Abstract
My favorite homework problems challenge students to grow beyond direct procedural mathematics, engaging their reasoning with higher level tasks from Bloom's Taxonomy: Analyze, Evaluate, & Create. This whirlwind tour presents four flavors of signal processing homework problems. First, graph-based conceptual problems require students to reason from figures, rather than deriving equations. These problems scaffold students as they transfer intuition from closed form analytic equations to the often messy time series and power spectra obtained with real world data. Second, Matlab problems require students to work with audio data on problems that are often deliberately underspecified. These problems introduce students to the cyclic trial and error nature of engineering design. Third, module problems require students to design systems to meet a specification, while also considering some form of cost. Real engineering is almost always a multivariate optimization problem balancing performance with constraints. Finally, story problems require students to identify common misconceptions in the mouths of fictional peers. Team dynamics reward engineers who can succinctly identify the shortcomings or misconceptions in poor approaches as well advocate on behalf of their own designs. This talk may include some audience participation.