Paola Ardon Ramirez
Human-robot collaborations are becoming prevalent. The International Federation of Robotics (IFR) reports that in 2016 Europe increased their sales for personal/domestic robots considerably by 29% and have a chair of 4% of the world market (world market of approximately 42.9 million units per year). Furthermore, they are projecting an increase of 30-35% growth per year until 2020 for domestic/household and 20-25% growth per year for entertainment robotics, which will create opportunities for human-robot collaboration in a home environment (e.g. cooking together, cleaning, online shopping) as well as elsewhere. In order to achieve these applications autonomy in humanoid robots is essential, specially for small tasks such as grasping. Hence it is of research interest to improve grasp affordances, specially in dynamic environments.
As a long term goal this research project seeks to answer two questions: How to identify the good grasping points on a target object? And once the environment context has been inferred, how to repeat the action on new objects with a high probability of success?