| Autor: |
Julius F. Tilly, Jurgen Dickmann, Ole Schumann, Fabio Weishaupr, Gerd Wanielik |
| Rok vydání: |
2021 |
| Předmět: |
|
| Zdroj: |
2020 17th European Radar Conference (EuRAD). |
| DOI: |
10.1109/eurad48048.2021.00039 |
| Popis: |
A complete understanding of the environments encountered in road traffic is crucial for automated driving. Hereby, one major challenge involves robust classification of road users which might interact with the driving strategy. One of the sensors used to face this challenge is the radar sensor. Using a newly available automotive polarimetric radar, more information including shapes and orientations of geometries scattering the emitted electromagnetic waves can be obtained compared to conventional radars. This work evaluates the performance gain in road user classification tasks by using the additional polarimetric information to generate new features, which are given to a random forest classifier. In experiments on a real-world dataset a significant improvement in classification performance is demonstrated when polarimetric radars instead of conventional radars are utilized. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
