Mandatory Internship - Feature-based synthetic in-cabin data generation - REF74569M

Continental develops pioneering technologies and services for sustainable and connected mobility of people and their goods. Founded in 1871, the technology company offers safe, efficient, intelligent and affordable solutions for vehicles, machines, traffic and transportation. In 2023, Continental generated sales of 41.4 billion and currently employs around 200,000 people in 56 countries and markets. The increasing level of automation in vehicles is shifting more and more driving tasks from the human driver to the car itself. Future vehicle interiors will have various driver and occupant monitoring sensors, which can provide intelligent in-vehicle HMI interactions such as automatic adjusting of in-vehicle control systems based on motion sickness detection, playing mood light or music based on emotion prediction, or ensuring automation level compliance based on driver activity recognition. We require a vast in-cabin driver dataset to deliver such sophisticated systems. However, collecting more data for in-cabin activity recognition is expensive. Therefore, synthetic data generation is an economical alternative. Due to advancements in large language models (LLM), Generative AI methods are widely used for synthetic data generation. However, such methods are computationally expensive, cannot generate diverse data, and are biased toward training data. This work aims to generate synthetic in-cabin data without additional computational cost using data augmentation methods in feature space.