Italdesign is looking for an enthusiastic and smart student passionate about automotive, eager to join an innovative and international business environment, working in an agile way and embracing a strong team spirit.
You will gain a great experience in learning all the instruments, procedures and technical aspects with the support of your tutor and colleagues of the Department.
Italdesign has a long-standing expertise in the development of Augmented Reality, Mixed Reality, and Virtual Reality solutions within its XR-Solutions team, part of the Strak & XR-Solutions Department.
You will be part of the XR-Solutions team and will actively collaborate to refine and improve an innovative system developed from previous research aimed at evaluating the feasibility of controlling a 3D avatar using IMU-type sensors instead of digital cameras and markers.
Thesis Project
As thesis student, you will be supported by your Tutor in order to analyse and develop the following thesis project about:
- Expanding the dataset by acquiring realistic movement data in the ConceptLab laboratory, using the ART (Advanced Realtime Tracking) system with 28 IR cameras and a full body-tracking suite, to obtain high-precision data and a wide variety of postures;
- Optimizing the AI model by training it on data acquired in realistic conditions, with the goal of improving the reliability of motion estimation. Both hardware (e.g., adding IMUs) and software (e.g., preprocessing, tuning, validation) solutions will be evaluated;
- Managing data drift (cumulative time error) by studying software and hardware approaches to compensate for it.
The different stages of the thesis may include:
- Study of the previous research activity, with a specific focus on the AI algorithm and on the sensor system;
- Experimental data acquisition at the Italdesign laboratory;
- Use of the acquired data to improve the training dataset;
- Modelling and optimization of the AI algorithm to reach millimeter-level accuracy compared to the real system controlled by AR;
- Validation in real time by comparing the avatar tracked by ART with the avatar tracked by IMU + AI (analysis of graphical stability, initial alignment, and motion consistency over time).