Radiobots: The autonomous, self-learning future cognitive radios

A futuristic autonomous self-learning cognitive radio (CR), also called as a Radiobot, is proposed. The Radiobot is defined to be a radio device that is capable of self-managing and self-reconfiguring in real-time to match its RF environment while continuously self-learning from its past experience to achieve: 1) autonomous communication and awareness of experienced RF environment, 2) spectrum coexistence/efficiency including dynamic spectrum sharing (DSS), 3) inter-operability in heterogeneous RF network environments, 4) multi-mode operability (simultaneous operation over multiple modes/networks), and 5) power efficient green communications. In this paper, we present a system level architecture of the Radiobot and our current work on self-learning guided wide-band spectrum-sensing. We also discuss future research directions in order to make the concept a reality, including necessary cognitive algorithms critical for its operation and the need for real-time reconfigurable hardware and RF antennas.