ETD

Archivio digitale delle tesi discusse presso l'Università di Pisa

Tesi etd-11192018-164145


Tipo di tesi
Tesi di laurea magistrale
Autore
BONAMINI, LUCA
URN
etd-11192018-164145
Titolo
Toward a shared autonomy control framework: application to Ego robot navigation
Dipartimento
INGEGNERIA DELL'INFORMAZIONE
Corso di studi
INGEGNERIA ROBOTICA E DELL'AUTOMAZIONE
Relatori
relatore Prof. Bicchi, Antonio
relatore Prof.ssa Pallottino, Lucia
Parole chiave
  • shared
  • autonomy
  • mobile
  • robot
  • teleoperation
Data inizio appello
10/12/2018
Consultabilità
Non consultabile
Data di rilascio
10/12/2088
Riassunto
A goal of autonomous robotic systems is to execute tedious or dangerous tasks, increase productivity, decrease production costs and preserve the environment. However, state-of-the-art systems are only able to solve a limited number of tasks in clearly defined environments robustly. These robots cannot handle complex tasks or unexpected events reliably and have problems adjusting to changes in an unstructered environment.

Robots are skilled at data acquisition, data storage, navigation and manipulation. They are able to execute tedious work which would bore a human quickly with a steady performance. On the other hand, many tasks that are highly challenging for robots such as perception, situation awareness and intelligent decision making, can be easily handled by humans. Therefore, it seems to be a very promising approach to combine both the skills of a robot and that of a human in order to overcome restrictions which arise from a system design that purely focuses on autonomy. The combination of human and machine is able to incorporate the strenghts of both sides and therefore increase the robustness of a system, decrease the development costs and offer the possibility that one can learn from the other. This idea is often referred to as Shared Autonomy, or Human-Robot Cooperation. As human input is cost-intensive, one optimization goal is to maximize the level of autonomy. Another goal is to obtain a system which is still reliable and robust. Furthermore, it is important to create an interface between human and robot which allows an efficient cooperation.

This thesis aims to improve the functionality and reliability of a soft dual-arm mobile platform, EGO-1, developed to operate in different environments and equipped with state of the art technologies for safe human-robot interaction, robustness and versatility. The main goals are to develop and implement algorithms, firstly, in order to make the robot able to create a map of an unkown scenario for autonomous navigation, and secondly, to improve the robot capabilities while it is teleoperated by a human pilot, so a contribution to shared autonomy is given.
File