Research On Energy Management Mechanism And Its Key Technologies Of Lunar Rover CPS

Lunar rover is a typical cyber-physical system(CPS).It is open for external physical environment,and couples their cyber and physical parts in network by tightly integrating with computing,communication and control to form an open feedback control,and provides critical services including real-time perception of physical environment,intelligent decision-making of behavior,effective coordination of resources,and so on.Energy management is the guarantee for lunar rover to carry out scientific exploration mission in real time and provide energy.The energy output of solar array and battery is a continuous physical process.The task scheduling of computing system is a discrete computing process.The computing process and physical process interacts continuously to guarantee the reliable operation and sustainable energy supply.Therefore,the energy management system of lunar rover is also a CPS.The interaction of computing process and physical process breaks the traditional separation design methods of scheduling and energy management.Thus,the energy management of lunar rover must combine scheduling theory and energy management techniques to develop new methods.The main contributions of this dissertation are as follows:1.Based on analyzing the energy management mechanism of lunar rover,combining the typical interaction features of cyber world and physical world,the interactions of physical entities and computing entities are studied.Finally,the energy management model of lunar rover CPS is proposed,and it provides reference for energy management system implementation of new generation lunar rover CPS.2.As the fixed priority preemption threshold scheduling(FPPT)doesn't take account of the harvesting energy,the energy level of battery and the energy consumption of computing tasks,FPPT is not suitable for computing system of lunar rover CPS.To solve this problem,an energy related preemption threshold scheduling(ERPT)is proposed,and the preemption threshold assignment algorithms are also presented.Simulation results indicate that the ERPT scheduling can reduce the tasks preemption effectively.3.Due to the tasks scheduling need to adapt the non-energy constraint case and energy constraint case in the energy management of lunar rover CPS,a group-based energy adaptation task scheduling algorithm(GATS)is proposed.Simulation results indicate that GATS can reduce the tasks preemptions in the case of non-energy constraint and reduce the battery-mode switches for increasing the average energy level of battery in the case of energy constraints.4.Because of the nonlinear feature of solar array energy output,the energy management feasibility testing is a challege for lunar rover CPS.In this dissertation,an energy management feasibility testing method is proposed.It is verified by four energy strategies with workload scheduling.Simulation experiment is taken to analyze the physical characteristics and cycle life of battery.Simulation results show that BEF can decrease battery cycle number and reduce battery capacity fade.Furthermore,to overcome the problem of software simulation methods can not truely reflect the system energy consumption and battery dynamic,a semi-physical simulation platform of energy management is built in this paper.We analyze the effect of context switches on computing system energy consumption,also analyze the effect of energy management strategies on battery dynamic.Under the conditions of present experiment,reducing the context switches of1% can decrease the energy consumption of 0.6%,and BEF strategy can obtain a higher voltage level.