|A multi-agent system architecture for mobile robot navigation based on fuzzy and visual behaviours (pdf)|
In the mobile robot navigation area, topological approaches for the representation of the environment have the advantage of manipulating the information at a high level of abstraction. On the other hand, geometric representations usually need more computational effort and react slower. Furthermore, geometric models are usually more dependent on the localization system requiring a high knowledge about the position of the robot to take the necessary decisions in the navigation. This project focused on the development of a multi-agent system based on behaviours for controlling the navigation task of a mobile robot in office-like environments. The set of agents is structured in a three-layer hybrid architecture. A high level of abstraction plan is created using a topological map of the environment in the Deliberative layer. It is composed by the sequence of rooms and corridors to transverse and doors to cross in order to reach a desired room. The Execution and Monitoring layer translates the plan into a sequence of available skills in order to achieve the desired goal and monitors the execution of the plan. In the Control layer there is a set of agents that implements fuzzy and visual behaviours that run concurrently to guide the robot. Fuzzy behaviors manage the vagueness and uncertainty of the range sensor information allowing to navigate safely in the environment. Visual behaviors locate a required door to cross and fixate it indicating the appropriate direction to reach it. Artificial landmarks are placed beside the doors to show its position. The system has been implemented in a Nomad 200 mobile robot and has been validated in numerous experiments in a real office-like environment.
Reactive following of a landmark: In this video the robot finds a landmark placed beside a door labeled as 00. A set of fuzzy and visual behaviours cooperate to lead the robot to the door. In the way to the door, visual behaviours "fixate" the landmark indicating the straight direction to the door. Meanwhile, fuzzy behaviours lead the robot in the direction indicated by the camera, avoiding possible obstacles found in the way: seguimiento_de_marca.avi.
Crossing a door: (Continuation of previous video) Once the robot is near enough the door, a set of fuzzy behaviours makes the robot cross it. The robot creates a temporal model of the environment detecting the gap of the door and cross it: cruzar_puerta_habitacion.avi.
Looking for next door: (Continuation of previous video) When the door has been crossed the robot is prepared to find the following door to cross. The robot aligns itself in the corridor and check in its topological map where is the next door (at the right-side in this case). The following door has a landmark labeled as 02:alineacion_y_moverse_pasillo.avi.
Follow corridor until find following door: (Continuation of previous video) The robot follows the corridor pointing the camera towards the wall where the following door is. A set of fuzzy behaviours keeps the robot in the middle of the corridor while a visual behaviour searches for the landmark that identifies the following door (02). When the landmark has been found, the robot cross the door entering in the desired room:moverse_pasillo_y_encuentra_marca_02.avi.
The rest of the videos can be seen at :http://www.decsai.es/~salinas/videosrobotmarcas/.