The purpose of this bachelor thesis is to analyze the curvature estimation problem along the trajectory of a mobile robot. An increased number of mobile robots are used, the safety and stability of operations are important. A key element of safety is reliable localization and part of it may be curvature estimation, especially for line following robots.
For the development, a DaNI robot platform and the LabVIEW software development environment are used. LabVIEW allows programing the operations of the mobile robot using graphical user interfaces. Two kinds of sensors are used to estimate the curvature: the optical quadrature encoders on the wheels and a digital compass.
Since the algorithm is realized on a real mobile robot and it has to be run in real-time, it is important to understand the geometry and kinematics of the robot that links the measured quantities to the curvature of the actual path. The calculations are presented for the specific robot. Thanks to the geometric and kinematic relations, an algorithm is presented and implemented for the curvature estimation.
Estimation errors occur for multiple reasons such as uncertain parameters, measurement noise and discrete time operation. It turns out that the estimated curvature using the digital compass sensor is less reliable as the one based on optical quadrature encoder measurements.
The implemented curvature estimation VIs can be used in the mobile robotics field for line following algorithms and indoor navigation of robots. It can also be combined with other sensor information to improve the performance.