Application of weighted centroid algorithm based on weight correction in node localization of wireless sensor networks.

Node positioning accuracy in wireless sensor networks (WSNs) directly affects the reliability of monitoring data. As the core technology of WSNs, node positioning technology is related to its normal operation and data monitoring performance. Traditional centre-of-mass algorithms are susceptible to RSSI ranging errors and anchor node distribution due to their reliance on simple geometric calculations, resulting in large positioning errors. In addition, high-precision algorithms are often accompanied by high energy consumption, making it difficult to balance accuracy and energy efficiency. Therefore, how to improve positioning accuracy and reduce energy consumption in complex environments becomes a key challenge. To improve the positioning accuracy and extend the network application range, the study proposes a weighted centre-of-mass algorithm based on weights correction, which is optimally designed from three aspects to improve the performance. Firstly, Gaussian-constant filtering is used to denoise the Received Signal Strength Indicator (RSSI) values to reduce the interference of environmental multi-path effects. Secondly, the anchor node density and communication range are considered to correct the weight factor. Finally, the chimpanzee optimization algorithm is introduced to implement the design of the centre-of-mass node localization algorithm, and the iterative idea is used to achieve the selection of the optimal value for the unknown node position. The results show that the improved weighted Gaussian filter can reduce the ranging error of the weighted centre-of-mass algorithm, and its minimum distance error is less than 1.0 m. The improved positioning algorithm's positioning error is better than that of the traditional centre-of-mass algorithm and the weighted centre-of-mass algorithm, and its minimum average positioning error is 0.15 m and 0.14 m under different communication radii and anchor node ratios, respectively. In addition, the average positioning error and the normalized average positioning error of the improved positioning algorithm are lower than those of the RSSI-PSO algorithm, RSSI-SSA algorithm and RSSI-Trilateration algorithm for different total node numbers, with the minimum values of 0.120 m and 0.152 m, respectively. The improved positioning algorithm has a positioning accuracy of more than 90% for different node coverages, and the difference in energy consumption between the improved positioning algorithm and the other algorithms is at least 90%. The difference in energy consumption between the comparative algorithms is at least 3 img/s and 1.5 J/img, indicating good data loading performance. This approach not only expands the theoretical framework of WSN positioning technology by addressing the vulnerabilities associated with RSSI, such as its susceptibility to interference and the rigidity of weight assignments, but also enhances the precision of center-of-mass positioning. Consequently, it offers a novel strategy for achieving low-power, high-accuracy positioning solutions.