Non-Stationary Vehicular Channel Estimation in Intelligent Transportation Systems
Dr. Thomas Zemen
Telecommunications Research Center Vienna (FTW)
DEI - SALA CONFERENZE
19 aprile 2012
This talk presents an iterative channel estimator for non-stationary vehicular channels as occur in intelligent transportation systems (ITS). For vehicle-to-vehicle communications high relative velocities and large delay spreads must be accommodated. We define a finite set of pre-calculated subspaces in the time and frequency domain spanned by generalized discrete prolate spheroidal sequences. These subspaces represent different hypothesis on the support of the Doppler power spectral density and the power delay profile. The best subspace is chosen using a probabilistic bound on the reconstruction error based on a single received data block allowing for fast adaptation to changing second order statistics of the fading process. We extend the subspace selection method to multiple observations within the stationarity region of the fading process to increase the probability for selecting the best subspace. We validate the subspace selection scheme in an IEEE 802.11p compliant link level simulation for a relative velocity range from 0 to 111m/s = 400km/h. A reduction in number of iterations and hardware complexity by a factor of three to five and in terms of bit error rate by one to two magnitudes is demonstrated.
Area di ricerca: