Martí-Franquès COFUND Fellowship Programme


Details

Login to the application
Reference:

2018MFP-COFUND-30

Area:

Engineering and Architecture

Supervisor name and surname:

Domenec Puig Valls

Supervisor email:

domenec.puig@urv.cat

Supervisor short biography

PhD programme:

Computer Science and Mathematics of Security

Title of the research project:

Autonomous vision-based system for assistive robots in a domestic environment

Description of the research project:

Assistive personal robots will become more useful in our lives by interacting with and assisting human beings in tasks of everyday life, such as housekeeping and butler services. Although many commercial products for home care assistance have been launched, such as Moley robot, their abilities are very limited due to being fully targeted at one specific task. Thus, research on multitask robots has recently seen much progress. These robots operating in domestic environments (e.g., home settings) should be able, for example, to pick up and safely transport objects from one place to another, to draw, or care. To achieve such tasks, the robot has to be able to identify and to recognise, as well as to grasp and manipulate different types of objects, in addition to cope with many challenges.

A first challenge is the nature of indoor environments in which robots can operate and navigate through avoiding unexpected objects inside these environments.This project will devise an obstacle avoidance method specially for wheeled mobile robots in unknown environments based on a consolidation of supervised machine learning (SML) and adaptive neuro-fuzzy inference systems (ANFIS).

A second challenge is the ability to analyse and identify risks related to the targets that are based on their shape, size, material, temperature and position. Finally, the grasp planning in general includes the determination of finger contact points on the object and the choice of an appropriate gripper configuration. Consequently, this project attempts to develop a novel mature vision-based grasping model for a mobile robot assistant within a kitchen-based experimental laboratory that robot is provided with only single colour camera and single intelligent arm.

This model emulates fine motor skills (i.e., skills and activities of a child using his hands) of early childhood learners that have two separate systems responsible for catching a stationary object: one for determining how the robot is approaching the target (i.e., approaching system) and another for how it is grasping it (i.e., grasping system).

Gross anual salary:

26443.80 €

Dedication:

Full time

Working hours:

37.5 hours a week

Expected start date:

European union This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 713679