Thesis project 30hp - Domain Adaptation for Robust Anomaly Detection in Industrial Systems

Background:

Scania is one of the world’s leading manufacturer of trucks and buses for heavy transports, as well as industrial and marine engines. Transport services and logistics services make up an increasing part of our business, which guarantees Scania’s customers cost-efficient transport solutions and high availability. Over a million Scania vehicles are in active use, in over 100 countries.

In the Connectivity section within Scania R&D, we develop new solutions for connected vehicles in our Internet of Things (IoT) platform, as part of Scania’s increasing focus on communication, services, and smart transport solutions. Advanced data analysis capabilities are a cornerstone enabler in this development. 

Target/scope:

Modern cyber-physical systems (CPS) — such as those used in transportation, manufacturing, and industrial automation, continuously generate large volumes of data from sensors, controllers, and human–machine interactions. This data offers immense potential for building data-driven models that can detect anomalies, estimate component health, and predict the Remaining Useful Life (RUL) of critical subsystems.

However, real-world data collected from these systems often lacks reliable healthy data and labels. Additionally, it is affected by sensor noise, operational variability, and domain shifts. These challenges limit the direct applicability of learning methods. To overcome this, simulation-based data generation and domain adaptation techniques are increasingly being explored to bridge the gap between simulated and real-world data.

This thesis aims to design and validate anomaly detection models that can learn from simulated environments and adapt to the data distribution of real-world systems. The student will use the Causal Chambers framework, a controlled experimentation setup for causal inference and data collection — to design experiments and test domain adaptation approaches for Scania’s physical systems.

The thesis will involve the following key steps:

1. Design experiments: design faults and collect relevant data from framework (simulator and physical system).

2. Build evaluation measures to evaluate performance.

3. Baseline model development, test baseline model’s performance on physical data by training on data from simulator.

4. Build models for domain adaptation.

5. Test the proposed approach on a Scania’s system.

Students will have access to appropriate computational infrastructure and expert supervision throughout the project.

What skill are for needed this thesis:

• Required

  • Knowledge on basic of Neural networks (NN) e.g. back propagation, loss, supervised learning, convolution networks, recurrent networks

  • Knowledge about generative models e.g. GANs, Auto-encoders.

  • Knowledge on gradient based optimisation.

• Bonus

  • Knowledge of domain adaptation or transfer learning.

Education/line/direction:

Area of education or direction: Masters programmes in Machine Learning, Data Science, Computational Mathematics, Complex Adaptive Systems, Computer Science or similar.

Number of students: 1-2

Start date for the Thesis project: January 2026

Estimated timescale: 20 weeks

Contact person and supervisor:

Abhishek Srinivasan, Data Scientist, 08-553 816 96,

Juan-Carlos Anderesen, Manager, 08-553 835 16,

Application:

Your application should contain the following:

1. CV.

2. Personal letter.

3. Copies of grades.

4. Optional: To propose tentative approach to the problem.

Applicants will be assessed on a continuous basis until the position is filled. Do not wait until the last date to apply. 

A background check might be conducted for this position.

Nothing is more important than our people. With a company culture that’s based on shared and inclusive core values. We have a supportive community, where everyone has opportunities to grow and succeed.