Modelling in-room radio channels using point processes

Point processes of various kinds have been applied to model multipath propagation
observed in a radio channel. Complex propagation medium, such as the indoor scenarios, give rise to multipath propagation, i.e. the transmitted signal arrives at the reciever via multiple paths. Each path is characterized by its socalled multipath parameter which in the simplest case include only a propagation delay and complex gain. The number of multipath components and their parameters are generally unknown and are thus modelled as random entities. Specically, since the early works 1970s, various types of (marked) point processes has been applied to model the set of multipath components. Experimentally, the points themselves are not observed, but the observed signal is a shotnoise driven by the point process.

In this talk, we develop a shot noise model for the in-room radio channel. We approximate mirror source positions for the inroom scenario using a homogeneous Poisson process in R3. Mapping this process to the propagation delays and gains, results in an inhomogeneous marked point process in time (R+) of delays (points) and complex gains (marks) of multipath components. The intensity of this point process increases quadratically and thus the early delays well separated while the later points are very dense. The received signal is a shotnoise driven by this marked point process. We apply the point process model to derive moment properties of the resulting shotnoise model.