Localization In Wireless Rechargeable Sensor Networks

Wireless Rechargeable Sensor Networks(WRSNs) technology is an emerging tech-nology which integrates wireless charging, sensing, communication and computation capabilities. Different from the traditional wireless sensor networks, the sensor nodes of the WRSNs are powered by the onboard energy supply and need to be charged above a threshold in order to support their functions. Such rechargeable scheme can effective-ly solve the problem of the limited life of the traditional WSNs caused by the sensor nodes powered by batteries. Therefore they are of great potentials to be widely used in warehouse inventory management, environmental monitoring and so on.In most of the applications and studies of the sensor networks, the location infor-mation of the sensor nodes is indispensable. For example, only based on the location information of the sensors, we can ascertain where the collected information is orig-inated from, thereby, we can optimize the geographic routing protocols and charging efficiency and so on. For this reason, research on the localization problem becomes the prerequisite of many applications in WRSNs.In this paper we investigate the problem wherein a WRSN is consist of a set of ran-domly distributed sensor nodes and a self-localizated charger can be utilized to locate all the nodes in the sensor network. We propose three localization methods based on the characteristics of WRSN, including charging power and charging region. Specifically, the work in this paper can be divided into three parts. First, we propose to leverage the relationship between the charging power and the distance to locate the nodes under the charging scenario without noise interference. According to the received power and the mathematic model of charging power and distance, the distance can be calculated and the coordinates of location can be obtained using mathematical methods such as trilateration. We propose a triangle covering algorithm which uses equilateral triangles to cover polygon to get the stop points of the mobile charger. To choose a reasonable length of the side, we make a study on how the required localization time is affected by the side length and prove that the time duration decreases with the length increasing. Second, we take the noise interference into consideration in the charging scenario. The noise in the charging power values will introduce deviation to the distance value and as a result the trilateration algorithm can not produce a valid solution. To solve this, we utilize the Least Squares method to suppress the result error. Third, considering the dif-ficulty to obtain the exact charging model and directional chargers are more commonly used in applications because of its more stable power output, we then propose a poly-gon division localization algorithm based on directional charging characteristics of the directional charger. The main idea behides this method is that the charging region of a directional charger is a sector rather than a disk and when the charger stops at different positions and turns to different orientations, it will charge different sensor nodes. We propose a polygon division localization algorithm based on Angle Division and Grid Division. Further, for the Angle Division algorithm, we study optimization for both the single node and multiple nodes scenarios. To evaluate our algorithms, we conduct both field experiments and large-scale simulations. The experimental results show that the algorithm proposed can locate the nodes with the error less than 20cm in an area size of 5m x 5m. The Simulations evaluate out algorithm from more aspects and the results show that our algorithm can effectively locate the sensor nodes and decrease the distance error of localization.