Multiple chaos generation by Neural-Network-Differential-Equation for intelligent fish-catching

Abstract

Continuous catching and releasing experiment of several fishes makes the fishes find some escaping strategies such as staying stationary at corners of the pool. To make fish-catching robot intelligent more than fishes' adapting and escaping abilities from chasing net attached at robot's hand, we thought something that goes beyond the fishes' adapting intelligence would be required. Here we propose a chaos-generator comprising Neural-Network-Differential-Equation(NNDE) and an evolving mechanism to have the NNDE generate plural differential equations as many as possible that can yield different kind of chaos. We believe that the fish could not be adaptive enough to escape from chasing net with many different chaotic trajectories, since unpredictable chaotic motions of net may go beyond the fishes' adapting abilities. In this paper we introduce chaos-generating system by NNDE, which has a possibility to yield uncountable kinds of chaos theoretically, then analyze the chaos with Lyapunov number, Poincare return map and initial value sensitivity.