Iterative APP decoding and channel estimation for multiple-input multiple-output channels

We describe an iterative decoding technique for bit interleaved coded modulation (BICM) systems where the modulator uses multiple transmitting antennas and the receiver employs space-time detection with multiple antennas. We investigate the performance of this technique on two families of error-correcting codes: convolutional codes and turbo codes, as well as two diierent types of fading channels: independent at Rayleigh and block fading channels. We also present an adaptation of the Expectation-Maximization (EM) algorithm as an alternative means to pilot symbols for the estimation of multiple antenna channel coeecients. Associated to the iterative detection, the EM channel estimation does not signiicantly increase the decoding complexity. A 64-state rate 1=2 convolutional code performs 1:5 dB from outage capacity with perfect channel state information and 3 dB with EM estimation on the block fading channel.