Federation of Indian Chambers of Commerce and Industry

The Robotics Institute Summer Scholars program is a eleven-week research program that allows undergraduate students from around the world, to work on cutting-edge robotics projects, driving innovation with real-world impact. RISS is one of the top robotics research programs for undergraduates in the world. This year we had five Indian students in the program supported by the Federation of Indian Chambers of Commerce and Industry (FICCI) – Mononito Goswami, Tushar Kusnur, Shiven Mian, Viraj Parimi and Vivek Roy.

Mononito Goswami

I am working on the RoboTutor Project (https://www.cmu.edu/scs/robotutor/) under Dr. Jack Mostow. RoboTutor is an intelligent tutoring system in the form of a tablet-application. It enables children in developing countries to learn basic reading, writing and arithmetic in 18 months, with minimal human-teacher guidance. As a part of this project, I am developing a tool that would use data from field testing of the robot, and identify the most pressing issues for developers to address (we call these hotspots). Hotspots can exist in reliability of the robot, student engagement or usability among others. This tool is expected to be a vital cog in the process of data-driven iterative design, thereby improving the RoboTutor and similar automated intelligent tutors.tal cog in the process of data-driven iterative design, thereby improving the RoboTutor and similar automated intelligent tutors.


Tushar Kusnur

As part of RISS 2018, I am with Prof. Maxim Likhachev in the Search Based Planning Lab at RI, CMU. Autonomous search and exploration have tremendous applications in Robotics. I will be working on path-planning for a multi-agent coverage problem. We intend to start out with a solution for up to five agents monitoring various dynamic regions of interest. Once a satisfactory planner is functioning, we wish to add learning to the solution, wherein the planner will use its previous experiences to develop better search strategies. I find this problem interesting and challenging as it demands sound theoretical knowledge in Artificial Intelligence as well as strong development skills.


Shiven Mian

I’m working with Dr. Jack Mostow on the RoboTutor project. The project is a $1M finalist in the Global Learning XPRIZE challenge, funded by Elon Musk, amongst others. RoboTutor aims to address the extreme shortage of teachers and books in developing countries and is currently being field-tested in Tanzania. It is an Android application made for kids aged 7-9 helping them to gain literacy and proficiency in basic arithmetic without adult assistance.


Viraj Parimi

I am working on two projects with Intelligent Coordination and Logistics Laboratory. One involves the integration of the Hierarchical Bayesian model that predicts bus dwell times to the traffic signal system. The idea is to first to examine the effectiveness of the model developed in a simulation and based on the results it would be deployed to run online. The second project involves developing a new algorithm which acts as an adaptive filter for robotics sensors to improve performance during unexpected environmental changes.


Vivek Roy

I am working with Dr. Martial Hebert on Object Detection and Tracking. We plan to have a deep learning approach to the problem. There are existing works on Object Tracking but none of them handle the case of occlusion to a level to be of particular use in the real world where occlusions are very common. We plan to tackle this problem by intelligently combining the output of a Tracker and a Detector. Objects that are partially or even completely occluded are not picked up by the Detector. Frames where the Detector fails to pick up are filled in by the Tracker. Thus the tracker and detector combined help us perform Robust Occluded Object Tracking (ROOT).