Modeling of the Adaptive Chemical Plume Tracing Algorithm of an Insect Using Fuzzy Inference

In this paper, we focus on the chemical plume tracing (CPT) problem, which is a known engineering challenge. In nature, animals solve the CPT by adaptively modifying their behavior according to the environment. Therefore, we propose a CPT solution method with high engineering value by modeling the CPT algorithm of an animal. In this paper, we consider a male silkworm moth as a model. To perform CPT in a turbulent environment, the adaptive selection of the behavior plays an important role. Therefore, we performed simultaneous measurement experiments involving CPT behavior of the brain and analyzed the links between the brain's neural activity and behavioral patterns. We measured the brain's neural response in the lateral accessory lobe (LAL), which generates motion commands. We employed fuzzy inference to analyze the relationship between CPT behavior and LAL neural activity. As a result of analyzing the relationship between CPT behavior and LAL, we found that the moth modulates the behavior of the state transition probability depending on the odor frequency. We modeled the obtained phenomenon and verified its effectiveness through a constructive method. As a result, the search performance was improved compared to the conventional moth algorithm.