SG-AMP: Scene-Graph-Guided Active Perception and Semantics-Aware Motion Planning for Pepper Plants
IROS 2026 · Workshop
From Innovation to Adoption: Hardware, AI, Policy, and Funding Challenges in Agricultural Robotics
Robotics Researcher
Shiva Rudra
Lolla
Helping robots see, navigate, and act.
I’m Shiva Rudra, a robotics researcher passionate about robot vision and intelligent systems. My work spans building systems that help robots understand complex environments, make reliable decisions, and operate effectively in the real world across both research and applied settings.
Robot Vision & 3D PerceptionPhysical AIAutonomous Systems
My background
From research to real-world robotics.
Research
Student Researcher
Humanoid Robots Lab, University of Bonn, Germany
December 2025 - PresentResearch Intern
Robotics Research Center, IIIT Hyderabad, India
June 2023 - December 2023
Industry
Robot Navigation Engineer
AlphaZ, Los Angeles, California, USA
May 2024 - September 2025
Education
M.S. in Mobile Robotics
University of Bonn, Germany
October 2025 - September 2027 (Expected)B.E. in Electronics and Instrumentation
BITS Pilani, Hyderabad, India
November 2020 - May 2024
Publications
Research papers and workshop contributions.
Projects
Ideas brought to life through code, hardware, and experiments.
*
Developed a geometry-based planner to find and rank leaf-clip placements with supported contact and clear insertion paths. Results include pose-quality heatmaps, clip-placement visualizations, and cuRobo motion demonstrations on a robot model with payload and B2 in Isaac Sim.
Explore project Developed a workflow to turn airborne and mobile LiDAR data into 3D urban building models through point-cloud processing, building segmentation, and LoD2 reconstruction. Parameter optimization increased valid reconstructions from 70% to 91% across 43 tested buildings.
Explore project The ESDF map is made using pointcloud data from the depth camera and rectified odometry from the VINS framework. Voxfield framework is utilized for this purpose
Explore project Designed the control system for translational and jumping motion of a wheeled biped robot. LQR controller is used for translational self balancing motion and PD-fuzzy logic controller is used for jumping motion
Explore project Designed and Simulated an underwater ROV control system using PID and LQR controllers and performed a comparative study of both the models
Explore project Implemented drone detection using Yolo algorithms and techniques like layer freezing, image augmentation. Performed comparitive study of various scenarios and test cases
Explore project Design and fabrication of a Resistive touch controlled manipulator
Explore project A spherical surveillance bot which can move around get a live video feed of the surroundings
Explore project Blogs
Notes from building, testing, and learning.
Contact
I’m always excited to connect over robotics, research, and ideas that push intelligent systems forward!