Timing Synchronization in MF-TDMA Systems for Geostationary Satellites[Topics in Radio Communications]

The most expensive costs in satellite communication are incurred by the space segment. Therefore, effort should be focused on the efficient use of this resource. One aspect is the optimization of the physical layer, to approach the Shannon limit of channel capacity. In IP-based networks, communication between arbitrary terminals can be established, which must hold for IP-based satellite networks as well. A commonly used topology is the star network, where data are routed via the master station, which in case of satellite communications introduces a double hop for terminal-to-terminal connections [1]. A double hop requires twice the bandwidth and twice the response time, which is very unattractive for interactive services. This drawback is avoided with a fully meshed topology, where terminals can communicate directly. Of course, an appropriately designed access system is required in this context. The well-known multi-frequency, time-division multiple access (MF-TDMA) scheme slices the channel capacity along both time and frequency axes. Each fragment (slot) can be used by any station for direct communication [2, 3]. This article describes the method on an abstraction layer, starting with the high-level timing that copes with frame numbers; the low-level timing that maps the frame numbers to time instants of the local time scale; and ends with the time stamp handled in the hardware, where calculation delays and filter flushing affect the received time stamp. Compared to commercial systems, the method described in this article uses differential time stamps, which reduces the effort of the synchronization procedure.