Information acquisition, controlled sensing, and active hypothesis testing

Measurements are ever increasingly ubiquitous, playing a central role in a wide variety of applications. However, the capabilities of any measurement device (e.g., sensitivity, resolution, dynamic range, reliability, etc.) are directly related to the deployed measurement resources/cost (e.g., size, weight, power, etc.), almost all information acquisition systems experience a tradeoff among competing goals. A novel unified mathematical formalism to understand this seemingly inherent and intuitive tradeoff is of paramount value and importance. A combination of techniques and results in stochastic control, information theory, decision theory, and statistics have proved powerful in advancing this goal and obtaining partial results. This presentation catalogues these advances and hopes to motivate further investigation of this research problem.