Skip to main content

MObile Wideband Global Link sYstem

Exploitable results

The overall technical objective of the Mowgly project was three-fold: - assess the business case; - define and specify an end-to-end 'industrial' telecommunications system for mobility applications taking advantage of existing or planned broadband satellite access architectures in Ku and / or Ka band; - develop prototype system for experimentation onboard aircraft, ships and trains. In particular: - The existing standards have been studied and analysed in detail. - Available results and theory on modelling the mobility environments, as well as existing mobility solutions (from other projects) were reviewed. - Three separate simulation platforms: link-level, packet-level and system level were developed for the purposes of the study.%l - The impact of mobility on the link, system and network performances were analysed through simulations.%l - In the physical layer the following main results were produced: a) A new model for the multipath fading channel was developed theoretically and simulated. Moreover the impact of accurate channel modelling on the link performances was simulated. b) Solutions (space transmit diversity) for solving the periodic fading effect in the return link in the rail environment, were proposed.%l c) Novel LDPC decoding and carrier acquisition algorithms for DVB-S2 were developed and simulated. - In the network-level the following main results were produced: a) Critical review for IP-level mobility & multicasting, and security indicated that the major obstacle for the successful adoption of these mechanisms in a mobile DVB-RCS system is due to the higher-than-terrestrial link errors that may be observed. With proper link budget dimensioning these mechanism can be applied in mobile DVB-RCS. b) Simulations conducted on burst synchronisation showed that cumulative time correction method with 3 sec guard times can be used to maintain burst synchronization even without periodic GPS updates. However, with periodic GPS updates, (although absolute performed better than cumulative) both absolute and cumulative time correction methods resulted in burst timing errors that cannot be compensated for by increased guard times. The interaction between periodic GPS update and SYNC-CMT based closed loop time correction self-induced timing errors. If periodic GPS updates are to be used, SYNC-CMT based correction should be disregarded. c) Simulations conducted on the proposed position-based spotbeam handover mechanism showed that the handover failure rate can be kept below than 1.4% for 95% of time as long as PER is kept less than 10-2, even with memoriless NCC handover execution design. With PER=10-3, the handover failure rate can be less than 0.28% for 95% of time. Majority of handover failures were due to TIM message loss on the forward link. With a simple improvement on the NCC side of the handover execution phase (NCC with memory), the failure rate can be less than 0.11% for 95% of time even with PER as high as 0.05. Regardless of the NCC-side of the mechanism, at PER=0.1, the observed handover failure rate was as high as 40%. d) Simulations conducted on handover delay showed that the handover delay on the return link was around 1.21 seconds for more than 90% of the successful attempts. A second peak in the histogram was ob-served at around 2.4 seconds. The second peak was created due to packet loss on PAT/PMT/SCT/FCT/TCT tables, which forced the RCST wait for the next table on the forward link. Even with PER as high as 0.1, the percentage of successful handover attempts that created more than 2 sec delay on the return link was less than 14%. The handover delay on the forward link for IP packets was upper-bounded by the physical layer synchronisation delay on the new forward link, which is about 100 msec. For non-IP content, the hand-over delay on the forward link depended on the correct reception of PAT/PMT tables. This exhibited forward link handover delay around 1 second for more than 90% of the successful attempts. e) Simulations analyses may result in slightly different results for larger uplinks. This is because SCT, FCT, TCT tables tend to be larger for higher number of carriers; and they may suffer from a higher loss ratio for the same PER. Similarly, PAT and PMT tables grow with the number of programs being broadcast in the system. However, we do not expect significant change because the TIM size is not a function of network size, and thus TIM loss probability will not be different within large networks. f) More sophisticated spot beam handover mechanisms can be devised. However, the work in this document showed that even with a very simple mechanism acceptable failure rates can be achieved provided that PER is less than 10-3. Further reduction on the handover delay is possible if either: ¾ SCT, FCT, TCT, TBTP PIDs are kept the same across all spot-beams, or/and, ¾ SCT, FCT, TCT tables are also encapsulated within TIM messages. We favour the first solution, since the latter significantly increases the TIM size and its loss probability for the same PER. - In the system-level: a) Feasibility study type of analysis produced for both forward and re-turn links of Mowgly system. b) Results for forward link showed that it is dimensioned correctly and can ensure link closure for 99.9% of the time when severe fading effects are neglected. c) Results for return link showed that the link is power limited and the achieved service availability time is reduced to 99% and 96% of the time, depending on the chosen target PER. This issue can be solved by increasing the terminal transmit power, or the terminal antenna directivity/gain, or the Increasing transponder gain, or by reducing the transmit symbol rate requirement. d) Potential for Mowgly terminals to exceed allowed interference into adjacent satellite systems evaluated. Results showed that Mowgly terminals can potentially exceed the allowed interference threshold for a significant portion of time. Solutions to this issue exist (e.g. spreading, more directive antenna) but have not been investigated in this work. e) Dominant performance degradation sources in maritime satellite and railway satellite environments identified as shadowing and rain attenuation respectively. f) Potential for fading mitigation techniques to counteract the fading effects mentioned above discussed. In particular, in the maritime satellite environment closed loop techniques such as power control and ACM have the potential to achieve performance improvement. In the railway satellite environment, the frequency and duration of shadowing from power arches makes closed loop techniques inefficient. Should this shadowing be tackled by other means, such as antenna diversity, then closed loop techniques will be able to produce performance improvements similar to the maritime satellite environment. - ACM/SNR estimation. a) SNR estimation technique proposed for use in the DVB-S2 standard (SNORE) modelled and performance simulated. b) New SNR estimation technique developed and performance simulated. Performance also compared to that of the SNORE algorithm. The analysis showed that the new technique performs on a par and in several occasions slightly better, than SNORE. The new technique is also significantly simpler than SNORE and poses reduced computational complexity. c) ACM algorithm modelled in MATLAB and performance evaluated in channel conditions relevant to the Mowgly environments. Simulations showed that ACM clearly has the potential to improve system throughput and availability times significantly. d) The effects of the large round trip delay on ACM performance were also investigated. Simulations showed that under the assumption of a 500 ms delay and a relatively fast varying rain attenuation performance is significantly degraded. The amount of performance degradation is proportional to the fading slope. It was concluded that at the system design fade the maximum slope that is likely to be encountered, or that the system needs to be able to handle needs to be identified and the ACM algorithm parameter values should be refined accordingly. Other solutions for mitigating the effects of the round trip delay, such as predictive SNR estimation and utilisation of the estimated local fading slope, were outlined and briefly discussed. Additional investigation of these techniques is required for better understanding of the performance effects they can yield. Standardisation Interaction with standardisation bodies and dissemination of results is essential for the Mowgly project. The use of existing standards contributes to strengthen the competitiveness of the European telecommunication companies in comparison with American competitors. Not only: a standardised mobility solution is a key factor for pushing forward the commercial objectives of a future MOWGLY system. By raising its profile within the European SATCOM community, the project succeeds in achieving its objective with respect to opening the DVB standards in order to support mobile-vehicular applications. Regulation Deregulation and the emergence of private satellite operators in the past decades have made that accessing appropriate orbit/spectrum resources and establishing safe spectrum regulations have become a critical issue when developing new telecom services such as Mowgly. In that respect, mobile applications face particular difficulties in sharing spectrum with other applications due to their mobile and wide reach nature. On the other hand, the regulatory path to the International Telecommunications Union (the 'ITU') has been opened in the past years for systems such as Connexion by Boeing and regulatory mechanisms are already in place at international, regional and national levels. More precisely, one remembers that in June 2003 the ITU has granted status and recognition to these applications at Ku-band, including designation a portion of Ku-band for satellite communications with ships and aircrafts, definition of the associated technical limitations to share spectrum with other applications and recommendations to administrations for their national licensing processes. Satellite communications with trains, on their side, involve reduced international regulations and should be handled under existing national VSAT regulations with possibly some adjustments. Since then, European and US national regulators in particular have been transposing, harmonising and complementing these international regulations for application in their national airspaces, waters and territories. It offered Mowgly the rare opportunity to have almost stabilised regulations at Ku-band in these two important regions at an early stage of the project, with surprisingly some differences between both sides of the Atlantic. From an industrial viewpoint, the Mowgly consortium has dedicated an important workforce to the analysis of these spectrum regulations so as to ensure full compliance of the system design and operations with applicable spectrum regulations and provide maximum confidence and safe and stable revenues to future operators and services providers. The consortium has also identified the regulatory promise of Ka-band for the provision of satellite based mobile communications even if the international regulations of the ITU require some further developments. These regulatory developments at Ka-band should follow similar paths as Ku-band in the recent years with thus limited concern on a favourable outcome. For the next stage, efforts will also be pursued to refine understanding of local restrictions that apply in addition to international, US and European regulations. In this activity, Mowgly strength relies on the implantation of its members in various countries and to their world-wide recognised expertise in that field. Mowgly strength also lies in the integration of the regulatory problematic in the earliest stage of design of the system providing the highest level confidence to future operators and service providers that regulatory risk on revenues is minimal.