Interesting article in IEEE Comms Magazine (embedded below) about the Moving Relay Node (MRN). 3GPP has done a study on a similar topic available in 3GPP
TR 36.836. To make the case for the MRN they provide a reference scenario of high speed train
The TGV Eurostar in Europe is 393 m long, moves at speed
reaching 300 km/h. The Shinkansen in Japan
has similar characteristics, with 480 m long, 300 km/h of commercial speed. The high speed train in China
is 432 m long moving at speed reaching 350 km/h.
Due to fast moving and well shield carriage, the network in
high speed train scenario faces severe Doppler frequency shift and high penetration
loss, reduced handover success rate and increased power consumption of UEs.
To improve the coverage
of the train deployment, access devices can be mounted on the high speed
train, providing a wireless backhaul
connection via the eNBs along the railway by outer antenna e.g.
installed on top of the train, and wireless
connectivity to the UEs inside carriages by inner antenna installed
inside.
MRN is a good solution but when it has to operate alongside with many other technologies can pose challenges. The IEEE article summarises it as follows:
Furthermore, new challenges regarding interference management arise due to the use of MRNs. As the distance between an MRN and the vehicular UE served by it is very short, the MRN and the vehicular UE can communicate with each other using very low power. In addition, the VPL can further help to dampen the signal of the MRN access link that propagates out from the vehicle. Thus, compared to direct transmission, the use of MRNs generates less interference from the access link, for both downlink and uplink, to UE outside the vehicles. This is appreciated in a densely deployed urban scenario where link availabilities are usually dependent on interference rather than coverage. For the backhaul link, however, the problem becomes complicated, as interference is expected both between different MRN backhaul links, and between MRN backhaul links and macro UE. The use of predictor antennas can improve CSI accuracy to enable the use of advanced interference avoidance and cancellation schemes for the backhaul links. Nevertheless, whether enhancements on the current intercell interference coordination (ICIC) framework in LTE are needed to support the use of MRNs still requires further investigation.
I have been thinking of possible use of 8x8 MIMO, this can be one possible scenario where the network may use 8x8 or even 4x4. Anyway, the complete article is embedded below: