Alamouti Space-Time Block Code

The magic lies in the orthogonal structure of the space-time code matrix. Even without knowing the channel, the code is designed such that the received matrix, when processed by the receiver, has the property that symbols can be separated via linear operations. The receiver does need channel estimates to decode optimally, but the transmitter does not need to adapt; the code works for all channels.

Directly extending to 3+ antennas is hard: you either lose rate (symbols per transmission time) or gain rate but lose orthogonality. Higher-order space-time codes exist (e.g., for 3 or 4 antennas) but are more complex. Alternatively, spatial multiplexing (MIMO) gives better rate. For 3+ antennas, MIMO is often preferred over higher-order space-time codes.

Alamouti provides diversity and robustness but no rate increase. Spatial multiplexing divides the channel into independent streams, increasing rate but sacrificing diversity. In poor channels, Alamouti is better. In rich scattering, spatial multiplexing is better. Many systems use an adaptive switch: choose diversity in fading, multiplexing when conditions allow.

Unlike some codes, Alamouti is relatively robust to modest channel estimation errors; the linear decoder naturally averages errors over the orthogonal structure. However, severe errors (e.g., from poor training) degrade performance. In fast fading where pilots are widely spaced, pilot interpolation errors accumulate and performance degrades.