Nikhil Sobanbabu

I'm an MTS intern at Amazon FAR, working with Carlo Sferrazza, Guanya Shi, and Rocky Duan.

I received my master's in Electrical and Computer Engineering from Carnegie Mellon University in 2026, where I worked with Prof. Guanya Shi at the LeCAR Lab, with a focus on system identification for legged robots and policy adaptation for humanoid whole-body control. I also collaborated with CMU's Search-Based Planning Lab on multi-agent quadruped planning.

I received my bachelor's in Electrical Engineering from IIT Madras. During that time, I worked at USC's Dynamic Robotics and Control Laboratory (DRCL) with Prof. Quan Nguyen on prior-guided reinforcement learning for agile legged robot control.

I was also the team lead and electrical subsystem lead of Team Anveshak, IIT Madras's competitive Mars rover team, during 2022–2023.

My research statement outlines the directions I wish to pursue in the future.

Email / CV / Google Scholar / LinkedIn / GitHub / Twitter / X / Blog

Research Interests

I am interested in enabling high-DoF robots to acquire contact-rich skills and adapt them for robust real-world deployment. In particular, my interests include learning-based control (agile whole-body control and loco-manipulation) and Real2Sim2Real (reliable simulation evaluations, large-scale sim-to-real transfer, and sample-efficient policy adaptation in the wild).

News

September 2026: FADA accepted to CoRL 2026! Code and a research timeline are now available.
July 2026: RIO published at RSS 2026.
2026: Completed my Master's in ECE at Carnegie Mellon University.
September 2025: Released HDMI, learning interactive humanoid whole-body control from human videos.
August 2025: SPI-Active accepted to CoRL 2025 (Oral)🎉.
July 2025: MAPF for Quadrupeds accepted to ICAPS 2025 (Demo Track)🎉.
May 2025: Preferential OGMP accepted to IROS 2025🎉.
April 2025: ASAP accepted to RSS 2025🎉.
August 2024: Started my Masters in ECE at Carnegie Mellon University.

Publications

* Equal contribution.

	FADA: Few-Shot Domain Adaptation via Dynamics Alignment for Humanoid Control CoRL 2026 Angchen Xie, Nikhil Sobanbabu, Ishayu Shikhare, Alan Wang, Max Simchowitz, Guanya Shi Webpage / arXiv / Code / Research timeline
	RIO: Flexible Real-Time Robot I/O for Cross-Embodiment Robot Learning RSS 2026 Pablo Ortega-Kral, Eliot Xing, Arthur Bucker, Vernon Luk, Junseo Kim, Owen Kwon, Angchen Xie, Nikhil Sobanbabu, Yifu Yuan, Megan Lee, Deepam Ameria, Bhaswanth Ayapilla, Jaycie Bussell, Guanya Shi, Jonathan Francis, Jean Oh Webpage / arXiv / Proceedings / Code
	Sampling-Based System Identification with Active Exploration for Legged Robot Sim2Real Learning CoRL 2025 (Oral) (All Strong Accepts) Nikhil Sobanbabu, Guanqi He, Tairan He, Yuxiang Yang, Guanya Shi [Webpage\|Arxiv\|Code]
	Preferenced Oracle Guided Multi-mode Policies for Dynamic Bipedal Loco-Manipulation IROS 2025 Prashanth Ravichandar, Lokesh Krishna, Nikhil Sobanbabu, Quan Nguyen [Website \|Arxiv \|Video \|Code]
	Towards Unstructured MAPF: Multi-Quadruped MAPF Demo ICAPS 2025 (Demo Track) Rishi Veerapaneni, Nikhil Sobanbabu, Guanya Shi, Jiaoyang Li, Maxim Likhachev [Extended Abstract \|Video]
	ASAP: Aligning Simulation and Real-World Physics for Learning Agile Humanoid Whole-Body Skills RSS 2025 Tairan He, Jiawei Gao, Wenli Xiao, Yuanhang Zhang, Zi Wang, Jiashun Wang, Zhengyi Luo, Guanqi He, Nikhil Sobanbabu, Chaoyi Pan, Zeji Yi, Guannan Qu, Kris Kitani, Jessica Hodgins, Linxi "Jim" Fan, Yuke Zhu, Changliu Liu, Guanya Shi [Webpage\|Arxiv\|Video\|Code]

Pre-prints

	HDMI: Learning Interactive Humanoid Whole-Body Control from Human Videos Haoyang Weng, Yitang Li, Nikhil Sobanbabu, Zihan Wang, Zhengyi Luo, Tairan He, Deva Ramanan, Guanya Shi [Webpage\|Arxiv\|Video\|Code]
	OGMP: Oracle Guided Multi-mode Policies for Agile and Versatile Robot Control Lokesh Krishna, Nikhil Sobanbabu, Quan Nguyen, [Webpage\|Arxiv\|Video\|Code]

Projects

	Augmenting Learned Centroidal Controller with Adaptive Force Control OCRL Project (Spring 2025) Nikhil Sobanbabu, Kailash Jagadeesh, Tony Tao, Bharath Sateeshkumar Improving Payload Adaptability of CaJun controller by Augmenting it with ℒ₁ Adaptive Control. [Report\|Poster \|Code]
	Terrain Traversability Analysis with Image Dehazing Deep Learning Project (Fall 2024) Nikhil Sobanbabu, Woojin Kim, Ritarka Samanta Investigation of different image dehazing techniques to improve terrain traversability analysis for Autonomous vehicles. [Report \|Video]
	Single-Pursuer Multiple-Evader Target-Guarding-Game Optimal Strategies found for a 1vn game by solving the HJI-PDE equation for a 1vn game with a reach-avoid objective. [Thesis Report]
	Multi-agent trajectory tracking for Crazyflie Quadrotors Coordinated autonomous control of multi-agent quadrotors in a regular polygon formation with varied orientations using the in-built mellinger controller. [Report] \| [Poster]
	Control of 5DoF Manipulator of a Mars Rover as part of Team Anveshak Nikhil Sobanbabu,Shivendra Verma Autonomous control of Manipulator using inverse kinematics and ROS moveit.
	E-yantra Robotics Competition. Theme: Krishi-bot Nikhil Sobanbabu, Gokul MK Autonomous navigation of an Agricultural robot to identify and pick and place bellpepers inside a greenhouse. [Video] \| [Github]
	Receding-Horizon mode planner for mode planning against perturbations Nikhil Sobanbabu, Lokesh Krishna Planner chooser behaviours encoded as latent modes using monte-carlo roll outs to be robust against perturbations generating emergent transitions between modes/behaviors. [Results Video]
	Multi-agent Game Theoretic Framwork for Target-Attacker-Defender game Nikhil Sobanbabu, Shivendra Verma Simulation environment for the single attacker, singe target, multiple defender pursuit evation differntial game.

Course Projects

Swing-up and Stabilisation of inverted pendulum.
Course EE6415 Non-Linear System Analysis

Swing-up is done using a control from an energy based Lyapunov function. After a reaching an appropriate angle, pole-placement based stabilisation kicks in.

[Report] | [Presentation video] | [Github]

Motion planning for a KUKA mobile Manipulator
Course ED5215 Intro to Motion Planning
Nikhil Sobanbabu, Balaji R, Kanishkan M S

Problem Statement: Optimal pick and place of multiple objects to a given destination with payload constraints for the mobile manipulator.
Solution: Dijkstra+Modified TSP for navigation. RRT* for manipulation.

[Report] | [Github]

Fun Mini-Project

Why WASD when you can play minecraft with ESP32

Miscellaneous
	Tech and Innovation Fair, coordinator 2022. Flagship event of Shaastra, IITM's Technical festival for the year 2022. Participants convert a prototype to a minimum viable product. [Link]

Discussion with Bob-Balram (Chief Engineer of Ingenuity Helicopter)	Former Indian Defence Secretary, Ajay Kumar

Services

Reviewer for IEEE International Conference on Robotics and Automation (ICRA) 2025
Reviewer for IEEE Robotics and Automation Letters (RAL) 2025
Reviewer for Conference on Robot Learning (CoRL) 2025
Reviewer for IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2025

Template from jonbarron.