Jagadeswara Pavan Kumar Varma Pothuri
Robotics PhD Student · Aerial Autonomy · Robot Perception · Active Mapping · University at Buffalo
Buffalo, NY
· jpkvarmapothuri@gmail.com
· +1 716 247 3865
github.com/pjpkvarma · linkedin.com/in/jpkvpothuri · Google Scholar
github.com/pjpkvarma · linkedin.com/in/jpkvpothuri · Google Scholar
Robotics PhD student building autonomy that runs on real UAVs, from GPS-free visual tracking on a Crazyflie to multimodal active mapping in field studies. First-author Best Student Paper, AIAA AVIATION 2025. Seeking a robotics research internship.
Research areas: aerial autonomy · robot perception · active mapping · multimodal sensing · vision-based control · reinforcement learning · sim-to-real transfer · multi-robot coverage planning · field robotics · hardware-in-the-loop simulation
Current directions: energy-aware multirotor planning, world models, physics-informed learning, diffusion and VLA policies.
Education
Ph.D., Computer Science & Engineering, Robotics
2025 to Present
M.S., Robotics · GPA 4.0 / 4.0
2023 to 2025
B.Tech, Electronics & Communication Engineering · GPA 9.05 / 10
2016 to 2020
Research Experience
Graduate Research Assistant · DRONES Lab
2025 to Present
- Exploring world models and learned dynamics for prediction and planning on UAVs, with a focus on policies that transfer from simulation to real hardware.
- Early work on physics-informed models of quadrotor dynamics for data-efficient, dynamically grounded control.
Graduate Research Assistant · ADAMS Lab
2023 to 2025
- Vision-based aerial autonomy, first author, AIAA 2025: trained a camera-only reinforcement-learning velocity controller in AirSim and deployed it sim-to-real on a Crazyflie with no GPS; one UAV actively tracks another using an efficient detector for target state estimation, beating a tuned PID baseline on tracking up-time and stand-off accuracy.
- Multi-UAV coverage planning, SCoPP, JCISE 2026: co-developed a scalable coverage-path planner that partitions non-convex survey areas with no-fly zones and load-balances workload across up to 150 UAVs; field-validated on a post-flood survey.
- Open-source flight-demo platform, AIAA 2024: contributed quadrotor flight-dynamics modeling to an F450 / Pixhawk FPV platform and a matching hardware-in-the-loop AirSim / Unreal digital twin for real-to-sim capture of human flight demonstrations.
- Systems & field work: PX4 / ArduPilot on Pixhawk and Jetson; ROS 2, MAVROS; flight tests with Crazyflie and Parrot Anafi. TA, MAE 550 Optimization in Engineering Design.
Professional Experience
Software Developer · Tata Consultancy Services
2020 to 2022
- Built ML-driven IT-operations tooling: a recommendation engine that cut ticket resolution time 35% and a forecasting dashboard covering 500+ applications at 92%+ accuracy.
Publications
- P. KrisshnaKumar, J. Witter, L. Collins, J. P. K. V. Pothuri, P. Ghassemi, E. T. Esfahani, K. Dantu, S. Chowdhury. “Efficient Planning for Scalable and Load-Balanced Area Coverage by Multiple UAVs.” ASME J. Computing and Information Science in Engineering, vol. 26, no. 9, 2026. DOI
- J. P. K. V. Pothuri, A. Bhatt, P. KrisshnaKumar, M. Oddiraju, S. Chowdhury. “A Learning Approach to Efficient Vision-Based Active Tracking of a Flying Target by a UAV.” AIAA AVIATION Forum, 2025. arXiv★ Best Student Paper, AIAA Intelligent Systems Student Paper Competition, AIAA AVIATION 2025
- H. Xiao, P. KrisshnaKumar, J. P. K. V. Pothuri, P. Soni, E. Butcher, S. Chowdhury. “An Open-Source Hardware/Software Architecture and Simulation Environment for Human FPV Flight Demonstrations for UAV Autonomy.” AIAA AVIATION Forum, 2024. arXiv
Technical Skills
- Languages
- Python, C++, CUDA, MATLAB, Bash
- Learning for Control
- deep RL, imitation learning, physics-informed learning, sim-to-real transfer, domain randomization, transformers / VLMs
- Dynamics, Control & Planning
- 6-DOF rigid-body dynamics, Newton-Euler modeling, system identification, nonlinear control, MPC, PID, trajectory optimization, motion & coverage path planning, Nav2
- Perception & Estimation
- object detection, visual tracking, target state estimation, SLAM, visual-inertial odometry, sensor fusion, OpenCV
- Robotics & Hardware
- ROS 2, PX4 / ArduPilot, MAVROS / MAVLink, NVIDIA Jetson, Pixhawk, Crazyflie, Intel RealSense, LiDAR, IMU / GPS-RTK
- Simulation & Tools
- AirSim, Gazebo, MuJoCo, PyBullet, Unreal Engine, Isaac Sim, hardware-in-the-loop; PyTorch, Docker, Git, TensorRT / ONNX, SLURM / HPC, Weights & Biases
Available for research internships · CPT / OPT
jpkvarmapothuri@gmail.com