Insect-Inspired Self-Motion Estimation with Dense Flow Fields--An Adaptive Matched Filter Approach.

摘要: The control of self-motion is a basic, but complex task for both technical and biological systems. Various algorithms have been proposed that allow the estimation from optic flow on eyes. We show two apparently very different approaches to solve this task, one technically biologically inspired, can be transformed into each other under certain conditions. One estimator based matched filter approach; it has developed describe function motion sensitive cells in fly brain. estimator, Koenderink van Doorn (KvD) algorithm, was derived analytically with background. If distances objects environment assumed known, estimators are linear equivalent, expressed mathematical forms. However, most situations unrealistic assume known. Therefore, depth structure needs determined parallel parameters leads non-linear problem. It shown standard least mean square approach used by KvD algorithm biased estimator. derive modification order remove bias demonstrate its improved performance means numerical simulations. For beneficial spherical visual field, similar many flying insects. case representation simplified. Based result, we develop an adaptive systems nearly field. Then only eight about memorized updated during self-motion.