10:30 - 12:30 | Wed 30 May | Civil Engineering Auditorium | Keynote
Historically, spectrum has been segmented and allocated for different services and systems by regulatory authorities in regions and/or countries. However, with the proliferation of wireless communications technologies for various applications, such static spectrum management has shown to be ineffective and inefficient. In commercial contexts, static spectrum allocations have shown to result in inexcusably wasteful underutilization of spectrum resources. In military, homeland security and emergency communications contexts, on the other hand, static spectrum management is increasingly becoming, simply, unacceptable. In these applications, dynamic spectrum management is not a luxury but a requirement. Dynamic spectrum management, however, requires improved spectrum situational awareness (SSA) capability: The ability to monitor, comprehend and act on RF spectrum._x000D_
Many years ago, we envisioned the wideband autonomous cognitive radios (WACR) as a technology solution to providing spectrum situational awareness (SSA). WACRs are radios that can sense and comprehend the state of the overall system made of the radio, spectrum, the network and the end-user and have the ability to self-configure the mode of operation over a wide non-contiguous spectrum range in response to this spectrum state. A WACR achieves spectrum awareness through a process called the wideband spectrum knowledge acquisition. Unlike spectrum sensing discussed in dynamic spectrum sharing (DSS) literature, wideband spectrum knowledge acquisition is aimed at fully characterizing a wide spectrum of interest by not only detecting signals, or the absence of them, but also classifying and associating signals in order to identify all spectral activities. Thus, WACR technology provides an ideal platform to build RF spectrum situational awareness systems needed for defense, military, homeland security and space applications as well as consumer wireless telecommunications._x000D_
Recently, there is an increasing desire to achieve spectrum situational awareness over several GHz-wide spectrum. This is a challenging task given the limitation of instantaneous bandwidths of state-of-the-art radios to only several hundred of MHz of spectrum. The WACRs provides a solution by developing a technology that integrates software-defined radios, machine-learning and reconfigurable hardware._x000D_
This keynote address will discuss how a comprehensive SSA solution can be developed around the spectrum knowledge acquisition capability of a WACR. Machine-learning aided, context-aware decision-policies are developed to effectively scan several GHz-wide spectrum within the instantaneous bandwidth limitations imposed by state-of-the-art radios. Detected spectral activities are then isolated and relevant features are extracted. The approach calls for on-demand, context-based feature extraction. To handle real-time requirements, a hierarchical SSA framework is developed with the aid of a hierarchical signal classification approach. The hierarchical SSA framework allows for the fact that SSA needs can be dynamic and thus it must be able to support redefining the SSA needs in real-time. For example, at one time the interest may be in distinguishing an out-of-place radar signal in a spectrum band primarily allocated by communications systems, whereas another time it may be to identify a possible GPS spoofing signal before it succeeds. In the hierarchical SSA framework, powerful machine learning techniques are combined with advanced decision-making algorithms to achieve these goals.
No information added