2D-Directional Location estimation in IoT using Fingerprint based Enhanced Kalman Filter

Abstract-In regular triangulation systems, problems of spatial alignment, whether sensor deployed things are colinear or placed in symmetry to anchors, frequently lead to drastic accuracy declines. In addition, errors are exacerbated by the presence of noise in the angle of arrival (AoA) measurements thus rendering the method inappropriate for high accuracy in applications. The integration of Kalman Filter into this process with a more specific setting of measurement noise covariance, has shown not only the promising results in reducing such estimation errors by iteratively adjusting predictions based on noisy observations in comparison to the earlier extended Kalman Filter based process but also in a low complex environment. The Enhanced filter though adds only a couple of steps in mathematical calculation, which indirectly reduces the number of iterations to reach the goal.