Aravind Sairam

An end-to-end pipeline for structural health monitoring. This project uses drone-captured footage to create high-fidelity 3D models via photogrammetry, then runs a semantic segmentation model (YOLOv8) to detect and visualize 19 classes of defects directly onto the 3D mesh.

Person-Following Drone with ESP32

An autonomous drone built on ESP-Drone firmware that uses an ESP32-CAM to stream video. A ground station running Python with YOLOv8 detects a person in the video feed and sends real-time control commands (Pitch, Roll, Yaw) back to the drone via UDP sockets to actively follow the target.

Designed a ROS2 simulation featuring a UR5e/xArm robotic arm for an automated kitchen cleaning task. The robot performs scene segmentation via an RGB-D camera and executes precise circular cleaning paths within a basin using MoveIt for motion planning.

Dual-Robotic SLAM Search & Rescue

Led a team to develop a search-and-rescue system with two TurtleBots, achieving a 78% success rate in complex Gazebo simulations. The project involved coordinated multi-robot SLAM, map merging, and human detection using Darknet (YOLO).

IoT Brake Pad Wear Warning System

Engineered an IoT-based brake pad wear warning system to enhance safety protocols for heavy vehicles. The system used custom sensors, an Arduino microcontroller, and a GSM module for real-time data transmission and visualization on the ThingSpeak IoT Server.

Disabled-Friendly E-Vehicle

Designed and prototyped an inclusive four-wheeled electric vehicle for wheelchair users. Focused on ergonomic design and accessibility, implementing ultrasonic sensors and DC motors for obstacle avoidance, with a Raspberry Pi running Python and OpenCV for control.