The University of Birmingham (www.bham.ac.uk) is one the largest research intensive universities in the UK. The School of Computer Science is ranked 8th in the UK (REF 2014). The Intelligent Robotics Laboratory (www.cs.bham.ac.uk/go/irlab) is a leading European lab in robotics and artificial intelligence. The IRLab has expertise in robot vision (Leonardis), task planning (Hawes, Wyatt), motor learning and multi-contact control (Mistry), software architectures for robotics (Hawes, Wyatt), manipulation (Wyatt, Kopicki), and machine and robot learning (Mistry, Wyatt, Hawes, Leonardis). The IRLab will draw on its experience in building systems that perform dextrous grasping, motor learning and human robot interaction (Burbridge, Hawes, Zito, Wyatt, Kopicki). The lab has a substantial code base for mobility and manipulation, which is used across multiple projects, including several pieces of software for integration (CAST, GOLEM, GRASP) that are used beyond Birmingham. The laboratory has significant experience in working with ROS, with designing integration frameworks linked with ROS, and in leading projects involving large-scale software integration for autonomy.
UoB is a Research member of the MTC-LU-UOB-Airbus challenger Team. UoB will lead T1, and in particular will lead the work on autonomous grasping. Birmingham will also contribute to the work on learning of trajectories and force-torque patterns for tool use.
UoB has significant experience in robotics projects having coordinated three European projects and participated in several more (FP6: CoSy; FP7 CogX, GeRT, PacMan, Strands, CodyCo; H2020: Cogimon). The relevant experience we will bring to this project is our experience of grasping (GeRT, PacMan) including results from the Boris manipulation system able to grasp novel objects from one-shot learning. In addition we will bring experience in force-based manipulation from PacMan, CodyCo and Cogimon, with an emphasis on learning interactions.