Research On Key Technologies Of Real-time Computing In Passive Systems

In recent years,the Internet of Things has achieved rapid development.Howto power a large number of Internet of Things devices has become one of the key issues that need to be solved urgently in the field of the Internet of Things today.Because the passive system using environmental energy harvesting technology has the advantages of small size,long life,and maintenance-free,it has been widely used in the Internet of Things.However,due to energy shortages and the mismatch between energy harvesting capabilities and comput-ing capabilities,passive computing systems still suffer from long task response times in practical applications.If reasonable transformation can be carried out to meet the response time requirements of real-time calculation,the passive system can quickly respond to the gateway,gain a stronger environmental awareness,and provide new development impetus for the Internet of Everything.In this thesis,the auxiliary energy storage device stores the environmental energy col-lected during non-real-time computing to provide an energy guarantee for real-time com-puting tasks,and an energy management module is added to realize low-power switching for non-real-time computing tasks.Energy sources from energy harvesters to energy stor-age devices.,Which solves the energy shortage problem faced by passive systems to realize real-time computing.To maintain the advantages of a long life of passive systems and improve environmental energy utilization,the intervention of auxiliary energy stor-age equipment is minimized in the real-time calculation process.Therefore,based on the checkpoint technology,this thesis designs two runtime auxiliary energy storage equipment intervention strategies to ensure the real-time performance of the system.By dynamically inserting checkpoints during task execution to determine real-time performance,not only ensures that real-time tasks can be completed before the deadline but also improves the utilization of environmental energy during execution.This thesis also models a passive system that supports real-time computing and describes the relationship between the volt-age threshold of the task cut-off operation and the battery intervention time and average response time.The optimization problem of battery intervention time is proposed.In re-sponse to this problem,a method to find the best cut-off operating voltage is proposed to minimize the intervention time of auxiliary energy storage equipment,thereby maximiz-ing the real-time life of the passive system.This thesis designs and implements a passive device that supports real-time comput-ing.The experiment verifies the low-power switching function of energy and the real-time computing capability of the system.Experiments have proved that,compared with the traditional load cycle strategy,two battery intervention strategies ensure the real-time operation of the system-the worst-case estimation strategy and the workload prediction strategy can complete the calculation before the real-time deadline;the time task,the cal-culation time is the same.Compared with the original execution strategy,the reduction was 32.26%and 38.82%respectively.Under the premise of ensuring the real-time perfor-mance of the system,the two execution strategies have reduced battery intervention time by 47.54%and 58.31%,respectively,compared with the direct use of auxiliary energy storage equipment for calculation.Compared with the worst-case estimation strategy,the battery intervention time of the workload prediction strategy is reduced by an average of 23.19%.This thesis solves the problems of an energy shortage,power mismatch,and uncertain response time faced by passive systems in real-time calculations expands the application scope of passive systems,and provides new possibilities for the interconnection of everything.