Ph.D. received on: 8/7/1997

E-mail: venditt@labrob.ing.uniroma1.it

Tutor: Prof. Giovanni Ulivi, University “Roma Tre”
 
 

 ___________________________________________________________________________________________________________

Exploration and planning algorithms for wheeled mobile robots   ___________________________________________________________________________________________________________

Advisor:

Prof. Giovanni Ulivi, University “Roma Tre”

Referees:

Prof. Giovanni Celentano, University of Napoli “Federico II”
Prof. Antonio Gallo, University of Catania
Prof. Paolo Caravani, University of  L’Aquila

Summary:

The research work carried out during the Ph.D. has led to the realisation of a complete autonomous navigation system for the mobile robot Nomad 200. This robot has the kinematic structure of a unicycle and is equipped with ultrasonic range finders. Facing the problem of realising a complete navigation system implies the study of planning and control problems. In particular, the thesis considers the problem of navigation in a completely a priori unknown environment. Several exploration and planning methods have been analysed and proposed in the thesis. The most significant navigation paradigm can be summarised as follows. The approach is based on the alternate execution of two fundamental processes: map building and navigation. In the former, range measures are collected through the robot exteroceptive sensors and processed in order to build a local representation of the surrounding area. This representation is then integrated in the global map so far reconstructed by filtering out insufficient or conflicting information. In the navigation phase, an A*-based planner generates a local path from the current robot position to the goal, that is safe inside the visited area and proposes directions for further exploration. The robot follows the path up to the boundary of the visited area, terminating its motion if unexpected obstacles are encountered. The most peculiar aspects of the method are (i) the use of fuzzy logic to build an environment map that is very efficiently computed and modified, and (ii) the iterative application of A*, that is a complete planning algorithm taking full advantage of local information. The holonomic path returned by the A*-based planner is approximated, during motion execution, by locally optimal feasible paths. The “nonholonomic” distance to the obstacles is computed using a new efficient geometric algorithm. The Experimental results of the implementation on a NOMAD 200 mobile robot show that the proposed method provides real-time performance both in static and moderately dynamic environments.
 

 _______________________________________