Knowledge Bank

This thesis introduces a real-time location system (RTLS) employing a hidden Markov model (HMM) map-matching scheme. In the hidden Markov model, hidden states are generated within the grids of the area of interest. At each time interval, two types of data are collected: the distance between a single anchor and the object, and the heading angle of the object. The object's heading angle is employed to compute the prior distribution of the object's next position based on its past position. The distance between the object and a single anchor is employed to calculate the observation likelihood of the current position using a normal distribution. By utilizing two probability matrices, the maximum likelihood of the current position of the object in the hidden Markov model state can be generated. Consequently, recursive Bayesian filters can leverage these estimated positions to create a map for the RTLS.