| In this thesis, we propose a cognitive information processing system (cognitive processing) on an evolutionary platform for retargetable applications such as facial image recognition, image feature extraction, evolvable filters, and environmental information tracking. Cognitive processing can process multiple-sensory information on an automated system such as an unmanned vehicle or a surveillance unit in a remote site to avoid harsh terrain. Evolutionary platform supports the ability to change information processing behavior to comply with ever changing environment in order to accomplish a mission objective. The cognitive processing model can overcome particular difficulties to traditional search, exploration, and engineering decision making applications. The proposed cognitive strategies emphasize the decomposition of multi-sensory information, the re-construction of internal representations, and the cognitive processing of combined information which yield sub-optimal solutions and indicate best local system direction. Several applications, such as facial image recognition and digital signal processing, are used to verify our models and compare them with other well-known approaches. The derived simulation and synthesized results show that the proposed cognitive processing model on evolutionary platform attains better performance than those of the conventional methods. |
