This project is designed and developed in CVPR Lab, IIT Ropar, India.

Guide : Dr. Subrahmanyam Murala

Sponsored By : Yamaha Labs, India.

In automatomibe industry, engine assembly is considered as one of the crucial part. During engine assenbly various parts assembled together to complate desired two stroke engine. Piston is a part which responsible for the production of stroke which prodce mechanical energy for acceleration. Piston contains various rings depending on different types of engines. During engine assembly, piston without rings can leads to the faulty engine. Due to this it is mandetory to take care of piston rings during engine assembly. Our project aims the real time detection of the rings in piston. Our objective is to develope a small device which real time detect piston rings. We consider the fact that the cost of our device should be less than 20,000 INR.

Reinforcement Learning (RL) is a subfield of Machine Learning where an agent learns by interacting with its environment, observing the results of these interactions and receiving a reward (positive or negative) accordingly. This way of learning mimics the fundamental way in which we humans (and animals alike) learn. Currently, Machines are not as intelligent as humans in terms of knowledge extraction and new discoveries with using previous learned knowledge. Reinforcement Training is the solution for training of Robots in real world like we humans does right from our childhood. In this project I used Reinforcement approach (Q Learning) to train the Robot so that it should follow line precisely without being explicitely programmed for it.

NVIDIA® DGX Station™ is the world’s fastest workstation for leading-edge AI development. This fully integrated and optimized system enables your team to get started faster and effortlessly experiment with the power of a data center in your office. For more info visit AI Workstation for Data Science Teams .Following article is about accessing NVIDIA DGX Station remotely from MAC with simple ssh and X11.

The backpropagation algorithm was originally introduced in the 1970s, but its importance wasn’t fully appreciated until a famous paper in 1986 by David Rumelhart, Geoffrey Hinton, and Ronald Williams. The paper describes several neural networks where backpropagation works far faster than earlier approaches to learning, making it possible to use neural nets to solve problems which had previously been insoluble. Today, the backpropagation algorithm is the workhorse of learning in neural networks. Although Backpropagation is the widely used and most successful algorithm for the training of a neural network of all time, there are several factors which affect the Error-Backpropagation training algorithms.