2A2-G07 Prediction system for stimuli timing within Chemical Plume Tracing by using Echo State Networks(Sense, Motion and Measurement (2))
Yohei Kinowaki, Y. Nakamura, Chew Jouh Yeong, Daisuke Kurabayashi · The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec) · 2013
In this paper, we realize a classification system in order to identify significant parts of the behavior exhibited by a silkworm moth, Bombyx mori. An adult silkworm moth has the ability to solve Chemical Plume Tracing (CPT) problem, which an individual robot hardly solves effectively. Since it is quite difficult to observe the chemical plumes in the air, we tried to estimate them based on the exhibited behavior. We employ Echo State Networks (ESNs) to realize an identifier, and estimate typical motions during the programmed behavior of a silkworm moth. We reported the estimated results according to a trajectory carried out by a living silkworm moth.