Adaptive AR modeling of multipath wireless communications channels

In the simulation of a communications channel one should be seeking a model that is simultaneously compact and easy to implement, but also one that sends the channels characteristics to a large degree. The approach that we follow here is to model the channel's impulse response by using two first order autoregressive (AR) processes. The two models are used to model the I and Q branches of a wideband channel sounder. The parameters of the model are determined using the LMS algorithm. The effectiveness of this approach is then determined by comparing the rms delay spreads of the measured data and that regenerated from a simulation.