My Projects
Here’s a selection of my robotics, AI/ML, and embedded systems projects that reflect my expertise in simulation, control systems, and intelligent autonomy.
Date: April 2025
- 🤖 Developed leader–follower architecture for Autonomous Surface Vehicle (ASV) and Autonomous Underwater Vehicle (AUV).
- 🌊 Implemented trajectory tracking, obstacle avoidance, and underwater–surface communication using ROS2.
- ⚙️ Designed cascaded feedback linearization controllers for underactuated dynamics in Gazebo simulation.
- 🛰️ Integrated IMU, depth, and GPS sensors for multi-agent coordination.
Implemented a cascaded PID and feedback linearization-based control for underactuated AUV in ROS2, managing surge, pitch, and yaw motions with stability guarantees.
- 🧠 Integrated outer-loop PID with inner-loop FBL control for precise trajectory tracking.
- 🧩 Developed custom ROS2 node for real-time control using Gazebo simulation.
An integrated robotic bin-picking system using AI-based object detection and ROS2 motion planning.
- 🧠 Used YOLOv5 for object detection and Point Cloud Library (PCL) for pose estimation.
- 🦾 Integrated GPD (Grasp Pose Detection) with MoveIt2 for grasp planning and execution.
- 🪄 Designed modular ROS2 nodes for perception, planning, and control layers.
Built a complete navigation pipeline in Gazebo with mapping, localization, and path planning using ROS2 and Nav2.
- 📍 Simulated realistic campus environments with dynamic obstacles.
- 🚗 Implemented SLAM, costmaps, and DWA local planners.
- 📡 Adapted Nav2 framework for differential-drive and drone-like motion.
- 📐 Modeled WMR kinematics for line and circular formations using Lyapunov-based control laws.
- 🧪 Implemented control algorithms in MATLAB for multi-robot simulation.
- 📊 Demonstrated robustness against dynamic disturbances and delays.
Developed an IoT-based home automation system with TensorFlow Lite-based voice recognition and sensor integration.
- 🎙️ Enabled real-time control of appliances via voice commands.
- 🌡️ Integrated temperature, gas, and motion sensors for intelligent automation.
- 💡 Built using Arduino, Bluetooth, and embedded C++.
Designed a computer vision pipeline for underwater inspection tasks.
- 🔍 Used LAB color space, Gaussian blur, Canny edge detection, and Hough Transform for robust detection.
- 💧 Optimized for low-light and noisy underwater imagery.
Customized TIAGo robot’s URDF/Xacro to operate only the base (excluding arm and torso) for differential-drive navigation experiments.
- ⚙️ Rebuilt simplified URDF for base-only operation.
- 🏗️ Tested integration with Nav2 and SLAM for indoor navigation.
Designed a MobileNetV3-inspired CNN model optimized for real-time agricultural disease detection on edge devices.
- 🌱 Achieved high accuracy with minimal model size.
- 📱 Deployed on mobile hardware for field testing.
