🛸 UAV Lab: Cyber-Physical Systems & Edge AI Roadmap

An engineering-focused repository tracking my progression into Unmanned Aerial Vehicles (UAVs), Robotics Middleware, Edge AI, and Cyber-Physical Systems.

Building upon a background in Software Engineering and IoT, this lab focuses on low-latency messaging, distributed ROS2 architectures, PX4 autopilot integration, computer vision on edge devices, and autonomous navigation.

---

🗺️ Roadmap & Epics (Pragmatic MVP)

📌 [EPIC-01] Block 1: ROS2 Middleware Core & Messaging Performance

• Objective: Master ROS2 cyber-physical software architecture, DDS QoS policies, and multi-threading execution in rclcpp using standard simulation models.
• Core Competencies: ROS2 Graph, Executors, Multi-threaded Spinners, DDS QoS profiles (Reliable vs. Best Effort, Volatile vs. Transient Local), Latency benchmarking, Shared-Memory IPC (rclcpp).
• Key Deliverables:
  • High-frequency (100 Hz) C++ Pub/Sub pipeline measuring DDS latency & QoS profiles.
  • Integration with standard Gazebo Harmonic robot models (PX4 X500 / TurtleBot3) controlled via ROS2 velocity topics (geometry_msgs/msg/Twist) while monitoring odometry.

---

📌 [EPIC-02] Block 2: Hardware Interfacing, Protocols & Go Edge Gateway (Core Focus)

• Objective: Bridge high-level edge software with PX4 SITL UAV autopilot, microservices, and secure Cloud IoT infrastructure using Go and ROS2.
• Core Competencies: MAVLink protocol, Micro-XRCE-DDS Agent integration, Go (Golang) telemetry ingestion service, WireGuard encrypted VPN tunnels, Time-Series Storage (TimescaleDB).
• Key Deliverables:
  • C++ ROS2 sensor telemetry parser node running on Raspberry Pi 5 / Edge environment.
  • Cyber-Physical Edge Gateway (Capstone Project): PX4 running in SITL mode, ROS2 bridge streaming encrypted telemetry over WireGuard from Pi 5 to a Go backend ingestion service, storing real-time data in TimescaleDB with a live React web dashboard.

---

📌 [EPIC-03] Block 3: Computer Vision & Edge AI (Python Pipeline)

• Objective: Implement real-time perception pipelines and Deep Learning inference running on edge hardware via Python (rclpy).
• Core Competencies: Python ROS2 nodes (rclpy), OpenCV video stream capture, Ultralytics YOLOv8 inference on Pi 5 / Edge CPU, target coordinate transformation, and ROS2 topic publishing.
• Key Deliverables:
  • rclpy camera node publishing target detections (geometry_msgs/msg/Pose2D) from live/simulated video streams to the ROS2 graph for downstream guidance.

---

📌 [EPIC-04] Block 4: Autonomous Navigation & State Estimation (Single-Agent Autonomy)

• Objective: Implement GPS/Odometry-based single-agent waypoint navigation and state estimation without multi-robot complexity.
• Core Competencies: EKF state estimation, Nav2 stack integration, Waypoint follow action servers, ROS2 action clients.
• Key Deliverables:
  • Autonomous single-agent navigation mission in a simulated Gazebo environment executing a sequence of waypoints received from the external Go API gateway.

---

🛠️ Quick Start & Build Workflow

Prerequisites

• Docker & Docker Compose installed on the host system.

1. Launch ROS2 Jazzy Container

From the repository root on your host machine:

./docker/run.sh

2. Build Workspace

Inside the interactive container environment (/workspace):

colcon build

3. Source Workspace Environment

To register compiled ROS2 nodes and environment hooks in your active terminal session:

source install/setup.bash

---

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.