Background

The Autodrive 3 (AD3) project is based on the Robot Operating System 2 (ROS 2) framework and utilizes a robotic model car platform. Currently, the AD3 architecture distributes computation between two systems: a Raspberry Pi 2 mounted on the model car, and a remote PC. The Raspberry Pi 2 is responsible for reading sensor data from a 2D spinning LiDAR and an IMU. This raw sensor data is transmitted via Wi-Fi to the remote PC, which performs sensor fusion, SLAM-based localization, and path planning. The resulting control commands are then sent back to the car for maneuvering.

While this architecture provides a working proof-of-concept, it is limited by the processing capabilities of the Raspberry Pi 2 and the latency introduced by Wi-Fi communication. Recent advances in embedded computing, particularly the Raspberry Pi 5, offer significantly improved processing power and memory bandwidth compared to its predecessor. This opens the possibility of consolidating all computation including sensor fusion, SLAM, and navigation directly on the vehicle. Eliminating Wi-Fi communication delays may not only improve real-time performance but also simplify the system architecture, making the model car a fully self-contained autonomous system.

Work description

This will include evaluating real-time scheduling of ROS 2 nodes to achieve deterministic execution and responsiveness.

• System Migration: Port the AD3 ROS 2 stack from a distributed setup (Raspberry Pi 2 + PC) to a single Raspberry Pi 5, ensuring full functionality of sensors (LiDAR, IMU) and local ROS 2 node communication.
• Real-Time Architecture: Implement deterministic execution using ROS 2 executor strategies, PREEMPT-RT kernel, CPU isolation, and scheduling for critical nodes like SLAM and motion control.
• Performance Evaluation: Benchmark latency, throughput, and determinism for both the distributed and integrated systems, assessing improvements from eliminating Wi-Fi and meeting real-time constraints.
• Analysis & Documentation: Compare efficiency, latency, and reliability, and provide recommendations for embedded ROS 2 deployment.

Qualifications

• Background in Computer Science, Robotics, Embedded Systems, or a related field.
• Experience with ROS 2, embedded Linux, and real-time computing is desirable.
• Good knowledge in both Swedish and English. 

To give you the best possible support during your thesis, we’d like you to be able to come to the office connected to the project and spend most of your time working from there.

Application:

We look forward to receiving your resume, and preferably, a personal letter in which you explain why you want to write your thesis with Syntronic.

We screen and evaluate applications on an ongoing basis. The thesis project may be filled before the application deadline.