Solar Energy Prediction And Routing Scheduling For Solar-charging Wireless Sensor Networks

In recent years,the development of the Internet of Things has been changing with each passing day,and it has produced practical effects in application scenarios in various fields.With the rapid development of machine learning algorithms and continuous changes in photovoltaic hardware,the optical charging Io T system has more room for optimization.Therefore,how to reduce network energy consumption and extend network life through a suitable model has become an urgent problem in the field of rechargeable Internet of Things.Existing research on the Internet of Things with optical charging usually has the following deficiencies: 1.The existing optical charging prediction algorithm model is single,ignoring the factors that small sensor nodes are susceptible to environmental influences,resulting in a decrease in the accuracy of the prediction algorithm.2.The existing optical charging prediction algorithm ignores the greater impact of the hidden factor of the "prediction period" on the prediction results.If the prediction period is too long or too short,the network energy consumption will increase or die early.3.The existing optical charging Io T routing scheduling strategy has a large computational burden and cannot calculate the network topology under frequent changes in illumination.In addition,it cannot be adapted to the application scenario where some nodes are allowed to die temporarily in the light charging scenario,and a certain percentage of nodes die after the light is restored.In order to solve the above-mentioned shortcomings in the research on the Internet of Things with light energy,this paper proposes a light energy prediction model JEKM based on machine learning for a small sensor node network,combined with a dynamic prediction period selection algorithm,and different if necessary.Periodic prediction reduces computational overhead while effectively improving the accuracy of the optical charging prediction model.In addition,a charging and discharging-aware routing and scheduling strategy,FAT,which is adapted to the scenario of the light-charged Io T is proposed.Experiments show that the model in this paper can enhance the accuracy of optical charging prediction,reduce computing overhead,increase network life,and bring optimization effects for optical charging sensor networks.Among them,the model in this paper increases the network lifetime by 20%-40%.