Soft Computing Techniques for Minimizing and Predicting Average Localization Error in Wireless Sensor Networks

Localization methods are used to approximate the position of unknown nodes in a network. Localization errors are calculated by comparing the estimated and true positions at each time step. Finding the best network parameters to minimize localization error during the network setup process while maintaining the requisite accuracy in a short period remains a difficult task. Both unknown and anchor nodes are strategically placed to reduce localization problems, which addresses a time series issue. Soft computing approaches such as Fuzzy Logic and Adaptive Neuro-Fuzzy Inference System (ANFIS) are used to address this issue. In this study, the number of nodes and network simulation area are used as de facto parameters for Average Localization Error(ALE) training and prediction. These feature values were obtained from simulations using the modified centroid localization technique with Kalman filter. This work tries to reduce localization errors by adjusting these parameters using soft computing techniques. The experimentation is carried out in MATLAB, demonstrating the suggested method's ability to improve reliability and reduce localization errors in wireless sensor networks.