We are focusing on developing the latest Advanced Driver Assistance System (ADAS) technologies to save lives on the road. To realize these functions, it is crucial to construct a high-quality representation of the vehicle's real-world environment. This can be done using sensor fusion, which consists of steps such as detection, tracking and fusion of surrounding objects in time and space. Alignment of these objects may also be necessary to further refine the detections, which are inherently noisy and imprecise. These steps can done using traditional algorithms and utilizing machine learning as well.
The task is to explore the possibilities of replacing some classical steps of sensor fusion with machine learning. To do this, you will research state-of-the-art neural networks, define interfaces, define heuristics, generate training data, inference said networks and evaluate performance.

The goal is to implement a Proof of Concept (PoC) solution for the clustering and alignment of objects (i.e. parking slots) that can be represented as 4-point polygons using machine learning.