This thesis is concerned with multiple-antenna wireless communication systems, initially, a forward-backward auto-regressive process is presented to predict the down-link channel of a LTE-TDD multi-user MIMO system. With the proposed channel prediction algorithm, an integrated block diagonalisation (BD) precoding scheme is proposed for improving the sum system capacity in a time-variant Rayleigh fading channel environment based on a Jakes model. Furthermore, a hybrid testbed, constructed using National Instrument PXIe platform, was used to compare the performance with that of the conventional BD precoding method where channel prediction is not present. Then various novel synchronisation and calibration schemes are presented and studied using another software defined radio (SDR) antenna array testbed which includes an array of USRP boards manufactured by Ettus Research and National Instrument. This testbed provides a flexible platform to perform a wide range experimental array processing studies with different geometries and a varying number of antennas. Furthermore, various localisation algorithms are evaluated within a controlled anechoic environment as well as outdoors. In particular, large aperture array (LAA) localisation techniques and multiple signal classification (MUSIC) techniques are studied for locating the position of a signal source operating in both in the far field and in the near-far field of the array. Finally, this thesis presents a novel array element pattern synthesis algorithm. Throughout the thesis, extensive experiments are presented to analyse the performance of the developed algorithms using an array SDR testbed made of off-the-shelf antennas.