The research and contributions presented focus on intelligent radio and signal processing techniques applied to MIMO systems. The HDR thesis focuses on the following points: (1) ergodic capacity of the MIMO channel (wireless, optical), (2) analysis and improvement of suboptimal techniques for MIMO detection through geometric interpretation of the detection problem and bio-inspired methods, (3) free spectrum band sensing techniques in the context of smart radio using cyclostationarity or eigenvalues of the covariance matrix of received signals, (4) analysis, development, and hardware implementation of sample rate conversion techniques (SRC), (5) analysis and optimization of the FPGA partial dynamic reconfiguration technique, (6) some results of experimental studies of free band detection techniques using USRPs platforms. The author also presents his future research perspectives related to securing mixed-signal electronic circuits (SOCs) (e.g., RFSOCs) against screaming channels, identification of transmitters through their radio frequency fingerprints, and securing future smart radio-based Internet of Things (CR-IoT) networks. The developed topics have been described in about 20 international journals and more than 60 papers in international conferences.