Computational Sensing

Reconstruction from sparse measurements and models.

Sensing and signal acquisition are indispensable technologies in our increasingly digital world. MERL exploits widely available computational power to overhaul the signal acquisition paradigm and significantly enhance sensing capabilities.

Sensing systems aim to capture physical signals and reconstruct them digitally or extract useful information from them. We aim to fundamentally understand how signals behave and propagate in the environment and the sensing systems. Using this knowledge, we develop reconstruction algorithms to recover signals with significantly improved fidelity and robustness.

Co-developing sensing systems with novel computational methods enables us to realize the full potential of the sensing hardware. Our research results in high resolution sensing of the environment at significantly lower cost.

We have successfully applied our methods to advance the state of the art in a number of applications, including high-accuracy low-cost active depth sensing, fusion of signals from multiple sensors with different modalities, accurate sound localization with microphone arrays, and improved resolution in radar imaging systems.