Performance criterion for space-time codes revisited

In this paper, we propose a new performance criterion for the design of space-time codes for the case of Rayleigh channels and coherent detection. This criterion is based on a simple expression for the upper bound on the pairwise error probability of space-time codes, and has an interesting interpretation by means of the "dual" capacity, i.e., the capacity of the code difference. The proposed criterion, as compared to the popular rank criterion of Tarokh et al, is not sensitive to the effect of finite precision in implementation; provides a tighter bound for the pairwise error probability; is a continuous-valued function of the code difference; and completely describes the performance of the code without need to be considered in combination with any other related criterion (e.g., the determinant criterion).