A new design metric for unitary space-time codes

Abstract

Multiple-input multiple output (MIMO) communication systems can have dramatically higher throughput than singleinput, single-output systems. Unitary space-time codes for MIMO fading channels have been shown to be optimal under certain conditions when neither the transmitter or receiver has channel state information. Previous work on constructing unitary space-time codes has focused on the diversity product or diversity sum of the unitary space-time code constellation. This paper shows how these min-max design criteria may produce sub-optimal codes in terms of minimizing symbol-error rate (SER). A union bound design metric based on a union bound of pairwise error probabilities is presented, and a bound on its optimal value is derived. Examples suggest this new metric can more accurately distinguish good constellations than the diversity product and diversity sum. The complexity of the union bound design metric is reduced by using recently derived tight bounds, and a desirable -incremental property of the union bound design metric is demonstrated in an example.

Multiary turbo code fitting for unitary space-time modulation and its MAP decoding algorithm

“…How to optimize Φ is still an open problem. At the present time, there exist three design rules: the diversity sum rule [17], the diversity product rule [18] and the union bound rule [19]. As our aim is to find the gain given by the coding, we only choose the simplest method to construct Φ [17].…”