Target localization and tracking based on improved Bayesian enhanced least-squares algorithm in wireless sensor networks

Classical tracking algorithms, such as the Bayesian algorithm, extended Kalman filter (EKF), and classical least-square (CLS) algorithm, have been extensively implemented at target localization and tracking in wireless sensor networks (WSNs). In this paper, an enhanced least-square algorithm based on improved Bayesian was developed for moving target localization and tracking in WSNs. We apply an improved Bayesian algorithm to obtain a set of sub-range probability based on target predictive location, and forming a range joint probability matrix. The range joint probability matrix is only automatically updated when the WSN testbed is in a dormant state. Then, the weight of every measurement is calculated and normalized based on the range probability matrix. Finally, the correction value of the target prediction position is calculated according to the weighted least-square algorithm. The experimental results show that compared with EKF, the weighted K-nearest neighbor algorithm (WKNN), the position Kalman filter (PKF), the fingerprint Kalman filter (FKF), variational Bayesian adaptive Kalman filtering (VBAKF), dual-factor enhanced VBAKF (EVBAKF), and variational Bayes expectation maximization (VBEM) algorithms, the proposed algorithm improves the positioning accuracy by 35%, 32%, 18%, and 13%, 9%, 6%, and 0.4% respectively. In addition, the proposed algorithm reduces the computational burden by more than 80 percent compared with the Bayesian algorithm.