Booth Id:
ROBO042T
Category:
Year:
2016
Finalist Names:
Islam Rafi, MD
Shithil, Shaekh
Abstract:
Programming robots can be slow and challenging. In trying to build a gesture based robotic manipulation , i asked myself whether it would not be possible to design an application that robot could watch a 2d video and learn the basic concepts from it. If successful , it should be of great assistance to those struggling with existing graphic and text based programming languages. Here i come with my idea, how it will be? If people can teach their robots by showing them videos from world wide web . As example , something like "youtube" video. With this in mind , I embarked on some research , with the aim of designing a tool that could help the robot understand basic tasks by watching videos. To prove that it was possible to develop such a functionality , I decided to limit design and testing to a prototype robot called "ANSME" and use KNN (K Nearest Neighbors) based novel algorithm for acquiring and comparing datas from the training videos .