- Titel
- Evaluation of Methods for Scenario-based Assessment of ADAS through Simulations
- Zitierfähige Url:
- https://nbn-resolving.org/urn:nbn:de:bsz:ch1-qucosa2-1031644
- Datum der Einreichung
- 25.06.2024
- DOI
- https://doi.org/10.60687/2026-0064
- Abstract (EN)
- Over the previous 20 years, there have been notable developments in the field of advanced driver assistance systems (ADAS). The primary objective of ADAS development is to increase road traffic safety and reduce the accident rate. This will require the observation and analysis of numerous traffic scenarios in real-world environments, as well as the development of an accident prevention system. In the actual world, a lot of unknown corner circumstances or potentially dangerous cases are tough to evaluate. It has been suggested that scenario-based assessment be used to focus testing difficult and important scenarios in a shorter amount of time and for less money. In virtual simulation environments, scenario-based assessment is used to analyze the response of the system being tested according to many criteria.Through the use of quasi-monte carlo and monte carlo techniques for parameter modifications, the scenario integrates unpredictability found in the real world. This research work demonstrates how an assessment concept based on scenarios might alleviate the impractical number of real-world evaluation that a traditional concept would require. As a result, a technique is developed and turned into a unique proof of concept framework. It comprises virtual testing, parameter variation to produce unpredictability, and actor behavior analysis in the scenario. By simulating scenarios and using sampling techniques like Monte Carlo and quasi-Monte Carlo, we may examine the unpredictable behavior of traffic participants. In order to facilitate the development, training, and evaluation of automated urban driving systems, scenarios are developed in the CARLA simulator. Python APIs are used in the development of the scenarios. Robot Operating System (ROS) along with ROS Bridge provide assistance for this implementation in terms of data collecting from scenarios and additional behavior analysis. This thesis work adds to the evaluation of several approaches of evaluating real-world scenarios in a virtual environment. The majority of the research work focuses on situations where the car’s ADAS is used.
- Freie Schlagwörter (EN)
- Scenario-based Simulation, Monte carlo, Quasi Monte carlo, Scenario Creation, CARLA Simulator
- Klassifikation (DDC)
- 000
- Normschlagwörter (GND)
- Assistenzsystem, Straßenverkehr, Simulation
- BetreuerIn Hochschule / Universität
- Prof. Dr. Wolfram Hardt
- Den akademischen Grad verleihende / prüfende Institution
- Technische Universität Chemnitz, Chemnitz
- Version / Begutachtungsstatus
- angenommene Version / Postprint / Autorenversion
- URN Qucosa
- urn:nbn:de:bsz:ch1-qucosa2-1031644
- Veröffentlichungsdatum Qucosa
- 18.03.2026
- Dokumenttyp
- Masterarbeit / Staatsexamensarbeit
- Sprache des Dokumentes
- Englisch
- Lizenz / Rechtehinweis
