Divyam Sood
Building resource-aware robust autonomy for real robotic systems
Undergraduate student working on robotics, control systems, and embedded hardware, with a focus on resource-efficient sytems and sim2real transfer. My current interests broadly include model predictive control, reinforcement learning, and real-world robotic systems.
Publications
SHIELD: Safe Hybrid Integration of PPO and MPC for Reliable Trajectory Tracking Oral presentation at RITA 2025, London.
TRaCe: Trajectory Tracking with Budgeted Localization using Meta-RL and Belief-Aware MPC Oral presentation at RITA 2025, London.
Research
Undergraduate Researcher — MuSiC Lab, IISER Bhopal Jan 2025 – Present
- Developed ROS2–Gazebo simulation pipelines for UGV and UAV systems.
- Implemented sim-to-real control and RL policies on TurtleBot3 and Crazyflie.
- Analyzed robustness of hybrid RL–MPC systems under localization uncertainty.
- Modeled soft robotic actuators using finite element analysis in COMSOL.
- Designing event-triggered and resource-aware control frameworks.
Skills
Robotics & Simulation: Robot Operating System (ROS2), Gazebo, Crazyflie, TurtleBot3, Sim-to-Real Workflows like SITL
Control & Learning: Model Predictive Control (MPC), Reinforcement Learning, Resource-Constrained Navigation
Programming & Embedded: Python, C++, Verilog, AVR Assembly
Digital Design & Hardware: FPGA Development, Digital Logic Design, Sensor Integration, Analog Circuits
Engineering Tools: Linux, Git, PlatformIO, LTSpice, Fusion 360, AutoCAD, LaTeX
Writing
I write about my experiences here at college, about workshops I attend and organize, troubleshooting experiences, and technical work in general.
Read the blog.