Thesis Project: 30 hp - RL-based decision-making for autonomous driving in heavy duty vehicles

Introduction
Thesis work is an excellent way to get closer to Scania and build relationships for the future. Many of today's employees began their Scania career with their degree project.

Background 
Autonomous driving technologies are poised to revolutionize transportation, offering enhanced safety, increased efficiency, and improved accessibility. In the context of heavy-duty vehicles, decision-making is critical for navigating complex environments and ensuring safe, reliable operation.
Reinforcement Learning (RL), unlike traditional methods, enables systems to learn optimal strategies through experience, eliminating the need for extensive training datasets. Additionally, RL-based decision-making can handle uncertainty about the behavior of surrounding vehicles by utilizing frameworks like Partially Observable Markov Decision Processes (POMDPs). This thesis will explore the use of RL to develop decision-making and control strategies for autonomous driving in heavy-duty vehicles.

Objective 
Investigate, implement and evaluate RL-based methods to perform autonomous driving with heavy-duty vehicles in realistic driving environments. Within this rather broad topic, you as the student have a lot of freedom to focus on areas that you find particularly interesting. Example focus areas are:

• Implement state-of-the-art deep learning architectures to enable better RL-based driving policies.
• Leverage a combination of real-world driving data and in-simulator driving data to narrow the sim-to-real gap.
• Enable the RL-agent to account for different vehicle configurations, e.g., with and without the presence of a trailer, different trailer lengths, etc.

You will have a “baseline” implementation that you can optionally start from.

Job description 
The thesis can roughly be divided into the following sub-tasks:

• Survey and summarize related literature.
• Based on your focus objective: propose, develop and implement an RL-based method to navigate a heavy duty vehicle in a simulated environment containing both dynamic and static obstacles.
• Systematically evaluate the learned policy of the RL-agent.
• Summarize the findings in a thesis chapter.

Education/program/focus
You are currently studying in the final year of a Master’s degree in computer science, robotics, engineering physics, electrical engineering, or applied mathematics, preferably with specialization in artificial intelligence algorithms, control theory, optimal control or optimization. Knowledge of programming, and reinforcement learning are a plus.

Number of students: 1
Start date for the thesis work: January 2025
Estimated time required: 20 weeks (full-time)

Contact persons and supervisors
Yunus Emre Sahin, Research Engineer in Autonomous Motion: 
yunus-emre.sahin@scania.com, +46 7 008 143 26

Oscar Palfelt, Development Engineer in Autonomous Motion: oscar.palfelt@scania.com, +46 7 208 399 49

Application:
Your application must include a CV, personal letter and transcript of grades 

A background check might be conducted for this position. We are conducting interviews continuously and may close the recruitment earlier than the date specified.
 

Thesis Worker