vishal73335238

Nice idea, bro. Best of luck:yes:

Oops! I forgot to mention one more thing in my previous post.

I've also submitted my Project Showcase Pull Request to the official Robotics Dream Keeper Challenge repository.

Project: RDK X5 Autonomous Vehicle
GitHub Repository: https://github.com/p...onomous-Vehicle
Project Showcase PR: https://github.com/D...allenge/pull/15

Thank you once again to the D-Robotics team for organising this challenge and providing such a great platform to learn and build with the RDK X5!

Stage 3 - Launch Submission | Robotics Dream Keeper Challenge 2026

Hi D-Robotics Team!

I'm excited to submit my Stage 3 - Launch project for the Robotics Dream Keeper Challenge 2026.

Project: RDK X5 Autonomous Vehicle

I built a 1/10 scale autonomous vehicle powered by the D-Robotics RDK X5 running ROS 2 Humble. The vehicle combines classical computer vision with AI to navigate a miniature road while detecting objects in real time.

Key Features
OpenCV-based lane detection with perspective transformation
YOLOv11 object detection
Stop sign and obstacle detection with automatic vehicle stop
Manual / Autonomous driving modes
Stereo Vision MIPI Camera
ESP32-C3 vehicle controller for steering and throttle
Modular ROS 2 architecture
USB serial communication between RDK X5 and ESP32-C3

During development, I initially planned to use the D-Robotics racing track detection model for lane detection. However, since it wasn't directly suitable for my custom miniature track and training a dedicated model would have exceeded the available project timeline, I implemented an OpenCV-based lane detection pipeline while using the RDK X5 BPU for YOLO object detection. This resulted in a reliable and responsive autonomous driving system.

GitHub Repository

https://github.com/p...onomous-Vehicle

Demo Video

Very close to completing the project, planning to submit the Stage 3 very soon.
 DKS_Pro_Car.jpg   1,29 Mo   0 téléchargement(s)
 RDK-X5_Autonomous_Car.jpg   1,53 Mo   0 téléchargement(s)

RDK Challenge - Stage 2 Build Challenge Completed

Hi everyone!

I'm excited to share that I've completed Stage 2 - Build Challenge for my project:

RDK X5 Autonomous Vehicle

During this stage, I focused on designing the complete software and hardware architecture before starting the actual build.

Stage 2 Progress
Designed the complete autonomous vehicle architecture
Selected all hardware components
Designed the ROS 2 software architecture
Created the system flow and hardware diagrams
Defined the AI perception and behaviour planning pipeline
Completed the Bill of Materials (BOM)
Created the engineering roadmap and development milestones
Completed the project proposal and risk analysis
AI Stack

The vehicle will use two AI models running concurrently on the D-Robotics RDK X5:

D-Robotics Racing Track Detection (ResNet18) for road tracking and lane following
YOLOv11 (planned) for vehicle and obstacle detection

The software is being designed using a ROS 2 architecture, with a DFRobot Beetle ESP32-C3 acting as the vehicle controller over UART.

Planned Features
Autonomous road tracking
Vehicle & obstacle detection
Autonomous lane following
Autonomous lane changing
Manual RC override using the RadioLink RC6GS transmitter

I'm now moving into Stage 3, where I'll begin building the hardware, integrating the software stack, and testing autonomous driving on a custom two-lane miniature road.

GitHub Repository:
https://github.com/p...onomous-Vehicle

I appreciate all the support from the D-Robotics team and the community. Looking forward to sharing more progress soon!

#RoboticsDreamKeeper #RDKX5 #ROS2 #EdgeAI #AutonomousVehicle #ComputerVision