Edgardo Ferretti

Integrating vision-based motion planning with defeasible decision making for the Khepera robot

Intelligent mobile robots need the ability to combine autonomous navigation with high-level reasoning. This paper integrates a defeasible decision making technique and vision-based motion planning. The problem faced comprises a scenario where a Khepera 2 robot has to perform removal tasks of different size objects. This problem can be conceptually divided in two stages: (a) to decide the objects transportation order, and (b) to plan and control the robot motion. The decision making task is carried out using the facilities provided by the cognitive layer of the Khe-DeLP framework, while the motion planning is performed using a smooth shortest path control generated from visual information.