Methodologies for following mobile robots trajectories

Identifier:  TFG:364

Handle: https://hdl.handle.net/20.500.11797/TFG364

Authors:  ARBONES CLUA, MARC

Abstract:
Methodologies for following mobile robots trajectories. AUTHOR: Marc Arbones Clua SUPERVISOR: Albert Oller Pujol marbones8@gmail.com This paper is an intensive study of all types and methodologies to evaluate the acceleration, speed, and position in a robot, specifically a ground wheeled robot. There is an exhaustive research about these methods, their main characteristics and their advantages and disadvantages. For the realization of the tests I have used the educative robot Mindstorms, from the company LEGO. This versatile tool, with the material provided by the community, gives to the project a wide range of elements allowing the development of the study. There is also an explanation about the environments where you can work and use the robot. I have chosen to work with the calculus program Matlab, a data processor which allows the communication, between the robot and the computer, the data processing and presentation of their results at the same time. There is a proposal about a specific method on how to fuse the different options shown and how to obtain a solution for the specific disadvantages of every model. This method named Linear Quadratic Approximation, known as Kalman filter, works predicting future states in order to compare them with the measurements made by the sensors so, there is an adjustment to obtain the best approximation. During the work specific tests have been performed in order to check the results of the theoric bases, to analyse different cases and to observe the behavior of the robot an the sensors in all kind of situations, from more usual, rectilinear movements, to more specific cases where the robot finds an obstacle or impediment in his movement. Study Program: Degree in Industrial Electronics and Automation Engineering Defense Date: September 2015