Failure-Aware Multi-Modal SLAM
Real-time sensor health monitoring and adaptive confidence-based weighting for robust indoor localization under RGB-D, LiDAR, and IMU degradation. Benchmarked against RTAB-Map with ATE/RPE metrics.
Sreeram M.V.
Robotics Researcher · SLAM · Legged Manipulation · HRI
Predoctoral Researcher at IISc HiRo Lab · Open to PhD 2026/27
I'm a robotics researcher who builds real-world autonomous systems — from SLAM and sensor fusion to legged locomotion and human-robot interaction. I care about robots that work outside the lab, and that know when to say “I don't know.”
Throughout 4 years of B.Tech, awarded highest department grade for projects. Ranked 1st every year. 4th year: 1st in department, CGPA 10.0/10.
drag joints to pose · click gripper
01 — Research
Research Directions
Digital Twin-Driven Multi-Robot Navigation
EKF-based multi-sensor fusion and event-triggered RL re-planning for GPS-denied environments. URDF-based digital twin validated in Gazebo/RViz with teleoperation and closed-loop feedback.
Quadruped Locomotion & Manipulation
Whole-body control for legged robots on Unitree Go2/B2 platforms. Human-robot interaction, teleoperation, and deploying state-of-the-art methods for real-world unstructured environments under Prof. Ravi Prakash at IISc HiRo Lab.
02 — Publications
Publications
Springer CCIS 2334, ANTIC 2024
Mining the Vaccine Pulse: NLP-Powered Twitter Analytics
Co-author. VADER sentiment analysis of 10K+ tweets.
03 — Projects
Selected Projects
Quadruped Locomotion — Unitree Go2/B2
Current work at IISc HiRo Lab. Whole-body control, imitation learning, RL-based locomotion, and human-robot interaction for Unitree Go2/B2 legged robots in unstructured environments.
Adaptive Digital Twin for GPS-Denied Navigation
EKF-based multi-sensor fusion with event-triggered RL re-planning for GPS-denied environments. URDF digital twin validated in Gazebo/RViz with teleoperation and closed-loop feedback. Built at IISc-ARTPARK.
TortoiseBot Pro Max: Full-Stack Autonomous Platform
ROS 2 packages integrating Intel RealSense D435i, YDLidar, and wheel encoders. Structured rosbag datasets for SLAM robustness analysis under depth lag, LiDAR dropouts, and IMU drift. RTAB-Map validation. 10 months at IISc-ARTPARK.
FlexiArm: 6-DOF Robotic Arm
A 6-DOF robotic arm we built from scratch to learn robotics hands-on. It implements the full forward and inverse kinematics — working out where the gripper is from the joint angles, and the joint angles needed to reach a target — and is driven through ROS, which we used to understand the whole robotics stack end to end.
MineSwarm: Multi-Robot Mining Digital Twin
Multi-agent RL digital twin for decentralized navigation of 3 LiDAR-equipped robots in GPS-denied mines. Curriculum learning, phase-aware replay, twin-critic stabilization. 91% goal-success across 8-phase curriculum.
WeedSense / AgroX: Autonomous Precision-Agriculture
Targeting ~34% weed-induced crop loss. Mask R-CNN (ResNet-101) for 92% accurate real-time weed detection. ROS 2 + IoT integration for autonomous irrigation. Multilingual dashboard and cloud monitoring.
04 — Mentored
Mentored
Teaching & Community
-
MIT Bengaluru — Semi-Autonomous Robotics Coursework
Mentored 75+ students through the full robotics stack — from scratch-built sensor-actuator interfaces to ROS 2 integration, state estimation filters, and high-level autonomous algorithms. Designed hands-on modules for a week-long intensive.
-
ARTPARK IISc — Robotics Hackathon Ground Engineer
Served as on-site robotics engineer during ARTPARK hackathon. Helped 20+ teams debug hardware (RealSense, LiDAR, motor drivers) and software (ROS 2, SLAM, navigation) issues under time pressure. Ensured working demos across diverse platforms.
05 — Experience
Experience
-
IISc HiRo Lab
Predoctoral Researcher / Research Associate
Quadruped locomotion and manipulation on Unitree Go2/B2 under Prof. Ravi Prakash. Whole-body control, HRI, teleoperation.
-
IISc-ARTPARK
Research Intern
ROS 2, TortoiseBot Pro Max, sensor integration, SLAM, CodeBot 2025. Advised by Prof. Amrutur Bharadwaj and Dr. Josephine Ruth.
-
MIT Bengaluru
Mentor
Mentored 70+ students in ROS, sensor integration, LeapMotion gesture control.
-
Eureka Club
Founder & Lead
Research culture, 10+ R&D team, 7+ student papers.
06 — Toolchain
Skills & Toolchain
- Languages
- Robotics Stack
- Simulation & Learning
- Tools
07 — Beyond
Beyond the Lab
A travel freak who clicks the moments — always chasing new places with a camera in hand. Off the lab, I run robotics workshops for local government-school students through NGO drives, play guitar, read and write, wander down philosophy rabbit holes, build little toy robots, and lose the occasional game of badminton.
“The best way to predict the future is to invent it.”
— Alan Kay
In Pictures
.jpg)
.jpg)
.jpeg)
.jpg)
.jpg)
.jpg)
08 — Resources
Resources
The Landscape of Robot Learning
Sim-to-Real Transfer
Key methods for deploying policies from sim to reality
Legged Locomotion
From Boston Dynamics to Unitree to academic control
Multi-Modal SLAM
Fusing LiDAR, visual, inertial for robust localization
HRI & Teleoperation
From teleoperation to collaborative manipulation
Imitation Learning
Learning control policies from expert demonstrations
On Reaching Out
When — and When Not — to Cold Email
- Research their recent papers first. Mention one specifically.
- Keep under 150 words.
- Attach CV + 2-line project summary.
- Don't email during conference deadlines.
- Follow up once after 10 days.
“Research is a conversation. The best emails start with listening.”
Wisdom
Tools
Let's Build the Future
Open to PhD positions 2026/27 and research collaborations in SLAM, legged robots, and human-robot interaction.
mvsreeramblr@gmail.com