Engineering a Reliable System

This project documents the design, development, and deployment of an autonomous mobile robot capable of exploring an unknown indoor maze, identifying heat sources, and launching a projectile at them using a dual-flywheel mechanism.

Inspired by real-world thermal response scenarios (disaster rescue), the bot was designed for full autonomous functionality.

Our Robot

Final navitgation run at 4x speed

🔧 Key Features

• ROS2-based Modular System Leverages Nav2 for exploration, MPPI for local control, and custom Python logic for mission flow.

• Thermal Detection & Localization Dual AMG8833 sensors fused with LIDAR to precisely locate and track heat sources.

• Autonomous Target Engagement Flywheel launcher actuated at runtime using servo-driven rack-and-pinion feed.

• Real-Time Decision Logic FSM architecture using multi-threaded execution (fast sensor loop + slow decision loop).

System Overview

Each subsystem is orchestrated via a centralized GlobalController node:

• State-driven transitions (Exploration → Goal Nav → Launching)
• IMU-based ramp detection & hazard avoidance
• Heat clustering via KMeans
• Visualization via RViz Markers

🚀 Mission Pipeline

1. Mapping: Navigate maze autonomously using LIDAR + SLAM.
2. Heat Detection: Thermal data processed & filtered to detect anomalies.
3. Localization: LIDAR returns from angular bins are fused with robot pose to get (x, y) coordinates.
4. Clustering: KMeans clustering reduces noise and sets launch targets.
5. Engagement: Robot drives to each heat source, aligns, and fires.

📎 Project Links

What I Learned

• The Art of building complex systems - With more moving parts, rate of failure grows exponentially
• Multi-threaded execution in ROS2
• LIDAR and thermal IR