Vamsi Kalagaturu

Robotics researcher, SECORO — University of Bremen

I am a PhD student with SECORO at the University of Bremen. I work on modelling complex, contact-rich manipulation tasks for mobile manipulators as geometry and force constraints. The models are graph-structured, so the pieces of a task compose and carry over to new tasks and platforms instead of being rewritten. From those models, the tooling generates correct-by-construction code, validated the same way in simulation and on real hardware.

GitHub Scholar ORCID LinkedIn

Based in
Bremen, Germany
Works with
C++ · C · Python · ROS · RDF/JSON-LD · JavaScript
Focus
Robotics manipulation · Integration · Software modelling

Research

Contact-rich manipulation
Controllers and skills for tasks where the robot must push, slide, and stay in contact, reasoning about forces rather than positions.
Composable robot models
Models that snap together and get reused across tasks and platforms, generating the code for a skill instead of leaving it to be hand-written every time.
VLMs & VLAs
Using vision-language and vision-language-action models to generate semantics and models rather than trajectories: the model says what a task means, and the robot's own models work out how to move.

Publications

  1. 2025

    From Composable Models to Correct-by-Construction Software for Contact-Rich Robotic Mobile-Manipulation Tasks

    S. Schneider, V. Kalagaturu, H. Bruyninckx, N. Hochgeschwender · IEEE Robotics and Automation Letters

  2. 2024

    b-it-bots: Winners of RoboCup@Work 2023

    G. Chenchani, K. Patel, R. Selvaraju, S. Shinde, V. Kalagaturu et al. · RoboCup 2023: Robot World Cup XXVI, Lecture Notes in Computer Science

  3. 2023

    Viewpoint Push Planning for Mapping of Unknown Confined Spaces

    N. Dengler, S. Pan, V. Kalagaturu, R. Menon, M. Dawood, M. Bennewitz · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Tools

Mujoco KDL Wrapper
A C++ library bridging MuJoCo physics simulation with KDL for robot kinematics and dynamics. Builds KDL chains straight from the compiled MuJoCo model, handles multi-robot scenes, and ships Python bindings.

Awards

Experience

Education

Contact

Collaborations, questions, or a chat about manipulation and control: my inbox is open.