Surviving the upcoming data deluge: A systems and control perspective

Arguably, one of the biggest challenges facing the systems and control community stems from the exponential growth in data collection capabilities, made possible by the development of low cost, ultra low power sensors. These developments have rendered feasible a spectrum of new control applications, ranging from zero emission buildings to reconfigurable, self aware environments, that can profoundly impact society. However, realizing this potential, requires endowing controllers with the ability to timely extract actionable information from the very large data streams generated by the sensors, a goal that challenges the capabilities of existing techniques. The goal of this paper is to show the key role that dynamics can play in accomplishing this task. This is accomplished by establishing a connection, largely unexplored until recently, between the problems of information extraction, manifold embedding and identification of switched systems, and showing that this connection allows for recasting the problem of decision making in 'data deluged' scenarios into a tractable convex optimization form.