A variety of Multiple Input Multiple Output (MIMO) techniques have been proposed for satellite communications [1]  As the cited paper notes, in addition to interference reduction, MIMO offers gains through: space and multiuser diversity; spatial multiplexing; along with array gain and coding gain.  The key observation is that “ . . . these merits come from an information theory point of view at no extra cost concerning transmit power or bandwidth.  Instead, MIMO-based systems take advantage of what is probably the last unexploited frontier in wireless communications, the spatial domain.” (Italics in original text)

​

MIMO techniques are built on the theory of space-time wireless channels as described in [2].  A basic approach is to use knowledge of the channel impulse response to modify the radiation patterns of antenna arrays at both the receiver and transmitter.  If there are M transmitter antennas and N receiver antennas, an M x N channel matrix can be formed with sufficient information.  Construction of this matrix is difficult and may require a fuller characterization of the propagation channel than envisioned by the WSCE program.  Also, the theory behind this approach is complex and encompasses ideas from electromagnetic wave theory, transforms, signal processing and stochastic modeling.  However, the potential benefits are significant.



For the case of the WSCE, a GEO satellite operating at frequencies in the V/W-band, it is expected that “the channel matrix will become rank deficient and MIMO system performance degenerates to that of the corresponding SISO system” [1].  This system can be aggregated to a SIMO system as a fade mitigation technique (FMT) by using multiple, interconnected ground stations communicating with the single GEO satellite as discussed elsewhere in this proposal.

Nevertheless, the multi-beam antenna approach outlined in paragraph 3.4 allows the use of several MIMO techniques.  Each beam can be thought of as the broadcast channel of multi-user MIMO scheme.



Continuing the discussion above, each signal is precoded based on knowledge of the M x N channel matrix before transmission and decoded on reception.  The codes for this are usually nonlinear.  However, within the multi-beam context of MIMO the European Space Agency (ESA) was able to do precoding based on minimum mean square error (MMSE) approach with good results.  Each ground-based gateway station was required to construct the entire channel matrix before precoding of the signal began, i.e. it needed Channel State Information (CSI).



Another approach called Multi-beam Opportunistic Beamforming (MOB) uses channel data from multiple ground stations that has been collected at a gateway and sent to the satellite to inform the precoding of the signal.  It introduces randomization into this procedure along with knowledge of the layout of ground stations and the spatial power density at each site.  Again CSI is a key requirement.

A variety of MIMO over satellite techniques are available.



The prospects/concerns of Multi-User MIMO for GEO satellite communications are summarized in [1] as:

a) Although this approach seems viable for near-term practical application, it does require Channel State Information at the Transmitter (CSIT), requiring some kind of feedback.

b) Improved channel modeling is necessary detailing temporal autocorrelation of the channel.  Recent work on rain attenuation time series synthesizers is noted.

c) Nonlinear precoding coupled with use of multi-beams.  The Tomlinson-Harashima Precoding  (THP) technique is highlighted.

d) Time overhead may be incurred for channel training and estimation procedures to run before the precoding can begin.

e) Hardware: Precoding schemes like THP induce nonlinear behavior in High Power Amplifiers (HPA); THP will also need a waveform distorter incorporated into the system; and the precoding must minimize the variances in the signal-to-noise plus interference ratio (SINR) of outputs within the same beam.

The development of the Experiment Plan will be cognizant of the issues raised by MIMO satellite communications, especially within the areas of channel characterization and in the construction of the modeling and simulation components.  This will require inclusion of first and second order statistics.

Inclusion of significant MIMO capabilities in the communications portion of WSCE will be at the discretion of the Air Force.

References

[1] Arapoglou, Pantelis-Daniel, Konstantinos Liolis, Massimo Bertinelli, Athanasios Panagopoulos, Panayotis Cottis, and Riccardo De Gaudenzi. "MIMO over Satellite: A Review." IEEE Communications Surveys & Tutorials (IEEE) 13, no. 1 (2011): 27 - 51.

[2] Durgin, Gregory D. Space-Time Wireless Channels. Upper Saddle River, NJ: Prentice Hall PTR, 2003.