Research On Gateway Based Heterogeneous Network Communication In Cyber-physical Systems

Cyber-Physical System(CPS)is a combination system which integrates sensor technology,embedded computing technology,distributed information processing technology,communication technology and control technology.It monitors and collects information of the environment or the objects of its working area and intervenes them through various actuators.In order to achieve the above functions,data exchanging between different equipment which supported by communication network of the system is necessary.The communication network of CPS is a combination of sensor networks,actuator networks,compute r networks and some other networks.These networks are usually heterogeneous,which makes devices from different network unable to exchange data directly.To connect these devices and heterogeneous networks,implementation of gateways which equipped with various wireless communication technologies and able to connect multiple heterogeneous networks is required.The cost of gateways is relatively high compared with these price moderate equipments like sensors,meanwhile,network coverage and data transmission capacity of gateways are limited.Arranging fewer gateways,using communication resourse of gateways more effectively and selecting the most suitable gateway to connect are the key means to reduce the cost of system communication network and improve communication network performance.This thesis will study the problems of gateway placement,bandwidth allocation of gateways and gateway selection of devices.The research issues and contributions of this thesis are summarized as follows.First,this thesis investigates the problem of how to reduce the number of gateways meanwhile keeping the communication requirement satisfied.There are a large number of various devices with different transmission range and demands of data transmission in CPS.Gateway placement must make sure that every single device is able to connect at least one gateway and their traffic demands are fulfilled.Arranging gateways at the right positions can satisfy above requirements with fewest gateways.Recent studies focus on the problem of how to select one among existing devices to be a gateway in order to improve performance and reduce the cost.These works are not suitable for heterogeneous networks in CPS.This thesis defines the problem of minimizing gateway placement while satisfyin g the communication requirement of heterogeneous networks,analyzes the computing complexity and propose two heuristic algorithms.One of them is based on plane geometry,which is suitable for the situation that gateways can be placed anywhere.The other is based on grid,which can be used when gateways can only be placed in some fixed positions.Second,this thesis investigates the problem of how to allocate the bandwidth of gateways in heterogeneous networks in order to maximize the delivery rate of packets in the system.In order to monitor and intervene the physical world effectively and properly,data in CPS is time constraint.If exceeding the time line,data becomes worthless.Data transmission is successful only if they have been delivered in time.Data exchanging between devices from different heterogeneous networks can only be fulfilled by gateways,but the bandwidth is limited.Only when the bandwidth has been fully used,the data transmission can be guaranteed.Considering the complicated and unpredictable working environment of CPS,this thesis proposes two static bandwidth allocation strategies and one dynamic bandwidth allocation strategy.These two static bandwidth allocation strategies are based on two considerations separately.One of them is that gateway uses all its bandwidth to forward every packet one by one.The other one is that gateway allocate minimal bandwidth for packets in order to serve more coming packets.The dynamic bandwidth allocation strategy adjusts the distribution of bandw idth depending on the traffic condition,which can fully use the bandwidth while capable to serve coming packets.Third,this thesis investigates the problem of how can devices select their gateways in order to achieve fair benefit.Bandwidth of gateways is important resource in CPS.Obtaining more bandwidth can improve the data transmission and reduce the delayed time.Meanwhile,the load balancing of gateways can improve the performance of communication network.This thesis proposes an evolutionary game theory based gateway selection algorithm and this algorithm can guarantee the fairness among devices in CPS.Further,by using evolutionary game theory,this thesis analyzes the behaviors of devices when they obtained fewer bandwidth than others.The bandwidth allocation model has been proposed,and formulate the process of changing gateways of devices as an evolutionary game in CPS.Players are devices in CPS and payoff is bandwidth of gateways.This thesis proposes the replicator dynamics of this evolutionary game and analyze the existence and stabilization of the evolutionary equilibrium.Forth,this thesis investigates the gateway selection problem which aims at maximizing total bandwidth obtained by devices from the same network.Through the above analysis,gateways' bandwidth is important resource of CPS.Although devices from different heterogeneous networks cannot exchange data directly,devices in the same network are able to communicate mutually.The more bandwidth a network can get,the more benefit a device in this network can achieve.And this will lead to bandwidth competition among different networks.This thesis has proposed the formal definition of maximizing sub-network benefit gateway selection problem,and formulate this problem as a non-linear integer programming problem.Based on the features of this problem,this thesis proposes a non-cooperative game model,in which devices change their connecting gateways to maximize their benefit.A game theory based distributed gateway selection algorithm has been proposed.This thesis analyzes the steady state of the system and Nash equilibrium of this gateway selection game.Simulation results demonstrate the effectiveness of proposed algorithm.