Experience in Industry: 14+ yrs
Experience in Academia: 8½+ yrs
Omron Research Center of America (ORCA), San Ramon, USA, robotics.omron.com
Mobile Robotics Research and Development Lead (Sr. Researcher) (Apr2018-current)
Part of the founding team of ORCA – leading the research and development of multiple mobile robot research programs.
Designed, developed, implemented and validated algorithms for simultaneous localization and mapping for industrial mobile robots.
(a) Developed end-to-end frameworks for sparse and dense metric, as well as topological SLAM pipelines with multi-sensor loosely and tightly coupled fusion, integrating multiple 2D and 3D optical, radio sensors – 2D/3D DToF Lidar, RGB-D, IMU, UWB, GPR, encoders etc. Driven by accurate Graph Optimization, Particle Filter, EKF variants to mapping and localization.
(b) Designed techniques for post-processing- classification, human tracking, ML and inference for on-board, on-edge (SBC/NUC), on-server 3D and computer vision applications.
(c) Crafted algorithms for distributed, decentralized, collaborative SLAM, collaborative autonomous navigation and mapping (exploration), GAN based predictive mapping, collaborative path planning, POMDP task planning, collision avoidance, deep SLAM feature learning and joint/ collaborative localization.
(d) Designed, developed, implemented and benchmarked multi-robot centralized fleet-managed map updating for incorporating incremental and periodic changes to mapped environment.
(e) Developed both proprietary and ROS1/ROS2-based stacks and for Omron's mobile robots and simulation platforms. Adopted standards including including VDA 5050, MassRobotics, OpenRMF, ISO3691-4 (older EN1525, RIA15.08), ANSI/RIA 15.06, ISO13849 etc.
(f) Represented ORCA at international conferences and contributed to ORCA IP generation and tech transfer processes (authored patents, whitepapers etc.).
(g) Collaborated with Omron's Product Management to build requirements, specifications, customer voice, user stories and test cases to guide ORCA mobile robot research
PreNav Inc., San Carlos, USA, prenav.com
Computer Vision Engineer, Owner (Dense 3D Reconstruction) (Nov2015-Mar2018)
Designed and developed PreNav’s cluster computing based photogrammetry pipeline for
(a) UAV/drone localization and camera pose estimation.
(b) Generation, measurement and analysis of billion-point metric accurate ultra- high density 3D point cloud and surface reconstructions of towers, buildings, turbines and other structures from drone imagery and laser data.
Also, built PreNav’s
(c) Ultra hi-res panorama pipeline
(d) RPN/CNN learning based tower components, breakage, crevice detection module
(e) UWB localization module for autonomous drones
Vision for Robotics (V4R) - Automation and Control Institute - TU Wien, Vienna, Austria, acin.tuwien.ac.at
Researcher (Nov2009-Nov2015)
Designed RGB-D/ range/ stereo sensors based object recognition and 3D scene reconstruction algorithms targeted at
(a) landmarks for indoor robot navigation (Robots@Home IKEA EU FP6)
(b) table-top scenes for robot arm manipulation (GRASP EU FP7).
Formulated evolutionary psychophysics based
(c) k-TR Theory to Visual Perception
(d) Recognition by Component Affordances (RBCA), an analog of Vision-Language-Action (VLA) framework
(e) AfNet-The Affordance Network for cognitive visual processing (theaffordances.net)
and allied components:
(f) AfRob (for Robotics)
(g) A-fACToR (ACT-R model)
(h) Af-kTRAANS (Cognitive language).
Supervision: Ishaan Gupta (Internship on SQ Fitting), Helmut Hansy (Masters thesis on Knowledge Bases)
Teaching: WS2013/14, WS2014/15 Ausgerechnet Elektrotechnik – Foundations of Electrical Engineering
Nvidia Corp., Santa Clara, USA, nvidia.com
Mobile Vision Research Intern (July2012-Sep2012)
Designed mobile image segmentation and modeling algorithms implemented on Tegra/Android devices using C/CUDA
Qualcomm Research – Vienna Corporate R&D Center, Austria, qualcomm.com
Augmented Reality Intern (July2011-Nov2011)
Designed and implemented real-time feature detection and tracking, SLAM algorithms for QC Augmented Reality (QCAR) SDK on Android, iOS environments
UtopiaCompression Corp., Los Angeles, USA, utopiacompression.com
Technical Research Synthesis Engineer (Research Scientist) (Jun2008-Nov2009)
Designed and implemented various algorithms for Vision based Automatic Target Recognition (ATR), Image & Video Compression for Electro-Optic (EO), Infra-Red (IR) & Sonar Imagery, Generic scalable cognitive object recognition; US Department of Defense (DoD) Small Business Innovation Research (SBIR) projects. Generated revenue ~$2 million and served as technical lead on several Phase I and II projects.
Scientific Research and Development Engineer (Research Scientist) (Oct2007-May2008)
Developed and implemented various algorithms for Robot obstacle avoidance, GIS, Airborne network modules for US Army, Air Force, Navy, NASA, Homeland Security SBIR programs
Center for Systems and Software Engineering (CSSE) -USC, USA, boehmcsse.org
Software Research Intern (Summer 2007)
Developed ASP & JSP extensions to CodeCount and DiffTool code profilers, maintained CSSE Bugzilla server interface
VulcanTech Software, PA, USA, vulcantechsoftware.com
Image Processing Research Intern (Sep2004–Jun2005)
Built DSP analysis plugins for ‘Vortex’, a neuro-image processing software built on a Matlab-SPM-VMat-Python pipeline
MIC Electronics, Electronic City, Hyderabad, India, mic.co.in
Systems Trainee (Summer 2003)
Trained in network management, Printed Circuit Board (PCB) design, embedded system design using CADSTAR and CamSoft
Miscellaneous Volunteer Roles
Developed ‘Demographic Enumerator’, an application for record maintenance of the Tsunami Relief Program of CARE&HELP (2005), using Visual Basic (VB) and OLEDB linking
Created ‘Regis’, a tool for registration, management of participants, with automatic slot allocation and on-line support, for PANORAMA 03, an AECE symposium (2003), using VB
Note: Company logos are trademarks of their respective owners and do not imply endorsement. These logos have been used solely to identify affiliation for the duration listed.