A benchmark collected by a fleet of Segway delivery robots in real world environment
Segway DRIVE benchmark is a novel and challenging benchmark collected by our fleet of Segway delivery robots deployed in real office buildings and shopping malls instead of the lab environment. Pedestrians walking in view of the camera, planer moving pattern of ground vehicle, and frequent environment/light changes will present new challenges to SLAM algorithms.Each Segway delivery robot is equipped with a global-shutter fisheye camera, a consumer grade IMU hardware-synced with the camera, two low-cost wheel encoders, and a high accuracy lidar for providing groundtruth values.Mutiple evaluation metrics has been provided as key indicators of algorithm performance on out dataset.