Coded diversity on block-fading channels

This paper considers coded diversity schemes over block-fading Rician chann els using random coding techniques. Two random coding upper bounds on the e rror probability of block codes are derived: a new bound and a simpler but looser bound assuming binary input distribution. Also, a new lower bound fo r any block code is derived using the strong converse to channel coding the orem. The lower bound shows that the new random coding upper bound is quite tight. Furthermore, it is shown that the maximum achievable diversity orde r in a block-fading channel with finite interleaving depends not only on th e number of subchannels L, but also on the code rate R and that the perform ance can only marginally be improved by increasing the block length of the code. The random coding upper bound and the lower bound are shown to conver ge to the capacity outage for large channel black lengths N, demonstrating that the capacity outage can be used for estimating the error probability o f coded diversity schemes.