HOME

SURAJ RAVAL

Empowering Intelligent Robot Motion for Real-World Impact

Projects (That Lead to Publications in Reputed Conferences/Journals)

Model-Based Simulators and Reinforcement Learning for Magnetic Surgical Robots

Zernike Polynomial Based Magnetic Field Estimation for Better Control

Modification of UNet Architecture for Segmentation and Keypoints Based Localization of Magnetic Tools in Cluttered Surgical Scenes

Enhanced Magnetic Field Articulation Through Stacks of Individually Addressed Electromagnetic Coils for Higher Force Generation

Image: Tesla Optimus by Tesla, CC BY-SA 4.0

Robotics Modeling & Simulation Intern at Tesla Optimus (June 2025 - Present)

Generating Motion Along Lie brackets for Singularity-Free Motion of Magnetic Robots

Time-Scale Separation Analysis for Surgical Needle Control in Electromagnetic Robotic Systems

Industry Experience During PhD

Human-Robot Physical Interaction Intern at Intuitive Surgical Inc. (June 2022 - Aug 2022)

FEA-Based Magnetic Field Estimation for Better Control

A view of Intuitive Surgical’s da Vinci 5 robotic system. Courtesy of Intuitive Surgical

Research Questions That Keep me Up at Night

Robust Pose estimation of rigid bodies from RGB images: Can we leverage the physics concerning rigid bodies to learn the features in a less data-hungry way? Can this lead to smaller-sized models that are more efficient?
Control Barrier Functions for Magnetic Robotics: Can we provide safety guarantees using Control Barrier Functions (CBFs) to actuate multiple magnetic tools and accomplish meaningful surgical tasks?
Behavior Cloning for Magnetic Robotics: Can we use the data collected from deployment of traditional controllers and learn out-of-distribution tasks without extensive modeling of the physics?
Pursuer-Evader Games: Can we guarantee capture of an evader with the use of implicit measurement signals?