Distributed Data Traffic Scheduling With Awareness Of Dynamics State In Cyber Physical Systems With Application In Smart Micro-grid

A cyber physical system is a system in which information processing and physical dynamics are tightly integrated such that it creates a good foreground for the realization of real-time sensing,intelligent information processing and physical dynamic control of large-scale engineering system.Specifically,this paper has taken smart mirco-grid as an example of CPS and studies the medium access control layer design of distributed multi-hop wireless networks for CPS in the presence of mutual interference among transmitters.In particular,this paper studies how the data traffic from voltage monitoring sensors should be scheduled with awareness of physical dynamics state for the voltage regulation of distributed energy generations access system in smart micro-grid.Aiming at the building communication infrastructure possessing the ability of well awareness of dynamic state stability characteristic,this paper concretely studies the impact of network imperfections(including band-limited channels,communication delays,packet dropouts and transmission delays)on the formulation of distributed data scheduling strategy.The main research contents are as following:1.Firstly,this paper constructs the cyber physical system model of mirco-grid,including the physical dynamics and wireless multi-hop communication network.In the case of distributed control,controllers are supposed to be able to harmoniously regulate the voltages at multiple points of common coupling.Multicast is used.Specifically,data packets containing measurements from each sensor are transmitted along a multicast tree to a defined set of destination controllers.In that way,each distributed controller can conduct state estimate using data from varying sensors.Correlation and coupling between controllers in multicast network environment will benefit both convergence and stability of the to-be-controlled system state.Rather than centralized scheduling over a single hop,this paper studies decentralized scheduling over multiple hops under the impact of network imperfections.2.Delay-Tolerant Kalman Filtering algorithm is designed where the effect of communication delays is taken into account.This algorithm takes advantage of the information form of Kalman filtering in order to maintain scalability with the number of sensors.Based on DTKF algorithm,an algorithm is developed that calculates mutual information between an observation and the state to be estimated in distributed real-time sensing applications.Then a multi-sensor fusion method that allows communication delays is proposed.To measure the effective information amount of different packets,we propose the concept of virtual queues(VQs)by defining the equivalent bits of each measurement for the purpose of stabilizing the physical dynamics.The proposed framework integrates packet scheduling in communication and control in CPS using the novel interface of VQ.The used algorithm takes advantage of Delay-tolerant Kalman filtering and allows us to achieve distributed sensor selection and scheduling.VQ based back pressure algorithm is proposed for the distributed data traffic scheduling with awareness of dynamics state in wireless multi-hop networks.The scheduling method is then applied in the context of distributed energy generations access system voltage control in microgrid.Numerical simulations demonstrate that the proposed scheduling algorithm can achieve better performance than general scheduling algorithm without the awareness of dynamics state.3.This paper studies how packet losses affect the VQ length for nodes.The total amount of delays original DTKF can tolerant ranges from 0 to L-1.The dropped packets could not reach the controllers,which means even those measurements having smaller delay than L cannot participate in the data fusion based on DTKF.L time delay increments are added to the existing delays(0~L-1),making sure the the total amount of delays for the lost measurements ranges from L to 2L-1.So we can rightly compute effective information for the measurements which may be lost in order to avoid miscalculation during packet scheduling.Based on modified DTKF,the effect of packets dropout on the number of effective information bits of packets is studied.Then,VQ lengths considering packets dropout are obtained.Simulation results demonstrate the validity of VQ lengths calculation method and scheduling algorithm considering packets dropout under the influence of packets dropout.By utilizing VQ lengths considering packets dropout,better MSE compared with original DTKF based scheduling method can be obtained.4.This paper studies how transmission time delays affect the VQ length for nodes as well as the impact of transmission time delays on distributed data traffic scheduling.The DTKF algorithm is modified to tolerant transmission time delays,and then we derive the equations for calculating the number of effective information bits of packets with transmission time delays based on modified DTKF.Then,we obtain the VQ considering transmission delays.The impact of transmission delays on the formulation of distributed data traffic scheduling strategy is also studied.In order to verify the scheduling strategy considering transmission delay for distributed energy generations access system voltage control,two different scenarios are considered respectively: the first one is that the two sensors with delays compute their VQ lengths correctly;the second one is that the two sensors compute the VQ lengths without consideration for transmission delays.The evolution of system state for the different scenarios is simulated to compare the converging rate of the system.Further numerical results obtained through contrast experiments demonstrate the validity of the scheduling algorithm which is established according to VQ lengths affected by transmission time delays.