Abstract:
Evolution of robotics have found its applications in labor extensive operations. The human hand serves as a highly dexterous unit capable of performing a multitude of operations such as grasping a wide variety of objects and use them in any devisable fashion. Thus it makes the best option for the development of a robotic equivalent. This Report presents a detailed account of efforts to develop a robotic hand with basic finger movement capability. Instead of relying on external trigger the project aims to use muscle generated potentials that could serve as control mechanism and allow movement in various directions of interest. The intent was to use different machine learning algorithms for classification of different finger movements and come up with one with the highest accuracy. The highest accuracy of 78% was observed during the process.
The project was divided into three segments from the beginning namely design and development of hand, signal acquisition and processing to finally micro-controller interfacing. A duration of 9 months was allocated and after extensive literature review and resources dedication a preliminary prototype was developed that could suffice sufficient goals and perform basic movements.
