Proceeding of the 7th International Conference on Ubiquitous Robots and Ambient Intelligence
巻, 号, ページ
pp. 180-183
出版年月
2010年11月
出版者
和文:
英文:
会議名称
和文:
英文:
開催地
和文:
英文:
Busan, Korea
アブストラクト
Localizing an odor source, known as the chemical plume tracing (CPT) problem is a difficult technical problem, that many animals are able to solve. Thus, understanding the strategies used by such animals is a promising approach. We focused on analyzing the pheromone source localization strategy of the silkworm moth, Bombyx mori. The main advantage of this model system is the virtual absence of noise in pheromone orientation that allows the acquisition of clear input-output relations. The responses of moths to pheromone consist of a distinctive series of movements that can be considered to be elementary building blocks. Using a behavior classifier based on Self-Organizing Maps (SOM), we have been able to reliably isolate the initial movement phase following reception of a pheromone stimulus from experimental data. Our SOM can thus be used as an estimator for stimulus timings. The SOM approach will be developed further for classifying other behavioral elements and merged with information from studies identifying neural mechanisms in the brain that are controlling pheromone source localization behavior in the silkworm moth.
各種検索
サポート
ヘルプ