Learn how to localize and track dynamic objects with a range of applications including autonomous vehicles. Autonomous vehicles, such as self-driving cars, rely critically on an accurate perception of their environment. In this course, we will teach you the fundamentals of multi-object tracking for automotive systems.
Class Deals by MOOC List - Click here and see EdX's Active Discounts, Deals, and Promo Codes.
Key components include the description and understanding of common sensors and motion models, principles underlying filters that can handle varying number of objects, and a selection of the main multi-object tracking (MOT) filters.
The course builds and expands on concepts and ideas introduced in CHM013x: “Sensor fusion and nonlinear filtering for automotive systems”. In particular, we study how to localize an unknown number of objects, which implies various interesting challenges. We focus on cameras, laser scanners and radar sensors, which are all commonly used in vehicles, and emphasize on situations where we seek to track nearby pedestrians and vehicles. Still, most of the involved methods are more general and can be used for surveillance or to track, e.g., biological cells, sports athletes or space debris.
The course contains a series of videos, quizzes and hands-on assignments where you get to implement several of the most important algorithms.
Learn from award-winning and passionate teachers to enhance your knowledge at the forefront of research on self-driving vehicles. Chalmers is among the top engineering schools that distinguish itself through its close collaboration with industry.
This course is part of the Emerging Automotive Technologies MicroMasters Program and part of the Sensor Fusion and Multi-Object Tracking Professional Certificate.
What you'll learn
- A thorough understanding of multi-object tracking (MOT) and its challenge
- Expert-level understanding of principles, theory and algorithms in modern MOT.
- Extensive know-how for solving various MOT problems in practice.
- Valuable experience from implementing different MOT algorithms.
