Series Sensor Data Processing and International Data Comparison in Automated Driving Safety Assessment
- Delivery
- Available on this site
- Format
- Price
- Non-members (tax incl.):¥1,100 Members (tax incl.):¥880
- Publication code
- 20219072
- Paper/Info type
- Other International Conferences
- Pages
- 1-6(Total 6 p)
- Date of publication
- Sep 2021
- Publisher
- JSAE
- Language
- English
Detailed Information
Author(E) | 1) Silvia Thal, 2) Jannes Iatropoulos, 3) Adrian Sonka, 4) Roman Henze, 5) Ryo Hasegawa, 6) Hiroki Nakamura, 7) Jacobo Antona-Makoshi, 8) Satoshi Taniguchi |
---|---|
Affiliation(E) | 1) Institute of Automotive Engineering, Technische Universität Braunschweig,, 2) Institute of Automotive Engineering, Technische Universität Braunschweig,, 3) Institute of Automotive Engineering, Technische Universität Braunschweig,, 4) Institute of Automotive Engineering, Technische Universität Braunschweig,, 5) Japan Automobile Research Institute, 6) Japan Automobile Research Institute, 7) Japan Automobile Research Institute, 8) Toyota Motor Corporation |
Abstract(E) | Collecting naturalistic driving data, detecting defined scenarios and analyzing statistical representative driving behavior is one of the main tasks in the scenario-based approach for safety assessment of automated driving systems. The scenario-based approach requires large, representative datasets which can be best obtained with test vehicles equipped with series environment sensors as radar and mono camera. The lower resulting data quality of series sensors require extensive data processing steps, which are not handled in related work. This study contributes in three ways. First, required processing steps and setting parameters for radar data are presented. Second, by using methods from Design of Experiments, a methodology for an efficient assessment of suitable setting parameters is presented. Thirdly, the subsequent processed datasets are input for a data comparison of cut-in scenarios from Germany and Japan. The differences revealed in the parameter distributions substantiate the importance of considering country-specific traffic characteristics in the harmonization of international safety assessment for automated driving. |