Modeling The Thermotaxis Behavior Of C.elegans Based On The Artificial Neural Network

Half a century ago,modern bionics was proposed,and it has been developed deep and widely.Specially,the most striking is Bionic robot,which combined Bionics with the application and requirement of the robot field.Depending on strong adaptability in complex environment and high dynamic stability,Bionic snake robot has been brought into sharp focus by many researchers in various countries.This thesis focuses on the biological nature,hopes to study the nervous structure and muscle structure of organism and brings to light the organism movement mechanism.The study of snake robot is a reference to the study of snake in nature,however,the neural structure and muscle structure of snake is too complicated to simulate.Fortunately,because the neuronal connections of C.elegans is full clarified by neuroscientist,it is a good research model.Furthermore,taxis behaviors is an important motion behavior for organism to find object and avoid the harmful environment.Thus,this thesis studies the thermotaxis behavior of C.elegans,and conducts related explorations about it.This thesis models the thermotaxis behavior of C.elegans by artificial neural network,conducts related simulation experiments,and mimics the thermotaxis behavior of C.elegans.The processes of this thesis are as follow: establish temperature distribution model to model the temperature environment where the nematode lives;analyse the muscle anatomy structure of the nematode;based on the movement process in a time period,construct the kinematics model;based on the logic of deflection movement,design nonlinear functions to model the deflection angle;learn the nonlinear functions by means of artificial neural network,realize the simulation experiment of thermotaxis behavior of C.elegans.This study is divided into the following two parts:1.This part adopts the nematode model as a point in environment model,constructs the BP neural network and RBF neural network that learn the motion logic,analyses detailedly the relevant parameters of the nematode's thermotaxis behavior in a time period,constructs several simulation comparison experiments.From the experiment results,this method models the thermal behavior of C.elegans successfully,and BP neural network is much closer to embody the actual thermal behavior and approach the essence of the movement of organisms that RBF neural network.2.Based on the muscle anatomy structure of the nematode,this part constructs the nematode body as the multi-joint rigid link model,constructs the BP neural network to learn the motion logic.Under the different time period,carries out the simulation experiments with nonlinear functions and BP neural network separately.The two kinds of deflection movement experiments are compared with the same initial condition and different time period,and time period is a quarter of motion period or half of motion period.The experiments can successfully model the thermal behavior of C.elegans by means of BP neural network,and the results demonstrate that the effect of thermal behavior approach the actual motion trail when time period is half of motion period.