INHALT & PFAD:
Experiments in Grasp Recognition through Surface Electromyography
Ort: Cartesium Rotunde
Vortragende(r): Prof. Shyamanta Hazarika Ph.D.
Multi-fingered hand prostheses control using Surface Electromyography (EMG) has appeared on the market. Nevertheless, it seems EMG based control is still rudimentary, being limited to a few hand postures or a simple proportional estimate of force. Further, control is non-intuitive, in the sense that the user is required to learn to associate muscle remnants actions to unrelated posture of the prosthesis. EMG based ‘grasp type’ classification holds promise for development of an embedded intelligent controller. I will discuss work undertaken at Tezpur University in pursuit of this goal. We have undertaken study of human grasping strategies to understand cognitive underpinning of constructing grasps with little or no effort during our everyday interaction with the physical world. Six grasp types have been identified as sufficient for evolving grasping strategies in daily living activities. With preference for 'temporal characteristics' in comparison to 'spectral characteristics' of EMG signals as a basis for a feature set, we explore EMG pattern recognition vis-à-vis the grasp types. Recognition of the six basic grasp types based on EMG signals has been achieved. Another key focus of our work is an attempt to integrate quantitative and qualitative modes of representation for grasp type recognition.
I am an Associate Professor of Computer Sc. & Engineering at Tezpur University, Tezpur, Assam, INDIA. I have an M.Tech in Robotics from IIT Kanpur and a PhD (on a Commonwealth Scholarship) from University of Leeds, England. My PhD is in the area of Qualitative Spatial Reasoning (QSR) i.e., reasoning akin to human cognition, in terms of qualitative abstractions rather than complete knowledge a priori.
My primary emphasis is in rehabilitation robotics with an aim towards Intelligent Assistive Systems. I am the Principal Investigator of an on-going DIT, Govt. of India funded Project on Design and Development of a Cost-Effective Bio-signals Controlled Prosthetic Hand. The project is researching to combine EEG and EMG signals coupled with exteroceptive and pro-prioceptive sensors to build an indegenous neuro-bionic hand.
I have more than 40 research publications in Journals and International Conferences. I have around 615 citations of my select publications that includes 459 citations of a paper on QSR in Fundamenta Informaticae published in 2001.
Since October 2009, I am on a six months visit as Substitute Professor for Cognitive Systems & NeuroInformatics, Cognitive Systems Group, University of Bremen, Germany.