How can I integrate my custom reinforcement learning agent with the PX4 Autopilot architecture in Simulink, train it within this setup, and then deploy it onto a real drone?

Hello MATLAB Community,

I am currently working on a project where I intend to integrate a custom reinforcement learning (RL) agent with the PX4 Autopilot system. My objective is to use MATLAB and Simulink to design and train this RL agent within a simulated environment and later deploy it onto a real drone running PX4 firmware.

Current Setup:

1. Installed Toolboxes:

• I have all the necessary toolboxes installed, including the Reinforcement Learning Toolbox, Simulink, PX4 Support from Embedded Coder, and UAV Toolbox.
• I am using MATLAB R2024a.

What I Need Help With:

Setting Up the RL Agent in Simulink:

• I need guidance on how to set up and design my own RL agent within Simulink.
• Specifically, I want to replace the position and attitude controllers from the PX4 Autopilot architecture with my RL agent and then start the training process. Could you provide steps on how to configure the Simulink environment for this setup?

Code Generation and Real-World Deployment:

• Once the RL agent is trained and validated in the simulation, what steps should I follow to generate code from Simulink and deploy it onto a real drone running PX4?
• Are there specific considerations or configurations needed for a smooth transition from simulation to real-world testing?

Real-World Testing Considerations:

• What additional precautions should I take when deploying and testing the RL model on an actual drone?

Any detailed guidance, documentation, or examples that you can provide would be immensely helpful in helping me achieve this integration and deployment successfully.

Thank you for your support!