Designing unique words for reliable channel estimation in multi-antenna block transmission systems

Single-carrier, multiple-input multiple-output wireless communication systems generally require knowledge of the channel to equalize a received message. In block transmission systems, such as those that utilize the frequency domain to facilitate channel equalization, short training sequences known as unique words (UWs) can be inserted into the data stream, thus providing a means for estimating and tracking the state of the channel. The problem of designing the UWs can be difficult when certain constraints are placed on the sequences. In this paper, a nonlinear optimization approach is taken to design near-optimal UWs under given constraints, such as a limit on the peak-to-average power ratio (PAPR) of the sequences. Optimality of the sequences is defined with respect to the Cramer-Rao bound on performance for an unbiased estimator