Celebrating innovation: Cihan Yurtsever's groundbreaking master’s thesis on autonomous driving safety, “Development of Autonomous Driving Control Algorithms Based on ASIL Scenarios. Safety-Centric Evaluation of Control Algorithms in CARLA Simulated Environments”
We’re thrilled to announce that Cihan Yurtsever, one of our talented interns has successfully completed a master’s thesis that delves into the cutting-edge realm of autonomous driving systems. This significant achievement not only underscores our commitment to fostering innovation but also highlights the exceptional capabilities of our team in driving the future of automotive technology.
Transforming the automotive sector with autonomous systems
The rapid advancement of autonomous systems is revolutionizing various industries, with the automotive sector at the forefront of this transformation. This disruptive technology is reshaping conventional paradigms of transportation and mobility, promising unprecedented levels of efficiency, comfort and safety on our roadways.
A systematic approach to autonomous driving
Cihan Yurtsever's thesis, titled "Development of Autonomous Driving Control Algorithms Based on ASIL Scenarios," focuses on implementing autonomous driving functionalities through a comprehensive approach that encompasses perception, motion planning, control and actuation. The primary objective of the study is to evaluate the efficacy and outcomes of different control and motion planning algorithms, with a critical emphasis on the safety of autonomous driving systems.
Adherence to ASIL standards for safety
To ensure rigorous safety assessment, the study adheres to the Automotive Safety Integrity Level (ASIL) standards outlined in ISO 26262. Hazardous scenarios identified through Hazard Analysis and Risk Assessment (HARA) are used to compare control algorithms of varying complexities, specifically PID and MPC. These scenarios are simulated in both city and rural settings within the CARLA Simulation environment, leveraging its advanced capabilities for a realistic evaluation.
Technical implementation
Path planning and perception tasks are facilitated by the CARLA Simulator API, utilizing both camera and obstacle sensors. The car model is based on the bicycle model, with motion planning incorporating adaptive cruise control. The project is developed in Python, adhering to PEP8 standards and executed on the Ubuntu operating system. Code quality and readability are ensured through Pylint linting analysis, with logging and output tracking maintained in log and CSV file formats to store operational data.
Practical exploration and comparative analysis
The thesis culminates in a practical exploration of real-life scenarios within the CARLA simulator, followed by a comparative analysis of the outcomes. Critical safety parameters, including cross-track and heading errors, are evaluated to ensure ISO 26262 safety assurance through HARA.
Looking ahead
This remarkable achievement not only showcases the depth of expertise within our team but also reinforces our commitment to pioneering advancements in autonomous vehicle technology. We are immensely proud of Cihan Yurtsever's dedication and contributions to this pivotal field and look forward to further innovations that drive us towards a safer and more efficient future.