Localization Based on Parallel Robots Kinematics as an Alternative to Trilateration

Abstract

In this work a new scheme for range-based localization is proposed. The main goal is to estimate the position of a mobile point based on distance measurements from fixed devices, called anchors, and on inertial measurements. Due to the non-linear nature of the problem, an analytic relation to compute the position starting from these measurements does not exist, and often trilateration methods are used, generally based on least-square algorithms. The proposed scheme is based on the modelling of the localization process as a parallel robot, thereby methodologies and control algorithms used in the robotic area can be exploited. In particular, a closed loop control system is designed for tracking the position of a mobile point based on range measurements from fixed anchors, and it is shown a peculiar structure of the tracking error dynamics, whose allows an intuitive gain tuning and ensures global exponential stability. Moreover, it is also shown a nice connection between tuning parameters and rate of convergence of the estimation error. Experimental results confirm the validity of the proposed localization method.