Signal Processing in Physical Environments: Crossing the Barrier of Traditional Assumptions

Опубликовано 17 августа 2016, 2:02

Traditional signal processing approaches are often built on well-known assumptions such as linearity, time-invariance, and incoherence among the signals involved. In reality, many of these assumptions are violated in physical systems built for challenging environments and therefore, conventional techniques fail to solve signal processing issues effectively in real-world systems. Examples of such problems include interference mitigation in the presence of amplifier saturation, tracking the time-varying delay spread of shallow water acoustic channels, oceanographic imaging in the presence of severe optical backscatter, among many others. My talk will focus on signal processing issues encountered in shallow water acoustic communications, as they relate to my research in sparse sensing and channel estimation techniques. While the techniques I will discuss have been developed with target applications in mind, the scope of signal processing innovations go beyond specific applications to the broader paradigm of other well-known problems with similar issues. Breaking the barrier of traditional assumptions and constraints will enable the scientific study of many physical processes hitherto difficult to observe as well as usher in the development of many new applications in the future. Therefore, I will also present my vision for extending signal processing theory beyond the limitations of these traditional assumptions.