Preparing for GBit/s coverage in 5G: Massive MIMO, PMC packaging by gap waveguides, OTA testing in random-LOS
Paper in proceedings, 2015
In this invited presentation I describe how we prepare for 5G in my research division. The background is that we have contributed to 3G and 4G developments with two commercial successes: The hat-fed reflector antenna for backhaul radio links, and Bluetest reverberation chambers for OTA (Oer-The-Air) characterization of devices with MIMO and OFDM. 5G means Gbit/s data rates, for which higher frequencies are needed towards the user terminal, may be up to 30 or 60 GHz. This means that the multipath will be weaker as it gradually diminishes when frequency increases. Therefore, we prepare by introducing a Random-LOS (RLOS) complement to the OTA testing in Rich Isotropic Multipath (RIMP) being provided by Bluetest's reverberation chambers. Further, high gain steerable beams will be needed, requiring new planar packaging solutions for closer integration of antennas and RF chipsets. Therefore, we prepare by research on gap waveguides. Massive MIMO have many technological uncertainties, so others also consider old-fashioned phased arrays, but they will not work in RIMP. The best hardware can only be chosen if we know how to characterize the system performance. The radiation pattern and realized gain cannot be used directly for this purpose due to all the statistical variations caused by the arbitrariness of the user. Therefore, we introduce instead the Probability of Detection (PoD) as a quality metric. This will be different for each desired number of bitstreams. Further, we quantify the difference between different PoD curves in dBiid in RIMP, i.e. in dB relative to the i.i.d. (idependent identically distributed) case, and in dBt in RLOS, i.e. in dB relative to an ideal polarization- and coverage- matched threshold receiver. The presentation will give an overview of this research.
Orthogonal frequency division multiplexing
Directional patterns (antenna)
Antenna phased arrays
Probability of detection