AI-APP-MidtermPaper

AI-APP-MidtermPaper - AN OVERVIEW ON BEHAVIOUR-BASED...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
AN OVERVIEW ON BEHAVIOUR-BASED METHODS FOR AUV CONTROL M. Carreras ψ , J. Batlle ψ , P. Ridao ψ and G.N. Roberts # ψ Computer Vision and Robotics Group, University of Girona Edifici Politècnica II, Campus Montilivi, 17071 Girona, Spain {marcc, jbatlle, pere}@eia.udg.es # Mechatronics Research Centre, University of Wales College, Newport Allt-yr-yn Campus, PO Box 180, Newport, South Wales, NP20 5XR, UK g.roberts@newport.ac.uk Abstract: Traditionally, deliberative reasoning methods based on a world model have been used to guide Autonomous Vehicles. Nevertheless, due to serious problems when dealing with complex and changing environments the alternative approach of reactive architectures appeared, directly coupling perception and action without any world representation. This paper describes the characteristics and variants of reactive architectures as well as four of the most suitable ones to guide Autonomous Underwater Vehicles: Schema-based approach, Subsumption, Process Description Language and Action Selection Dynamics. Each architecture is described and evaluated using an underwater simulator with an AUV model. Copyright 2000 IFAC Keywords: Autonomous Vehicles, Artificial Intelligence, Control Systems, Marine Systems, Robot navigation, Vehicle Simulation. 1. INTRODUCTION A great deal of work concerning Autonomous Underwater Vehicles has been carried out in recent years, mainly when human supervision is not strictly necessary. In this context, one of the most important subjects for developing the autonomy of the vehicle is the design of the control architecture. There are three different kinds of architectures. The first is deliberative architectures, which are based on planning using a world model. A mission is specified to achieve a set of goals, and each goal is executed by a control system. These architectures are suitable for structured and highly predictable environments. Nevertheless in complex, non-structured and changing environments, like the majority of underwater environments, serious problems appeare while trying to maintain, in real-time, an accurate world model. The second, the reactive architectures, appeared in response to the difficulty of the deliberative architectures to deal with non-structured and dynamic environments. As Arkin said, “Reactive control is a technique for tightly coupling perception and action, typically in the context of motor behaviours, to produce timely robotic response in dynamic and unstructured worlds” (Arkin, 1998). This technique tries to avoid any kind of representation of the world. Basically, a group of parallel behaviours act independently generating outputs to the actuators according to the value of the sensors. For this reason they are also called Behavioural architectures. Reactivity, real-time response and easy implementation are the most relevant advantages. The disadvantages are the difficulty in designing a set of behaviours in order to plan and to specify high-level goals. Finally, there are
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 6

AI-APP-MidtermPaper - AN OVERVIEW ON BEHAVIOUR-BASED...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online