Research On Optimal Charging Strategy In Wireless Rechargeable Sensor Networks

Nodes in traditional wireless sensor networks(WSNs)are equipped with batteries with limited energy for collecting data.To small-scale WSNs,we can replace batteries for nodes periodically to extend network life,but this approach is unrealistic after the network scale increases.In recent years,as an important method for solving the problem above,wireless charging technology receives much attention.It allows a mobile charger(MC)which carries energy source to traverse the network and charge all nodes,which can finally keep the network running forever.The essential problem for charging the whole networks based on wireless charging methods is how to decide the mobile charging strategy.In other words,that is how to calculate the charging location,charging time and charging trajectory for MC.This paper designs a charging strategy for MC based on wireless charging methods.Firstly,in order to improve the charging efficiency and shorten the charging time,this paper needs to make nodes with similar physical distance and residual energy more likely be divided into one cluster.Then we propose the energy-sensing clustering algorithm and finally get the mixed center of each cluster.Secondly,according to the mixed center set,a priority-based charging trajectory design algorithm is proposed to calculate the optimal charging path of MC based on the reinforcement learning.The main work of this paper is as follows:(1)An energy-sensing clustering algorithm is proposed.In order to make full use of the one-to-multiple charging technology(that is,MC can charge multiple nodes simultaneously),this paper cluster the whole networks after measures physical distance and residual energy of nodes comprehensively.Different from the traditional clustering algorithm which only take the physical distance into consideration.The clustering idea in this paper can reduce the charging energy and charging time for every single cluster,and finally improve the charging efficiency.Energy-sensing clustering algorithm is based on k-means algorithm,and improves performance in aspect of iteration dimension,noise node processing mechanism,k-value selection strategy and mixed center definition.The simulation results show that the energy-sensing clustering algorithm performs better in terms of total charging time and charging efficiency.(2)A priority-based charging trajectory design algorithm is proposed.Firstly,the algorithm divides the cluster into different charging priorities according to the residual energy of the nodes in the cluster.Secondly,this algorithm converts the path planning problem to the reinforcement learning MDP mathematical model,and determines the reward function by minimizing the mortality of the node as the optimization goal.Finally,a movement trajectory of the mobile charger is calculated by iterating.The simulation results show that the priority-based charging trajectory algorithm performs better on the node mortality and can optimize the charging time of the mobile charger to some extent.