Projects

  1. Reinforcement Learning using Neural Networks

    The goal of this project is to design a reinforcement learning algorithm using Q-learning with function approximation. A Multi-layer neural network is used as the Q function approximator. The environment which is trained is taken from the OpenAI Gym package. The Q-learning algorithm is then compared with other algorithm(SARSA)



  2. Checkpoint-based Fault Mitigation Techniques for GPUs

    Checkpoint mechanism has been very successful in achieving fault tolerance for many scientific applications. Conducted survey on the existing fault tolerant algorithms for CPU-GPU and proposed a new fault tolerant algorithm to aid Check-pointing which attempts to reduce the communication overhead between CPU and GPU



  3. High Performance Cache Simulation on GPU

    Cache simulation plays a major role in academia and industry. Various techniques have been proposed to reduce the space that store memory reference and improve the performance of sequential trace-driven simulation but couldn’t improve the computation time. Therefore, it is desirable to use some parallel computation for cache simulation in order to reduce the simulation time. The goal of the project is to develop a CPU-GPU based cache simulator and compare it with traditional CPU-only simulation. Due to high parallelism of GPU, the simulation time was found less in CPU-GPU compared to CPU-only device.



  4. The Effect of Mobility on IPTV (VOD) over Wi-max using OPNET

    In this project, the effect of Mobility of Mobile Wimax Subscribers on Video-on-Demand (VOD) over Wi-max is measured. OPNET simulator is used to analyse the performance of Internet Protocol Television (IPTV) over WiMAX as the software provides a flexible environment to model the communications networks with high accuracy. Different scenarios are considered with different mobility



  5. Undergraduate Thesis: Development of Embedded System based Device for Detection of Fluorine in Water

    This projects attempts to design and develop simple, low cost and easy to use device to determine the amount of fluoride in drinking water



  6. IBM Hackathon (India) - Internet of Things - Smart Traffic Control System

    Developed a device which changes the traffic signals according to the traffic density at a given place and sends text message or alert to mobile phone about the flow of traffic