Hossein Roufarshbaf

Research Interests

Detection Estimation Theory in Uncertain Environments

Detection and estimation of the system parameters of interest has been widely addressed in many statistical signal processing applications such as wireless communications, biomedical image processing, and sonar target tracking. In some applications, the uncertainty of the system state space model, non-linear models, and noisy measurements are combined to make these kind of problems specifically challenging. We have successfuly developed tree search techniques in detection and estimation problems that can be applied to non-linear system models in uncertain environments.

Modulation Classification in Wireless Communication Systems

Automatic recognition of digital modulation format of an incoming signal, which has aleays been important in surveillance applications, has seen renewed interest in intelligent devices such as cognitive radios. Most existing work on modulation classification has considered only flat fading channels while frequency selectivity of wireless channels adds more complexity since incorporating blind equalization is necessary in this case, but most blind equalization algorithm require knowledge of the modulation scheme.

Joint Blind Channel Equalization and Decoding

Implementation of blid equalization algorithm for fast fading frequency selective channels makes systems more efficient by deleting training sequences. Usually in receiver equalizer and decoder are processed separately. Joint combination of channel equalization and decoder increases the complexity while improves the performance of the receiver. Among the algorithms that have been proposed for joint blind equalization and decoding, algorithms that rely on less observation symbols are more preferable specially in time varying wireless channels where increasing observation window size leads to unstationarity of the wireless channel model.