About me

I am a PhD candidate in the Department of Engineering Cybernetics at the Norwegian University of Science and Technology (NTNU), where I am advised by Prof. Kostas Alexis in the Autonomous Robots Lab. I am co-advised by Prof. Eleni Kelasidi and Prof. Margarita Chli. My research lies at the intersection of robot perception, state estimation, and long-term autonomy, with a focus on enabling robots to operate reliably in environments where conventional vision-based methods fail.

During my PhD, I have developed resilient perception and state-estimation methods for underwater robots, spanning refractive visual-inertial odometry, learning-based proprioception, and long-term navigation under severe visual degradation and complete visual blackout. Looking ahead, I aim to bridge resilient perception with scene reasoning and foundation models, starting by tackling the data gap, to enable autonomous systems that can understand, reason about, and interact with complex marine environments.

While underwater robotics is my primary research platform, I am broadly interested in robust perception for autonomous systems operating in extreme environments. This has also led me to work on vision-based lunar navigation during a research visit to NASA's Jet Propulsion Laboratory. My long-term vision is to develop resilient robotic systems that perceive, reason, and operate autonomously where existing robotic systems remain fundamentally limited.

Computer Vision Underwater Robotics Visual-Inertial State Estimation Camera Models Space Robotics Deep Learning

Recent

News & Updates

  • 2026

    Co-organizing and presenting at RSS 2026 workshop on “Mapping the Reef: Underwater 3D Reconstruction for Coral Ecosystems”, Sydney, Australia.

  • 2026

    BEVIO (Bird’s-Eye-View sparse-update VIO for lunar day–night navigation) accepted and presented at IEEE ICRA 2026, Vienna, Austria.

  • 2026

    RadVIO (tightly-coupled radar-visual-inertial odometry) accepted at the European Control Conference (ECC) 2026.

  • 2026

    Awarded the Qualification Grant from the Research Council of Norway to commercialize Bionic Neuromast Inertial Odometry (patent pending).

  • 2025

    Summer at NASA JPL / Caltech (JVSRP) developing vision-based odometry for lunar navigation; presented my research at the Vision AI Seminar, at NASA JPL.

  • 2025

    Received Best Contributed Talk at the IEEE ICRA 2025 Amphibious Robotics Workshop.

  • 2025

    DeepVL: Deep Velocity Learning accepted and presented at the IEEE ICRA 2025, Atlanta, USA.

What I work on

Research Highlights

Online Refractive Camera Model

A novel online refractive camera model that adapts standard computer-vision methods to unknown refractive media, enabling trans-refractive visual-inertial odometry across air and water. IROS 2024.

DeepVL — Deep Velocity Learning

Dynamics- and inertial-measurement-based deep velocity learning for long-term underwater proprioception through extended visual blackout. ICRA 2025.

OmniPlanner

Universal exploration and inspection path planning across robot morphologies — one planner spanning aerial, legged, and other platforms in large, cluttered environments.

Ariel Explores

Vision-based underwater exploration and inspection via generalist, drone-level autonomy — enabling an AUV to explore and inspect unknown scenes fully autonomously.

BEVIO

BEVIO — Lunar Day–Night Navigation

Bird’s-Eye-View, sparse-update visual-inertial odometry for fast vision-based lunar rovers, developed at NASA JPL — maintaining tracking at frame rates well below conventional baselines. ICRA 2026.

Bionic Neuromast Inertial Odometry

Bionic Neuromast Inertial Odometry

A passive, bio-inspired sensor and framework for blind underwater navigation over 300 m with errors below 1% — patent pending, backed by a Research Council of Norway grant.

Selected work

Publications

Online Refractive Camera Model

Online Refractive Camera Model Calibration in Visual-Inertial Odometry

Mohit Singh, Kostas Alexis

IEEE/RSJ IROS 2024

DeepVL

DeepVL: Dynamics and Inertial Measurements-based Deep Velocity Learning for Underwater OdometryAccepted

Mohit Singh, Kostas Alexis

IEEE ICRA 2025, Atlanta, USA

OmniPlanner

OmniPlanner: Universal Exploration and Inspection Path Planning across Robot MorphologiesUnder review

Angelos Zacharia, Mihir Dharmadhikari, Mohit Singh, Kostas Alexis

Preprint · arXiv 2026

Ariel Explores

Ariel Explores: Vision-based Underwater Exploration and Inspection via Generalist Drone-Level AutonomyAccepted

Mohit Singh, Mihir Dharmadhikari, Kostas Alexis

ICRA 2025 Workshop on Field Robotics

BEVIO

BEVIO: Efficient Bird’s-Eye-View-based Sparse-Update Visual-Inertial Odometry for Lunar Day-Night NavigationAccepted

Mohit Singh, Shehryar Khattak, Ashish Goel, Michael Paton, Kostas Alexis, Issa Nesnas

IEEE ICRA 2026, Vienna, Austria

RadVIO

RadVIO: Tightly-Coupled Radar-Visual-Inertial OdometryAccepted

Morten Nissov, Mohit Singh, Kostas Alexis

European Control Conference (ECC) 2026

Bionic Neuromast Inertial Odometry

Bionic Neuromast Inertial OdometryUnder review

Mohit Singh, Kostas Alexis

Under review at Nature Portfolio Journal, Robotics · Patent pending (EP 26387035.4)

TartanAir

pip install tartanair: A Dataset to Push the General Robot PerceptionUnder review

Wenshan Wang, Y. Hu, Y. Shaoul, Mohit Singh, Y. Qiu, Y. Zhang, S. Zhao, Sebastian Scherer

Under review at Nature Scientific Data

Full list available on Google Scholar →

Open data

Released Datasets

Hugging Face

Multi-Camera Underwater Visual-Inertial Dataset

Synchronized multi-camera and inertial data for underwater visual-inertial odometry and perception research.

Access dataset ↗

Hugging Face

DeepVL Proprioception Training Data

Dynamics and inertial measurement data used to train DeepVL for deep velocity learning and underwater proprioception.

Access dataset ↗

Invited & contributed

Talks & Presentations

Resilient vision-driven underwater autonomy· Invited

RSS 2026, Workshop "Mapping the Reef: Underwater 3D Reconstruction for Coral Ecosystems", Sydney

Advancing Underwater Autonomy with Resilient Perception

Vision AI Seminar 2025, NASA Jet Propulsion Lab, Caltech, Pasadena

Visual-Inertial Navigation for Amphibious Systems

Amphibious Robotics Workshop, ICRA 2025, Atlanta

Resilient, Exclusively Vision-Driven Underwater Autonomy · Invited

IROS 2024, Workshop on Autonomous Robotic Systems in Aquaculture, Abu Dhabi · Slides

Recognition

Awards, Grants & Patents

2025

Qualification Grant — Research Council of Norway

$50,000 to commercialize Bionic Neuromast Inertial Odometry (European Patent Application No. 26387035.4).

Patent Pending

BNIO: Bionic Neuromast Inertial Odometry

Passive, bio-inspired long-term underwater state estimation — blind navigation over 300 m with errors below 1%.

2025

Best Contributed Talk — ICRA 2025

Amphibious Robotics Workshop, for refraction-aware vision-based perception and autonomy across air and water.

2021

Finalist — IROS 2021 Best Paper (Mechanism & Design)

Multi-body simulation and flywheel stability mechanism for a spherical robot.

2019

First Place — 11th International MAV Competition, Madrid

Aerial robot framework for autonomous warehouse inventory management; competed against 14 international teams.

2021

Silver — DRDO Vision-Based Obstacle Avoidance Drone Event

9th Inter-IIT Tech Meet: graph-based exploration planner for narrow, high-clutter passages.

Mentorship

Teaching & Supervision

Master’s Student Supervision — NTNU 2024–Present

Supervising 4 MSc students: trans-refractive-media visual odometry, uncertainty-aware underwater mapping, underwater quadruped locomotion, and learned inertial proprioception for high-speed aerial robots.

IEEE Robotix Winter Workshop — IIT Kharagpur 2018

Designed a 7-day computer-vision course (feature & lane detection) for 250+ newcomers; mentored 30 students in OpenCV.

Technology Robotix Society CV Workshop — IIT Kharagpur 2018

Mentored 30+ students on stereo cameras and disparity; built a blob-following ground robot with color segmentation.

Get in touch

Let’s talk robotics.

I’m always happy to discuss underwater & field robotics, perception, collaborations, or research opportunities.

Trondheim, Norway