CSC2125: Safety and Certification of Autonomous Vehicles

Winter 2018

Course Overview

---

Objectives

Before the dream of autonomous vehicles can be fully realized, we need to understand what their safety means and how to assure it. Yet, since autonomous vehicles typically include components constructed via machine learning, existing techniques and even existing standards are not directly applicable and need to be adjusted. In this course, we will explore some challenges in this field and survey recent work on safety and certification, specifically in the model-driven context. The application focus will be self-driving cars.

Assumed Background

Knowledge of at least one of the following: safety, autonomous driving, machine learning, testing and verification is strongly encouraged.

List of Topics

Autonomous driving, safety analysis (hazard analysis, assurance arguments, safety standards), design and synthesis, evidence generation via testing and verification adapted for ML-created systems.