Kalman Filter based position estimation using "optical mouse movement sensor" and differential drive robot model

Abstract

This article describes the creation of a mathematical model for the kinematics, dynamics and electronics of a two-wheel-steered robot. As a result, it is possible to use a previously created, potential field-based and fuzzy navigation-based robot control system ([1], [2], [3], [4]) with two-wheel-driven robots as well. Using the results presented in this article the current location and the driven path can be estimated more accurately. This can be achieved by using the Kalman Filter with the wheel encoder data and using Optical Flow-based movement measurement devices that are similar to the ones known from the optical mouse peripherals. The established equations define the bases for controlling and navigating robots in indoor environments (flat surface, no sliding). According to these, they reflect the kinematics of the ground unit, the mathematical model of the electronic motor, and also the models of the sensors installed on the robot (odometer and optical mouse movement sensor).