Digital communication based on Multiple Input-Multiple Output (MIMO) has drawn considerable attention in the communication society these last years thanks to the enormous gains they promise. MIMO systems are characterized by the use of multiple antennas at both the transmitting and the receiving ends. Thanks to the intelligent combination of signals at the transmitter and the receiver and making use of the random properties of the channel, MIMO systems increase data rates and improve link reliability simultaneously. A key assumption for the good working of such systems is that the fading between pairs of transmit-receive antennas be independent and identically Rayleigh distributed. This assumption is reasonably accurate when the environment is extremely rich in scatterers and the antennas are sufficiently far apart. However channel measurements have shown that the properties of the MIMO channel may highly deviate from these ideal propagation conditions. This thesis is devoted to the design of new space-time signaling schemes that take into account the non-ideality of the channel. Therefore characterizations of the MIMO channel based on analytical and physical MIMO models and based on outdoor and indoor measurement campaigns are carried out. Those models are then used to evaluate accurately the performance of space-time signaling in the presence of real-world conditions. New code design criteria are derived for the case when the transmitter does not possess any information about the channel and when the transmitter can make use of some statistical channel properties. The new derived schemes are shown to perform much better in the presence of real-world channels than codes originally developed for ideal propagation conditions. ...