PhD in Robot Gripping and Hands

About the project

About the Project Traditional robot grippers have tended to be mechanically simplistic with limited dexterity and sensing ability. However, whilst their mechanical robustness has made them well suited to industrial use, they tend to only be capable of grasping a small range of specific products and if the product to be grasped is changed or unknown the gripper will likely need to be changed. Dexterous grippers, often based on the human hand, are able to grasp a much broader range of objects but selecting an appropriate grasp is a challenge, particularly when the object to be grasped is not known in advance. The approach often adopted is to use vision systems to determine the best grasp, however, there is only so much a vision system alone can tell us e.g. an objects weight or its geometry if it is partially obscured. As human we are easily able to rapidly pick up a range of unknown objects presented to us, even if they are partially hidden or overlapped by other objects. We do this using multimodal sensor information from our eyes and sense of touch but also from our experience of grasping many objects previously. The aim of this research is to develop a low-cost intelligent dexterous gripper which is able to grasp unknown object autonomously without the need for any input from a human, or prior knowledge of the object. The project will involve mechanical design, programming and AI. This project can be tailored to the student’s particular area of interest. Applicants should have a first-class degree in an engineering subject. An undergraduate or MSc in robotics or a closely related subject would be desirable but is not essential. This PhD position is open to all students who are self-funded. For further details please contact Dr Steve Davis ( s.davis.2@bham.ac.uk ).

Research areas