Ashish Kumar

Visual Perception | Neural Networks | Robotics | Autonomous Drones | Autonomous Vehicles

About Me

PhD in Autonomous Robotics and Neural Network Architectures from Indian Institute of Technology, Kanpur, India (IIT Kanpur), Supervisor Prof. Laxmidhar Behera. My research interests are Space Robotics | Autonomous Drones and Vehicles | Machine Vision | Neural Network Architectures | Machine Learning | Deep Learning. I specialize in developing visual perception algorithms | neural network architectures | GPU optimizations | Developing hardware systems including electronics. Please see CV for more details.

CV Publications Robotics

News

2024

  • Feb:   A paper on Fast-Fruit-Detector got accepted to IEEE Robotics and Automation Letters (RA-L), 2024.
  • Feb:   PiX: Dynamic Channel Sampling for ConvNets got accepted to Computer Vision and Pattern Recognition (CVPR), 2024.
  • Jan:   Jetson-SLAM is accepted to be presented at International Conference on Robotics and Automation (ICRA), 2024.

2023

  • Dec:   A paper on the autonmous fruit harvesting UAVs (DroneBee) got accepted to IEEE Robotics and Automation Letters (RA-L), 2023.
  • Dec:   Defended PhD Thesis.
  • Nov:   A paper on gravity agnostic thrust microstepping control (TMDC) got accepted to IEEE Robotics and Automation Letters (RA-L), 2023.
  • Oct:   A paper on GPU-accelerated high-speed stereo visual SLAM (Jetson-SLAM) got accepted to IEEE Robotics and Automation Letters (RA-L), 2023.
  • Jun:   Presented open seminar of my PhD research. Please watch the Teaser Video of my research.

2022

  • May:   Our team in DJI Robomaster challenge ranked 3rd worldwide.
  • Jan:   DeepMI: A Mutual Information Based Framework For Unsupervised Deep Learning of Tasks (arXived).
  • Jan:   I am mentoring undergarduates at IIT Kanpur for DJI Robomaster University AI challenge, 2022.

2021

  • May:   One paper got accepted to International Conference on Systems, Man, and Cybernetics (SMC), 2021.

2020

  • Oct:   I'll present state-of-the-art seminar of my PhD research.
  • Jun:   Our work on challenge-2 of pick-transport-place using UAVs in MBZIRC 2020 has been accepted to International Conference on Intelligent Robots and Systems (IROS).

2019

  • Jul:   I will serve as computer vision and electronics system lead for challenge-2 of pick-transport-place using UAVs in MBZIRC 2020, Abu Dhabi.
  • Jun:   One paper got accepted to International Conference on Intelligent Robots and Systems (IROS).
  • Jan:   Our vision system for Amazon Robotics Challenge 2017 has been accepted to International Conference on Robotics and Automation (ICRA), 2019.

2018

  • Aug:   I have received IROS, 2018 SDC travel award to attend IROS conference.
  • Jul:   I am joining as PhD student at IIT Kanpur in Robotics with Prof. Laxmidhar Behera.
  • Jun:   One paper on textured 3D maps of cities using autonomous vehicles is accepted in International Conference on Intelligent Robots and Systems (IROS) 2018.

2017

  • Jun:   We ranked 3rd, 4th, 5th worldwide in pick-task, stow-task and pick-cum-stow task among teams MIT, Princeton, Delft, in Amazon Robotics Challenge 2017.
  • Jan:   I will serve as computer vision and electronics system lead in team IITK-TCS for Amazon Robics Challenge 2017, Leipzig, Germany.

2016

  • Sep:   I have joined with Prof. Gaurav Pandey on Autonomous Driving Project funded by FORD autonomous driving, california, USA.
  • Aug:   I have defended my M.Tech. research which served as visual perception system in Amazon Picking Challenge 2016.
  • July:   We ranked 5th worldwide among teams MIT, Princeton, Delft, in Amazon Picking Challenge 2016.

2015

  • Dec:   I will serve as computer vision and electronics system lead in team IITK-TCS for Amazon Picking Challenge 2016, Leipzig, Germany.

2014

Publications

Please visit my Google Scholar for full list

2024

Pick-or-Mix: Dynamic Channel Sampling for ConvNets

Ashish Kumar, Daneul Kim, Jaesik Park, Laxmidhar Behera

Computer Vision and Pattern Recognition (CVPR), 2024

Paper Code
High-Speed Detector For Low-Powered Devices In Aerial Grasping

Ashish Kumar, Laxmidhar Behera

Robotics and Automation Letter (RA-L), 2024

Paper Code Video

2023

Design, Localization, Perception, and Control for GPS-Denied Autonomous Aerial Grasping and Harvesting

Ashish Kumar, Laxmidhar Behera

Robotics and Automation Letter (RA-L), 2023

Paper Code Video
Thrust Microstepping via Acceleration Feedback in Quadrotor Control for Aerial Grasping of Dynamic Payload

Ashish Kumar, Laxmidhar Behera

Robotics and Automation Letter (RA-L), 2023

Paper Code Video
High-Speed Stereo Visual SLAM for Low-Powered Computing Devices

Ashish Kumar, Jaesik Park, Laxmidhar Behera

Robotics and Automation Letter (RA-L), 2023

Paper Code Video

2022

DeepMI: A Mutual Information Based Framework For Unsupervised Deep Learning of Tasks

Ashish Kumar, Laxmidhar Behera

arXiv, 2022

Paper

2021

Shape Back-Projection In 3D Scenes

Ashish Kumar, Laxmidhar Behera, K.S. Venkatesh

arXiv, 2022

Paper
End-To-End Real-Time Visual Perception Framework for Construction Automation

Mohit Vohra, Ravi Prakash, Ashish Kumar, Laxmidhar Behera

International Conference on Systems, Man, and Cybernetics (SMC), 2021

Paper

2020

Towards Deep Learning Assisted Autonomous UAVs for Manipulation Tasks in GPS-Denied Environments

Ashish Kumar, Mohit Vohra, Ravi Prakash, Laxmidhar Behera

International Conference on Intelligent Robots and Systems (IROS), 2020

Paper Video

2019

Domain-Independent Unsupervised Detection of Grasp Regions to Grasp Novel Objects

Siddhartha Vibhu Pharswan, Mohit Vohra, Ashish Kumar, Laxmidhar Behera

International Conference on Intelligent Robots and Systems (IROS), 2019

Paper Video
Semi Supervised Deep Quick Instance Detection and Segmentation

Ashish Kumar, Laxmidhar Behera

International Conference on Robotics and Automation (ICRA), 2019

Paper Video

2018

Real Time Incremental Foveal Texture Mapping for Autonomous Vehicles

Ashish Kumar, Dr. Gaurav Pandey, James McBride

International Conference on Intelligent Robots and Systems (IROS), 2018

Paper Video

Robotics