Job Description

Project Overview

We are looking for a highly motivated R&D Intern to support the integration of AI-based flight control algorithms into the PX4 Autopilot framework.

The project focuses on deploying pre-trained neural network models (including Feedforward Neural Networks, LSTM, and GRU architectures) for UAV attitude control and payload control. The selected intern will primarily research, design, and implement the most suitable approach for integrating these trained models into the PX4 ecosystem, enabling deployment and validation on a real quadcopter platform.

This is an excellent opportunity for candidates interested in autonomous systems, UAVs, artificial intelligence, robotics, and embedded flight control.

Key Responsibilities

• Study the PX4 Autopilot architecture and identify suitable integration points for AI-based controllers.
• Research and recommend the most effective methodology for integrating pre-trained neural network models into PX4.
• Integrate trained AI models (Simple Neural Networks, LSTM, and GRU) into the PX4 flight control framework.
• Support deployment and testing using PX4 SITL/Gazebo simulation before hardware implementation.
• Assist in deploying and validating the integrated controller on a quadcopter.
• Collaborate with the research team to troubleshoot integration challenges and improve system performance.
• Document the integration workflow, implementation details, and experimental results.

Required Skills

• Good understanding of control systems fundamentals.
• Basic knowledge of machine learning and deep neural networks.
• Familiarity with recurrent neural networks such as LSTM and GRU.
• Basic understanding of Reinforcement Learning (RL) and Deep Reinforcement Learning (DRL).
• Strong programming skills in Python and C/C++.
• Mandatory: Working knowledge of the PX4 Autopilot framework and its development environment.
• Familiarity with Git and Linux-based development.

Preferred Skills

• Experience with UAVs, drones, or robotic systems.
• Knowledge of ROS/ROS2 and MAVLink communication.
• Experience with Gazebo or PX4 SITL simulation.
• Familiarity with embedded systems and flight controller hardware.
• Experience with PyTorch or TensorFlow.

Eligibility

• Final-year / undergraduate or postgraduate students in Aerospace Engineering, Robotics, Computer Science, Electronics, Electrical Engineering, Artificial Intelligence, or related disciplines.
• Candidates with relevant personal or academic projects in PX4, UAVs, robotics, or AI are encouraged to apply.

Expected Deliverables

By the end of the internship, the intern is expected to:

• Develop a robust methodology for integrating AI-based controllers into PX4.
• Successfully integrate the provided trained neural network models within the PX4 Autopilot framework.
• Validate the integrated controller through simulation and, where feasible, real-flight testing.
• Prepare clear technical documentation covering the integration process, implementation, and testing results.

Duration

3 months (extendable)

Desired Candidate Profile

We are looking for a self-driven and research-oriented candidate who enjoys solving challenging engineering problems. The ideal candidate should be eager to learn, comfortable working with open-source software, and capable of independently researching technical solutions while collaborating effectively with the R&D team.

Skills