P25 - Road Traffic Safety Measures: Lab Testing on Collision Avoidance
- Event
- 22. GMA/ITG-Fachtagung Sensoren und Messsysteme 2024
2024-06-11 - 2024-06-12
Nürnberg - Band
- Poster
- Chapter
- Poster
- Author(s)
- N. Ravishankar, C. Cammin - Universität der Bundeswehr Hamburg,Hamburg
- Pages
- 520 - 527
- DOI
- 10.5162/sensoren2024/P25
- ISBN
- 978-3-910600-01-0
- Price
- free
Abstract
The increasing traffic volume necessitates robust collision avoidance. Modern vehicles employ advanced sensors like radar, LiDAR, and cameras for real-time data processing (5G collision avoidance). However, not all road users are equipped with such technology, especially vulnerable ones like pedestrians. The proposed approach for setting up a collision avoidance system involves utilizing a camera including internal object recognition capabilities. Additionally, the camera features functions allowing the definition and observation of specific regions as vehicles traverse them. In a laboratory setup, model vehicles follow predefined paths, and collision-prone regions along these paths are identified. The camera's extended functions, accessible via its dashboard, capture and analyze the behavior of moving objects within these regions. This approach enables a comprehensive examination of potential collision scenarios and facilitates the development and testing of collision avoidance algorithms. This paper provides an overview of utilizing the Bernard Mobility Analyzer (BMA) which is an intelligent optical system for identifying objects for monitoring, control, and guidance. It consists of a camera system that, supported by an internal controlling unit and artificial intelligence, recognizes objects, and only processes and transmits this anonymous information. The camera is deployed for functions such as trajectory prediction and collision avoidance. The paper covers the experimentation methodology, including challenges, along with settings for obtaining relevant data. Insights are given into the communication protocol, platforms used, and overall architecture. In conclusion, a simplified collision warning system could be set up on a laboratory scale.