Dataset: Driven By Sound - Automated Valet Parking

Abstract

Developing a safe automated vehicle navigation platform has been a research hotspot in automated driving. A key area in ensuring safe navigation is the perception systems that map environments and detect obstacles in real time. The Driven by Sound project, by Calyo, Benedex Robotics, and Cranfield University, focuses on building a functionally safe sensing platform using 3D ultrasonic sensors and Lidar for low-speed applications. Cranfield’s role includes developing a 3D mapping and obstacle-detection system using Lidar, a navigation planner, and a longitudinal and lateral controller. The project then integrates an ultrasonic detection into the Lidar map and explores ultrasonic sensors as a redundant sensor during Lidar failure. The project has successfully demonstrated an automated valet parking system, first validated using LiDAR and then replaced the LiDAR with ultrasonic sensors in a LiDAR-mapped car park.