Punarjay Chakravarty

About

Punarjay Chakravarty is a robotics and autonomous systems researcher whose work spans trajectory forecasting, multi-agent behavior prediction, mobile robotics, and reinforcement learning-based navigation. He is perhaps best known for his contributions to the Trajectron++ framework — a generative, dynamically-feasible trajectory forecasting system that models heterogeneous agent interactions in complex environments. This work, which has accumulated over 150 citations across multiple publications, has become a significant reference in the autonomous driving and human-robot interaction communities, addressing the critical challenge of predicting how humans and vehicles move through shared spaces. Earlier in his career, Chakravarty laid important groundwork in sensor fusion and surveillance robotics, developing systems that combined panoramic vision with laser range finders for multi-person tracking (45 citations) and building collaborative indoor surveillance frameworks using mobile robots and overhead cameras. His anomaly detection work demonstrated an early interest in robots that can learn and patrol environments autonomously. More recently, he has pushed the boundaries of uncertainty propagation in forecasting pipelines and curriculum-based reinforcement learning for indoor navigation. Spanning nearly two decades of research, Chakravarty's work consistently bridges perception, prediction, and safe autonomy — making him a notable contributor to the foundations of intelligent robotic systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Trajectron++: Multi-Agent Generative Trajectory Forecasting With Heterogeneous Data for Control

   86 citations · 2020
2. 2
   Trajectron++: Dynamically-Feasible Trajectory Forecasting with Heterogeneous Data

   64 citations · 2020
3. 3
   Panoramic Vision and Laser Range Finder Fusion for Multiple Person Tracking

   45 citations · 2006
4. 4
   Propagating State Uncertainty Through Trajectory Forecasting

   20 citations · 2022
5. 5
   Anomaly detection and tracking for a patrolling robot

   20 citations · 2007
6. 6
   External cameras and a mobile robot: A collaborative surveillance system

   9 citations · 2009
7. 7
   An A* Curriculum Approach to Reinforcement Learning for RGBD Indoor Robot Navigation

   6 citations · 2021
8. 8
   Person tracking, pursuit & interception by mobile robot

   4 citations · 2006
9. 9
   Trajectron++: Dynamically-Feasible Trajectory Forecasting With\n Heterogeneous Data

   3 citations · 2020

Key Collaborators

Contact & Links