ARGO: An extended jason architecture that facilitates embedded robotic agents programming

Abstract

This paper presents ARGO, a customized Jason architecture for programming embedded robotic agents using the Javino middleware and perception filters. Jason is a well known agent-oriented programming language that relies on the Belief-Desire-Intention model and implements an AgentSpeak interpreter in Java. Javino is a middleware that enables automated design of embedded agents using Jason and it is aimed to be used in the robotics domain. However, when the number of perceptions increases, it may occur a bottleneck in the agent’s reasoning cycle since an event is generated for each single perception processed. A possible solution to this problem is to apply perception filters, that reduce the processing cost. Consequently, it is expected that the agent may deliberate within a specific time limit. In order to evaluate ARGO’s performance, we present some experiments using a ground vehicle platform in a real-time collision scenario. We show that in certain cases the use of perception filters is able to prevent collisions effectively.