Earth Signals & Systems Group

PROJECTS (essg.mit.edu):

  • CAOS Cooperative Autonomous Observing Systems  - Visit a new initiative to use small unmanned aircraft systems to better understand environmental and ecological phenomena.
  • Animal Biometrics(SLOOP):  One of the oldest and largest systems for identifying individual animals and the first to propose adaptation of generic visual features into specific animal identifcation systems that learn from relevance feedback gathered from experts and crowds.
  • Statistical Theory of Inference for Coherent Structures (STICS):  A statistical theory of inference for coherent structures, which posits that by considering the pattern information associated with fields inference problems that are difficult for classical methods to solve become tractable, including Data Assimilation, Uncertainty Quantification, Downscaling (Super-resolution), Nowcasting, Velocimetry, and Mapping among others. This perceptual framework for fluids requires us to quantify and infer from mutual information between observational and physical variables at multiple scales, leading to a scale-recursive formulation for information theoretic inference, for which we develop tractable solutions.
  • FLUid eXperiments (FLUX): Experiments in Fluids -  Discover new techniques to image and understand fluid behavior in the laboratory and in the field.
    • Planet-in-a-bottle Project: The use of coupled physical-numerical systems to study geophysical fluids in the laboratory is demonstrated in this differentially-heated rotating fluid experiment. 
    • Particle Tracking: The classical multi-subject tracking problem extended to particles in a fluid!
    • Synthetic Aperture Imaging of Bubbly Flows, for example see, Lightfield imaging of bubbles! 
    • 3D imaging under water
  • HAZards, METeorological (HAZMET):
    • Hurricane risk in present and future climates. Visit windrisktech.com.
    • Adaptive sampling of storm surges and rapid uncertainty quantification.
    • Flood hazard prediction: A statistical-method for river flood prediction using sensor networks.