Description

The thesis consists in developing a fundamental component of a complex system for disassembling electromechanical devices with a robotic arm. It is our purpose to develop a planner that takes care of deciding the sequence of actions that the low-level controllers should execute to successfully retrieve the valuable parts of the device, given the current state percepts. This presents many difficulties that require more sophisticated techniques that those studied in deterministic planning. We want to address the problem through a probabilistic formalization that factors in stochastic outcomes.

The work is under the scope of the following projects:

• IMAGINE: Robots understanding their actions by imagining their effects (web)
• HuMoUR: Markerless 3D human motion understanding for adaptive robot behavior (web)