System Capacity and Spectral Efficiency

Satellite channelization

TRW's Odyssey communication system was an audio/low-data-rate system proposal similar to Iridium, with the primary difference being that it used a 12-satellite MEO constellation rather than a 66 sat LEO constellation.

The satellites were mostly of bent-pipe design, but required a mapping from feeder links to the individual satellites beams, called a "channelizer". The following diagram shows how 3 feeder links, each from a ground station, could be mapped to groups of beams. Each beam could have been mapped independently to one of the feeder links, but this would have increased power/weight/complexity. With reduced flexibility comes implementation loss though, and the design task became how to design the channelizer for reasonable capacity loss while still meeting the SWaP allocation for the satellite.

Since users can typically see multiple satellites, it is best to distribute load to underutilized satellites. The following is a visualization of ground users from MEO satellite, as well as availability of ground stations. Superimposed on video are beams.

Odyssey availability and loading

It was found that choosing beam sets with a wide spatial distribution of beams created the best capacity given the user availability and ground traffic model:

The following diagram shows the capacity carried by the constrained channelizer, with traffic having to be rerouted and sometimes blocked.

Here is a diagram of where the blockages were most prevalent in the simulation

Compare this to a fully-agile matrix channelizer, with hardly any blockage.

Software-defined radio

"Universal radio" design,spanning DOCSIS, LMDS, and MMDS systems. Used individually programmable coding, interleaving, and choice of QAM mapping.

See also evaluation of bit error probability using the channel coding tool

Filter design and other signal processing techniques

  •  FIR filter design
    • remez algo
    • canonical signed digit
    • spectral mask constraints
    • group delay metrics
  • Spread spectrum waveforms
    • Favorable correlation techniques
      • Periodic correlation and cyclic difference sequences. QR, TwinPrime, M-seq, Kasami, Gold
      • Finite correlation. Barker, Frank, Chu.
      • CDMA codebooks
    • Async vs sync CDMA capacity
  • Amplifier design
    • Intermod products and other non-linear modeling
    • Lifespan and power efficiency tradeoffs
  • Sampling and quantization
  • Spectral techniques
    • Phase noise classification / characterization
    • Spectral windows (although often not needed when correct spectral estimation metric is fully defined)
  • Eigensystems and decomposition
  • Source coding for audio and video

Satellite latitude coverage -- business decisions

For large communication systems such as Teledesic and Iridium, the system capacity is a function of not only the number of satellites and shared-link capacity, but also on the traffic model as well as satellite power capacity. One component of the traffic model is elasticity of demand of the communication service, which is dependent on the wealth density of the possible markets. If wealth were uniform on the earth, then the only geometric considerations in traffic have to do with busy times vs local solar time. But as can be seen below, the wealth and hence demand density are far from uniform in terms of latitude. These are signficant influences on satellite orbit choice.

Jump to next section: Orbits