John Williams, CTO and co-founder of Kudan, presents the "Commercial Grade SLAM Frameworks for Indoor and Outdoor Applications" tutorial at the May 2019 Embedded Vision Summit.
SLAM (simultaneous localization and mapping) is an essential technology for any device that requires an understanding of its location and orientation in the physical world. Kudan is the only publicly listed SLAM algorithms vendor in the world licensing commercial grade SLAM frameworks. In this talk, Williams briefly introduces SLAM and explains why it is superior to alternative approaches. He explores the different types of sensors (including cameras, IMUs, LiDAR, GPS, etc.) that can be used for SLAM, and how they can be combined for maximum benefit.
Williams then explains the advantages of accelerating the execution of SLAM algorithms, and approaches for doing so on various types of embedded and cloud processors. Finally, he describes Kudan’s licensable SLAM technology and customization services, and illustrates the practical use of this technology in a variety of real-world products and environments, including cars, robots, drones and AR/VR devices.