribution of RF signals are currently running. Similarly, fibre optics is envisaged for on-board high speed digital interconnect which has been identified as a severe limitation in the implementation of high capacity processors. Terrestrial experiments have already demonstrated that the performance is there with terabit rates… so the main issue is again the qualification of suitable building blocks. Enhanced on-board Processors Today mobile payloads are including transparent digital processors which allow performing flexible channelization, frequency reuse plan, beam forming as well as automatic level control. The increased processed bandwidth and the number of satellite beams require an evolution of the digital processor to enhance its capabilities with lower mass and power consumption [3], [4] (see Figure 3 and Figure 4). Key transparent OBP functionality is digital beamforming which allows single reflector AFR configurations and proves flexibility in: • Satellite coverage • Beam shape (e.g. sidelobe control), pointing • Beam/frequency allocation Advanced processor functions can include: • Payload/antenna self-calibration • Interference reduction by mean of “dynamic” or “adaptive” beamforming. Key driving technologies to support this processor enhancement are: • Technology update to latest DSM (0.65 μm under development on an ESA TRP contract [5]). • Modular and scalable architectures (reduced NRE, increased throughput/beams/feeds) – this solution is now available in Europe. • Exploitation of available ASIC resources (reduced # parts). • Increase of integration which requires new thermal solutions/materials. Highly integrated pre/post processors are playing an important role in the advanced mobile payload to interface the RF front-end with the digital processor. The pre/post processor unit can benefit from improvements which can leverage on the following technologies/techniques: • IF Sampling: Possibility of reduction of number of down conversion stages. • Frequency Conversion / ADC Integration – Reduction of payload complexity, cost, mass. • DBFN-based electronic antenna mispointing and thermo-elastic distortion compensation • New generation of high speed A/D and D/A developments (currently under development with ESA TRP fundings [5]). • Per user beam processing dB Carrier Level Carrier Level Carrier Level Interference Level Interference Level Interference Level Noise Floor Figure 3: Evolution of the payload processor capabilities INMARSAT4 INMARSAT XL NEXT GEN MSS Forward link A/D demux feeder links 10 x 12.6 MHz D/A demux D/A demux switch + gain control mobile-tomobile links A/D demux mux D/A mux D/A DBFN TC mobile feeds 120 x 30 MHz controller switch + gain control TM mux A/D mux A/D DBFN Return link Figure 4: Evolution of the MSS payload digital transparent processor architecture [3], [4] – On-board beamforming for reducing the number control signals. – On-ground beamforming and/or advanced multiuser techniques for capacity and flexibility increase. • On-board waveform digitization (WAVE-SAT): – On-board A/D conversion at feed level, vector quantization / data compression of the feeds’ signals and high-order modulation transmission of the data-flow (applicable for the return link: userto-gateway). – Digital transmission of the data-flow from the gateway, on-board demodulation, decompression and/or quantization re-mapping (e.g. from nonuniform to uniform), D/A conversion at feed level (applicable for the forward link: gateway-to-user). Hybrid Space/Ground processing consists in moving to ground the bulk of the digital processing by transferring the antenna feeds signals to ground [6] (see Figure 5). This future-proof approach allows for the on-ground implementation of advanced algorithms, as for example Multi-User Detection (applicable for the return link: user-togateway) and Multi-User Pre-Distortion (applicable for the forward link: gateway-to-user) in addition to the functionalities which can be supported by the on-board processor. One of the major drawbacks of the approach is related to the increase of feeder link bandwidth as the full band of each antenna feed has to be transferred to ground and vice versa. Techniques for feeder-link bandwidth reduction can be classified in three main classes [7]: • Hybrid on-board/on-ground beamforming: e1 B2 BN f eN LO-N A D Re(e1) Im(e1) Re(eN) DIGITAL BEAM FORMING Σ B1 DIGITAL DEMULTIPLEXER (Poliphase Network + FFT) LO-1 Beam/User #1 Beam/User #K Im(BN) WEIGHTS GENERATOR Figure 5: High level block diagram of a flexible broadband payload based on a Digital Bent Pipe Processor [6] References Enhanced waveform and on-ground signal processing Key waveform evolution axes which have been retained from current terrestrial trends are OFDM(A) and MIMO. OFDM(A) can be justified in future mobile satcom networks not only to be aligned with 4G terrestrial systems but to provide enhanced performance in hybrid networks and SFN mode of operation. The LTE [8] SC-FDMA [9] waveform provides limited PAPR compared to conventional OFDMA and a high degree of flexibility in terms of data rate supported which can be of interest for the satellite return link. OFDMA allows also the simplification the on-board frequency demultiplexing. Multiple Input Multiple Output (MIMO) is an emerging technique for increasing the wireless data rate and/or link reliability, and it is know to provide the highest performance gains in uncorrelated multipath rich environments. Although at first glance MIMO looks of limited attractiveness for satellite networks where multipath is mild and fading/shadowing uncorrelation is difficult to achieve, there are good reasons making it attractive: • Over terrestrial channels capacity doubling (almost) with substantial increase in power efficiency. • Over satellite almost double the capacity – MIMO shall bring some further power efficiency increase. • Over hybrid conditions: same capacity increase target with enhanced Q.o.S. (power efficiency). There is a need to understand the physics of the Land Mobile Satellite channel (LMS) MIMO channel for optimising the MIMO technique in satellite links [9]. The classical Alamouti [9] does not seem to exploit the potential capacity increase (rate ½ coding) and provides an oversized cross-polar robustness (no loss up to 0 dB cross-polar). Golden codes [10] may provide the right mix between capacity and diversity increase. In practice simpler to decode Golden like codes with reduced decoding complexity may be the right solution for the satellite MIMO application. 3 Conclusions In the authors’ opinion the identified systems techniques, payload architectures and technologies are good candidates for reinforcing satellite offer attractiveness in complementing the terrestrial networks. In line with the ESA’s Long Term Telecommunication Plan [2], ESA has initiated preliminary studies and prototyping activities in several of the identified areas but the momentum shall grow to ensure that the technology maturity will be at the right level when the commercial request will emerge. The preliminary results contained in this paper related to the application scenarios and the performance impact for the various techniques and technologies are expected to trigger the discussion and support decision makers to choose R&D lines to invest. [1] P. Angeletti, F. Coromina, F. Deborgies, R. De Gaudenzi, A. Ginesi, A.Vernucci, “SATCOMS 2020 R&D Challenges: Part I: Broadband Fixed Communications”, Proceedings of the 27th International Communications Satellite Systems Conference, 2009. [2] ESA’s Telecommunications Long Term Plan 20092013, ESA/JCB(2007)47, rev. 5, 2008. [3] R. Hughes et alii, “The Inmarsat 4 Digital Processor and Next Generation Development”, Proceedings of the 23rd AIAA International Communications Satellite Systems Conference (ICSSC 2005), Rome, Italy, 25-28 Sep. 2005. [4] C. Topping, A. M. Bishop, A. D. Craig, D. M. Howe, J. Hamer, P. Angeletti, A. Senior, “S-UMTS Processor Key Technolgies Demonstrator”, Proceedings of the 10th International Workshop on Signal Processing for Space, Rhodes Island, Greece, 6 - 8 Oct. 2008. [5] L. Hili, P. Angeletti, M. Nikulainen, “Overview of ESA Developments on High-Speed and Deep Sub-micron Digital Technologies”, ESA Workshop on Advanced Flexible Telecom Payloads, Noordwijk, Netherlands, 18-20 Nov. 2008. [6] P. Angeletti, G. Gallinaro, M. Lisi, A. Vernucci, “Onground digital beamforming techniques for satellite smart antennas”, Proceedings of the 19th AIAA International Communications Satellite Systems Conference (ICSSC 2001), Tolouse, France, 2001. [7] P. Angeletti, N. Alagha, “Space/Ground Beamforming Techniques for Emerging Hybrid Satellite Terrestrial Networks”, Proceedings of the 27th International Communications Satellite Systems Conference, 2009. [8] 3GPP TS 36.211 V8.1.0 (2007-11), “3rd Generation Partnership Project; Technical Specification Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA); Physical Channels and Modulation. [9] P.R. King, and S. Stavrou, “Land mobile-satellite MIMO capacity predictions”, IEE Electronics Letters, vol. 41, no.13, pp.749-751, June 2005. [10] K.P. Liolis, J. Gómez-Vilardebó, E. Casini, and A. Pérez-Neira “On the Statistical Model of MIMO Land Mobile Satellite Channels: A consolidated Approach”, Proceedings of the 27th International Communications Satellite Systems Conference, 2009. [11] S. Alamouti, “A simple transmit diversity technique for wireless communications”, IEEE J. Select. Areas Commun., vol. 16, no. 8, pp. 1451-1458, 1998. [12] G. Hyung Myung, Junsung Lim, J. David Goodman, “Single carrier FDMA for uplink wireless transmission”, in IEEE Vehicular Technology Magazine, IEEE Sept. 2006, Vol. 1, Issue 3, pp. 30-38. [13] J.-C. Belfiore, G. Rekaya, E. Viterbo, “The golden code: a 2×2 full-rate space-time code with nonvanishing determinants”, IEEE Trans. On Information Theory, April 2005 Vol. 51, Issue 4, pp. 1432- 1436.