Master thesis scope

One of the ways to demonstrate the safety benefit of automated driving (AD) is comparing the performance of the AD function with a competent and careful human driver, namely a reference driver. The UN ECE R157 ALKS (Automate Lane Keeping System) regulation defined a driver model and asked for comparison. The model is the result of 20+ years research on accident prevention and traffic flow dynamics. 

However, more and more models have been presented in the recent years targeting ALKS verification, like BMW, VW. In 2022, Waymo presented a Non-Impaired road user with their Eyes ON the conflict (NIEON) model as the reference response time model for evaluating the collision avoidance functions on their Robotaxi service (SAE L5).

This project is to implement a reference driver model based on existing models and set the parameters according to collected data. The model needs to run in the simulations to compare with AD functions. The scope of the project including:

1. Literature review
2. Data analysis
3. Model implementation and validation
4. Comparing the performance of AD function and Reference Driver Model
5. Write a technical report
   

Qualifications

• Good modeling skills and knowledge of driver modeling is a merit
• Basic knowledge on scenario-based AD verification
• Have experience with simulation tools like Carla, IPG CarMaker, or others
• Good programing skills on Python, C/C++, Matlab, etc.

Where? Göteborg

When? 2023

For how long? 6 – 9 months

Send your questions to: Jianlin Shi jianlin.shi@huawei.com