High Level Motion Planning for a Robot

Detta är en Kandidat-uppsats från KTH/Skolan för elektroteknik och datavetenskap (EECS)

Författare: Axel Kelsey Lundqvist; Fabian Tidner; [2019]

Nyckelord: ;

Sammanfattning: Robots and automated systems are becoming an ever present part of modern life. The simplicity with which a technology can be used by people without a technological background is paramount to its adoption by the general public. What is presented here is a simple framework for users to give high level motion control commands without sacrificing technological complexity. This is accomplished by using a control synthesis for high level phrases. A user enters high level commands expressed in Linear Temporal Logic which the system interprets and then a robot goes on to perform. As such the entirety of the logical inputs needed for completed motion planning is encompassed within one simple phrase. The problem has been approached two different ways, one that prioritizes computational speed and another with a higher degree of optimization. Both of these framework works well for a multitude of different rooms and commands and form a basis for further works in the subject. What sets this research apart is the inclusion of Büchi minimisation based on the research by [1] in a high level motion planning system. This notably reduces computation time for more complicated LTL phrases. [1]

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)