Research On General Channel Modeling Method And Signal Detection Scheme Of Molecular Communication Via Diffusion

Molecular communication(MC)is one of the feasible communication schemes for building nanomachines networks(nanonetworks).Nanonetworks could expand limited capability of individual nanomachine and bring huge progress to various fields such as biology,medicine,health,national defense and industry,etc.The scheme of molecular communication via diffusion(MCv D)attracts the most attentions due to its advantages of good energy efficiency and easy implementation among MC.At present,the research of MC is mainly focused on theoretical researches,of which the channel modeling is foundation.Although various channel modeling methods for MCv D have been proposed,the domain still lacks the general channel modeling method to unify various cases after a lot of literature surveies.Furthermore,some promising and important MC applications,such as precision drug delivery,require nanomachines to communicate in mobile.In mobile MC,the received signal is stochastic and more complicated which brings difficulties to signal detection.Meanwhile,nanomachines have limited computational capability due to their limited size.Therfore,mobile MC scenarios need signal detection scheme with low computational complexity.This thesis studies the above two problems and has two main innovative contributions:(1)Propose two general modeling methods of channel model in MCv D.Based on Newton's second law of mechanics,this thesis analyzes the force of information molecules and obtains the force equation-general Langevin equation.By demonstrating the relationship between Brownian motion and white noise,the mathematical theory tool of the force equation-Ito? stochastic differential equation is derived rigorously,that is,the first general channel modeling method.The mathematical laws express the physical laws of complex diffusion processes under different conditions.Meanwhile,by probability theory and Markov process,the Fokker-Planck equation of the probability density function of information molecule about position is derived,that is,the second general channel modeling method.The differential equation is used to describe the physical diffusion process of information molecule under different conditions.In addition,this thesis also proves the equivalence of Ito? stochastic differential equation and Fokker-Planck equation.According to above theoretical work,this thesis presents three examples,using the two equations for each example,and verifying the validity and advancedness of the two equations by correctness of the two methods.In this thesis,through a large number of Monte Carlo simulation experiments,the three examples are used to simulate the random movement of information molecules of Ito? stochastic differential equation and the theoretical curve simulation of Fokker-Planck equation.The validity of the two equations is verified by the agreement of the two experimental curves.(2)Propose low complexity adaptive signal detection method.According to the convexity feature of the received signal,this thesis proposes a mathematical formula to theoretically analyze and quantify the local maximum convexity indicator of received signal.Under the influence of intersymbol interference(ISI)and noise,this indicator also automatically adapts to the dynamic and stochastic change of mobile nanomachines and correctly judge the received signal.A large number of Monte Carlo simulation experiments verify that the proposed method is effective.At the same time,the computational complexity of the proposed method is O(n),and has the advantages of low computational complexity and good detection performance.Through rigorous theoretical derivation and numerical simulation experiments,this thesis analyzes and certificates Ito? stochastic differential equation and Fokker-Planck equation are two equivalent general channel modeling methods in detail.This work completely solves the problem of the lack of general theoretical framework and channel modeling methods in MCv D.The significance of this work for MC is similar to that of Maxwell's equations to conventional communications via electromagnetic wave.That is it provides theoretical framework and general modeling methods which will enhance the systematization and simplify the complexity of theoretical studies in MC.In addition,the proposed method avoids ISI suppression and dynamic distance estimation,and is effective for mobile MC systems with limited computational capability.