Measurement and modelling of human presence, through-wall attenuation and link diversity at millimetre-wave frequencies

Exploitation of the millimetre-wave (mmWave) bands at 28 GHz and above is a key part of the fifth generation wireless (5G) strategy to address the exponentially growing demand for high throughput and capacity radio access. However, mmWave signals are highly susceptible to blockage by people and/or building structures. Previous work has presumed that when the direct path is blocked, communication will be conducted by secondary paths due to reflection or scattering by the environment but has not shown how much performance would degrade. Here, our objectives are: 1) to develop a physical-statistical model of the effect of human presence/blockage at 60 GHz; 2) to statistically characterize the manner in which through-wall attenuation varies between wall types, different walls of the same type, and different locations on a given wall together with an indication of how these results scale between 10 and 30 GHz; and 3) to provide an accurate assessment of the relative quality and capacity of direct and secondary paths at 30 GHz in both indoor and outdoor environments. We used mmWave channel sounders of various types and configurations to conduct comprehensive, accurate and efficient measurement campaigns. The physical-statistical model for human presence is computationally efficient and shows very good agreement with measurements. It provides an accurate estimation of the reflection coefficient and the diffraction correction factor corresponding to indoor measurements at 60 GHz. Our through-wall attenuation measurement results demonstrate that through-wall attenuation falls into separable classes and is amenable to statistical characterization. The through-wall attenuation generally follows a Gaussian distribution in all building materials. Frequency dependence of the form of f¹·⁴⁵ is observed between 10 and 30 GHz through-wall attenuation (linear) values. Our results for the assessment of link quality show that the relative path loss on the four strongest secondary paths follow a Log-normal distribution while the Ricean-K-factor and SISO/MIMO channel capacity follow exponential distributions. They clearly show that the quality and capacity of secondary (reflected) paths at mmWave frequencies is dramatically lower than the direct (LoS) path with results obtained in outdoor microcells to be almost twice as worst as in indoor environments of comparable size.Applied Science, Faculty ofElectrical and Computer Engineering, Department ofGraduat