Pranav Thakkar

Hi! I am a research engineer at Honda R&D, where I work on robot perception.

Previously, I graduated from the Department of Aerospace Engineering at IIT Bombay, where I was advised by Prof. Leena Vachhani and Prof. Hemendra Arya. I am fortunate to have worked at the Autonomous Robots & Multi-robot Systems Lab at IIT Bombay, and at the Aerospace Systems Lab at the University of Texas at Arlington.

Email  /  CV  /  Github  /  LinkedIn

profile photo
Research

My research interests lie in robust perception and safe autonomy. I currently work on developing real-time multi-sensor odometry and mapping algorithms, tailored towards their respective use case. With regards to safety, I want to use formal methods and concepts from nonlinear system theory to make learned models used for robot vision and navigation, more reliable in their behaviour.

Unobservable Spaces for Bearing-Only Localization
Pranav N. Thakkar, Prashant V. Patil, Leena Vachhani
American Control Conference, 2021

We link geometric intuition with nonlinear observability analysis while discussing joint state & parameter localization problems for robots with access to only bearing measurements.

An India-specific Compartmental Model for Covid-19: Projections and Intervention Strategies by Incorporating Geographical, Infrastructural and Response
Heterogeneity

Sanit Gupta, Sahil Shah, Sumit Chaturvedi, Pranav Thakkar, Parvinder Solanki, Soham Dibyachintan, Sandeepan Roy, M. B. Sushma, Adwait Godbole, Noufal Jaseem, Pradumn Kumar, Sucheta Ravikanti, Aritra Das, Giridhara R. Babu, Tarun Bhatnagar, Avijit Maji, Mithun K. Mitra, Sai Vinjanampathy
arXiv, 2020

A concerted effort to model the spread of COVID-19 under government intervention, in a step towards data-driven public policy in India.

Bearing-only Localization in Uncertain Environments
Pranav Thakkar,
Master's Thesis, 2022

A compilation of my work over two years at the ARMS Lab, ranging from observability analysis to robust estimation for bearing-only robots.

State Estimation for Vision-based Localization under Uncertain Conditions
Prashant V. Patil*, Pranav Thakkar*, Leena Vachhani
arXiv, 2019

We show that a class of tunable nonlinear Kalman filters work well towards robust state estimation without having to re-estimate landmark locations.


* denotes equal contribution from authors.
Optimal Landmark Selection for Bearing-Only Navigation
Pranav Thakkar, Leena Vachhani
Advances in Robotics, 2019

A comparative analysis of heuristics-based algorithms that select a subset of landmarks for robot navigation with access to only bearing measurements.


Website template borrowed from Jon Barron.