AYUSH DEWAN Robotics Research Center, IIIT, Hyderabad http://researchweb.iiit.ac.in/~ayush.dewan/ EDUCATION International Institute of Information Technology,Hyderabad MS by Research in Computer Science CGPA: 7.83 Dehradun Institute of Technology Bachelors of Technology in Applied Electronics and Instrumentation Percentage Aggregate: 69.3
TECHNICAL SKILLS Languages: C,C++,Python,Assembly Platforms : Keil, Matlab, Labview, MPLABIDE, MikroC, PICC Operating System : Windows, Ubuntu Libraries: ROS, OpenCV, OpenSceneGraph Robots Worked With: Pioneer Robots(P3DX,Amigo), Parrot ArDrone,Turtlebot Sensors Worked With: Microsoft Kinect, Flea2, SICK laser, Sonar. WORK EXPERIENCE KUKA Robotics June-July,2010 Intern Pune, India · Completed 6 weeks internship. · Completed Basic and Advanced Robot Training Modules. · Developed a Labview application for TCP/IP communication between KUKA Robot and a Joystick. Robotics Research center Graduate Research Assistant · Working as a Graduate Research Assistant.
August,2011-Present IIIT Hyderabad, India
IIIT August,2013-Present Teaching Assistant Hyderabad, India · TA for Mobile Robotics Course(CSE-483) · My responsibility is to evaluate assignments, projects and conduct ROS tutorial classes PUBLICATIONS Suryansh Kumar, Ayush Dewan, K Madhava Krishna ”A Bayes Filter based Adaptive Floor Segmentation with Homography and Appearance Cues”(ICVGIP-12) Ayush Dewan, Aravindh Mahendran, Nikhil Soni, K Madhava Krishna” Optimization Based Coordinated UGV-MAV Exploration for 2D Augmented Mapping” (Extended Abstract,AAMAS-2013). Ayush Dewan, Aravindh Mahendran, Nikhil Soni, K Madhava Krishna ”Heterogeneous UGV-MAV Exploration Using Integer Programming”(IROS-2013) Ayush Dewan, K Madhava Krishna ”Distributed Asynchronous Multi Robot Exploration Using MaxSum”(ICRA-2014, Under Review). GRADUATE COURSE WORK Computer Vison (CSE5765) Artifical Neural Network (CSE4772)
Mobile Robotics (CSE483) Linear Control System (ECE451) Intro to Robotics: Mechanics and Control (ECE452) RESEARCH PROJECTS Vision based exploration Abstract: At present I am working on monocular camera based exploration system for ground robots and micro aerial vehicles. I use PTAM for interacting with environment. The objective is to densify sparse map pointcloud and improve the efficiency of PTAM by introducing a planning algorithm to dictate camera trajectory. The challenge to overcome is twofold, PTAM only provides sparse map information which in insufficient for exploration, thus some other clues like planar information are needed to generate a denser map. Secondly, the efficiency of PTAM depends on movement of camera and is very prone to rapid rotational motion, thus a planning algorithm is required to dictate trajectory of camera so that algorithm works efficiently for a large scale environment. Status: In Progress. Utilities: ROS, C++, Matlab Coordinated UGV-MAV Exploration for 2D Augmented Mapping Abstract:The objective of this project was to explore an unknown area using heterogeneous agents(UGV and MAV) and build a 2D augmented map. The laser sensor on ground robot builds 2D ground map and using monocular camera on MAV a pointcloud for elevated area is generated. The exploration strategy was formulated an Integer Programming problem. Using extensive simulation it was the shown that the proposed method performs better than previous method in multiple scenarios. This work was presented in IROS,2013. Status:Completed. Utilities:ROS, OpenCV, OpenSceneGraph,PCL,C++ Distributed and Asynchronous multi robot exploration using Max-Sum Abstract:The objective of this project was to exploit the distributive and asynchronous nature of a multi robot system for exploration. The problem of task/goal assignment is represented as factor graphs and is solved using Max-Sum algorithm. This work is under review for ICRA,2014. Status:Completed. Utilities:ROS, OpenCV, OpenSceneGraph,PCL,C++ Floor segmentation using monocular camera Abstract:The objective of this project is segment out the from a sequence of camera. Using homography and optical flow, set of sparse features lying on floor are tracked and using color segmentation image is divided into multiple segments. The tracked features are then used for selecting segments lying on floor. This work was published in ICVGIP,2012 Status:Completed. Utilities:OpenCV,C++ COURSE PROJECTS Velocity Obstacle EKF Localization FASTSLAM Image Based Search Building a Panorama 4DOF Robotic Arm For details please visit http://researchweb.iiit.ac.in/~ayush.dewan/projects
ACHIEVEMENTS AND ACTIVITIES Robotics Operating System(ROS) Contributor with high ranking w.r.t Karma(2nd in India). Received a travel grant from conference committee to attend IROS,2013. Reviewed a paper for IROS,2013. Member of team winning National level Micromouse Competition at NIT, Trichy. As an IEEE student member helped in organizing different technical competitions for college Techfest. Winner of online FIFA Gaming Competition.