Computational Investigations of Temporal Odor Processing Mechanisms to Serve Chemotactic Orientation: Olfactory Guided Orientation of the American Lobster HOARUS AMERICANUS

点击翻译按钮获取中文摘要

9631665 Atema The mechanisms by which animals detect objects by their sense of smell are largely unknown. In natural settings odor plumes are broken up by turbulence into many discontinuous odor patches, hence locating an odor source is a very hard problem. Computational approaches, using artificial neural nets have proven highly effective in recognizing complex patterns. Although most studies of animal pattern recognition are based on visual or acoustic inputs, the research funded this grant is focused on recognition of temporal odor patterns. It is proposed that the rapid concentration fluctuations that animals encounter in turbulent odor plumes contain spatial information and that animals use such dynamic patterns to guide them to the source of the odor. The computational research funded by this grant explores candidate search strategies for such gradients and tests these strategies under various plume conditions. In addition to increasing our understanding of how animals use their olfactory systems, the results obtained by this study are expected to be applicable to robot guidance in underwater search missions, such as locating sources of pollution in the natural environment.