Mobile robot path planning with explicit consideration of uncertainty due to state space and action space discretisation

Konferenz: ISR 2018 - 50th International Symposium on Robotics
20.06.2018 - 21.06.2016 in München, Germany

Tagungsband: ISR 2018

Seiten: 8Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Autoren:
Demeester, Eric (KU Leuven, Campus Diepenbeek, Dept. of Mechanical Engineering, ACRO research unit, Belgium)

Inhalt:
In order to enable mobile robots to perform complex manoeuvres such as parallel parking or driving through a narrow passage, it is important that navigation algorithms take the robot’s size and kinematics carefully into account. Furthermore, it may be necessary to plan multiple paths to various goal locations in an efficient way, for example to decide how obstacles should be avoided best or to estimate driver intents in teleoperated or shared control vehicles. In this work, we adopt a dynamic programming technique on a discretised configuration space that comprises robot position and orientation. The novelty with respect to prior work is that the robot’s shape and state transition probability are adapted to deal with the chosen action and configuration space’s discretisation. Furthermore, a look-up table is used to speed up continuous-path collision checking. Preliminary experiments in a simulation environment demonstrate the potential of this approach, but also reveal that further improvements are required.