Due to the evolutionary nature of computing and communication technologies, smartness and intelligence have become inevitable requirements for all futuristic systems. How to network these futuristic systems and make them deliver their services effectively is a major challenge. This challenge could be achieved by using wireless communication technologies. But, these smart systems need to be interconnected in seamless manner with facilities such as dynamic connection and disconnection, re-configurability, self-configurability, bandwidth optimization, etc. The existing wireless communication provides few of these facilities for the interconnection of generic computing devices such as server computers, desktop computers and laptop computers. A smart system highly depends on the individual real-time data monitoring for their effective performance. Hence, existing wireless communication technologies need to be customized for heterogeneous, event-driven and proactive smart systems. Hence this paper proposes a learning automata based cross layer framework for wireless networks using context awareness which aids in the reduction of energy consumption and better management of resources. The proposed technique has been tested on a simulated wireless multimedia network.