Semi-blind methods for channel estimation and data recovery in multi-carrier communications

Most communications systems require channel estimation for equalization and symbol detection. This task is currently achieved by using pilot symbols which consume a non-negligible part of the throughput and power resources. Semi-blind (SB) methods allow us to reduce the pilot size while preserving the overall system performance by exploiting jointly the information from pilot and data symbols. The very aim of this project is to design efficient SB identification methods dedicated to multi-carrier/MIMO communications systems. Three main objectives are targeted: (i) Investigate the performance limits and robustness of the SB approach in different scenarii and compare them to pilot-based and differential-modulation based methods; (ii) Develop appropriate (batch and adaptive) algorithms for SB channel estimation/equalization/shortening in multi-carrier MIMO systems; (iii) Investigate SB methods with side information including induced cyclostationarity and partial learning