The Optic Flow Group researches Biologically-Inspired Flow Field Computation for Sensing and Control of Ground Vehicles.
Our research is split into two broad areas:
- Computational Analysis of flow fields
- Systems modeling for visual navigation
James J. Gibson introduced the concept of optic flow during World War II. He defined optic flow as information carried by light resulting from environmental structure and the animal’s path through the environment. His ground breaking work identified how the optic flow field can be analyzed to determine the 3D motion of the animal (or camera) through the environment, and determine the structure of the environment. Nowadays, the term Optic flow is generally used synonymously with a vector field representation of motion in the image plane.
Humans and animals appear to have brain cells dedicated to the computation of optic flow and its analysis, particularly with respect to heading estimation, time-to-contact estimation, obstacle detection, and the structure of the environment. Our group analyzes data describing these neurons to determine how, and why, they compute what they do. We also analyze behavioral data to determine how humans, and animals, perform visual navigation tasks. By combining these analyses we construct computational models to understand the nature of optic flow computation in the brain and how it relates to behavior.
These models, and the insights learned during their construction, lead to the development of algorithms for visual navigation which we apply to robotic control.
Our group is housed at the Center for Computational Neuroscience and Neural Technology (CompNet) at Boston University, affiliated with the Program in Cognitive and Neural Systems, and we are partnered with Olin College who provide robotic systems support.
Our group was formed in March 2011 and is primarily funded by the Office of Naval Research ONR N00014-11-1-0535.