Projective Synchronization Analysis Of Different Neuron Models Based On Memristor

Neuron is the basic construction unit of neural network.The research on neurons by constructing neural model can greatly promote the development of bionic control,neuro-medicine and data prediction.The memristor has many nonlinear characteristics,which is very suitable for the simulation of nonlinear behavior of neuron circuits in engineering application.Synchronous behavior is an important behavioral feature of neurons.The study of the synchronization of memristive neural models can effectively promote the development of neuron related research in the field of medicine and communication.In this thesis,a non-smooth feedback controller is constructed to realize the projective synchronization of Fitzhugh-Nagumo(FHN)memristor neuron system and Hindmarsh-Rose(HR)memristor neuron system.The synchronization mechanism and characteristics of the system are analyzed based on the discontinuous dynamics theory,the specific research contents are as follows(1)A novel memristor neuron circuit of FHN is constructed,and a non-smooth feedback controller is designed to complete the projective synchronization of the memristive neuron circuit of FHN under the same structure.In the traditional FHN model,a smooth hyperbolic tangent mediated nonlinear ideal memristor is added to construct a new FHN neuron model.After selecting the appropriate system parameters by using the dynamic analysis method,the influence of the control parameters of the designed non smooth feedback controller on the synchronization effect is analyzed.The parameter diagram is used to complete the parameter analysis of the partial and complete synchronization of the system,and the simulation is carried out by Multisim.The simulation of synchronous circuit is realized on the real platform to verify the physical realizability of the circuit.(2)The dynamic behavior of FHN memristor neuron circuit in vectorial domain is analyzed.The synchronization mechanism of the system is discussed by using discontinuous dynamics theory,and the synchronous digital circuit is realized.The dimension reduction of the proposed FHN neural model circuit is carried out.The dynamic characteristics of the system in vectorial domain are analyzed after dimension reduction.The system parameters are selected for the proportional projection synchronization.The influence of the control parameters of the non-smooth feedback controller on the synchronization effect is studied.The control effect is analyzed by using the time sequence diagram and phase diagram.Finally,the synchronization circuit is realized by using Multisim and FPGA digital circuit theoretical verification.(3)The influence of control parameters on FHN and HR memristive circuits synchronization is analyzed by introducing HR memristive neuron circuit and FPGA digital circuit.Based on the introduction of the HR memristor neuron circuit,the analysis of the dynamic characteristics of the HR neuron system,the appropriate system parameters are selected to analyze the heterogeneous proportional projective synchronization of FHN and HR,and the control effect of different control parameters in the non-smooth feedback controller is analyzed.Finally,FPGA is used to simulate the same step circuit.