Background

Autonomous vehicles rely on vast amounts of multi-modal sensor data—including vehicle control signals, GPS, lidar, radar, and camera inputs—to perceive and navigate their environment. Sending all this data to central storage is often impractical due to high bandwidth and storage costs, and annotating it manually is nearly impossible because of the sheer volume. Federated learning (FL) provides a distributed, privacy-preserving alternative, allowing multiple entities to collaboratively train a shared model without exchanging local data.

Despite recent advances, FL faces challenges in large-scale, real-time vehicle deployments, particularly the need for labeled data. Deep learning-based perception modules require high-quality annotations, which are costly and complex to obtain. Self-supervised learning (SSL) offers a solution by leveraging mostly unlabeled data with minimal labels. Early studies show that federated self-supervised training can achieve performance comparable to centralized approaches, with potential improvements as larger unlabeled datasets are used.

This project aims to advance federated learning for autonomous vehicles by integrating self-supervised methods with robust aggregation techniques to develop models that are efficient, generalizable, and capable of handling both common and rare driving scenarios, while reducing reliance on manual annotation and avoiding the costs of central data storage.

Description

In this master thesis project, you will focus on: 

• A novel self-supervised learning approach to exploit all available data on the central server, even with limited labels
• A hybrid federated learning scheme combining self-supervised and supervised techniques, adapted to local and global learning rounds
• Validation through extensive comparisons with fully supervised learning within the same federated scheme.
• Demonstration of the efficacy of combining self-supervised and supervised learning on ZoD dataset ((https://github.com/zenseact/zod) under various federated scenarios.
• Present findings to the project partners

Key Responsibilities

• Design, implement, and optimize federated learning systems that efficiently scale across heterogeneous devices while maintaining privacy and compliance with data protection regulations.
• Collaborate with cross-functional teams to integrate machine learning models into distributed environments, ensuring high communication efficiency, fault tolerance, and robustness against model drift.
• Conduct research and development of novel algorithms for secure aggregation, differential privacy, and model compression to enhance scalability, convergence speed, and energy efficiency in federated networks.

Qualifications

We are looking highly motivated students with a good general background in machine learning and computer vision. The following skills would be essential:

• Deep learning
• Computer vision
• Python programming
• Reading scientific papers
• Handling complex systems
• Federated learning (would be a bonus)

Terms

• Industry supervisor:
  • Sima Sinaei, PhD (sima.sinaei@ri.se)
  • Mina Alibeigi, PhD (mina.alibeigi@zenseact.com)
• Location: RISE, Kista, Stockholm
• Application deadline: November 15^th, 2025
• Starting date: As soon as possible, not later than December 1^st, 2025.
• Credits: 30 points
• Compensation: 30 000 SEK upon a successful completion of a high-quality thesis.

Welcome with your application!