Design of a 2D/3D Positioning System and Its Real-Time Application With Low-Cost Sensors

In this article, we introduce a positioning system developed for two- and three-dimensional motion tracking. The system is based on a recursive Bayesian estimator with a dynamic naive Bayesian classifier map matching scheme. The states of the dynamic naive Bayesian classifier are created by using the map information and partitioning the region of interest into grids. The developed positioning system considers three types of measurements of the platform at each time instant: the heading measurement to determine the prior probability distribution; the single-anchor distance and altitude measurements to determine the observation likelihood. A recursive Bayesian estimator takes advantage of these measurements to obtain the posterior probability distribution. Ultimately, via the obtained posterior probability distribution, the most probable projection of these measurements onto the states of the dynamic naive Bayesian classifier is estimated as the current position of the platform. To avoid the potential ambiguities in the estimation process, the estimator exploits a design parameter that characterizes the platform’s maximum attainable speed. Simulations and real-time application results are given to illustrate the effectiveness of the developed system for positioning applications in two- and three-dimensional indoor and outdoor environments with constraints, such as corridors, roads, or flight paths.