Visual Navigation technology consists of simultaneous localization/mapping module and AI object detection module. The main features are obstacle detection on self-driving roads and the robustness of localization based on dynamic object recognition. The image sensor is monocular/stereo camera on the vehicle for image capture and feature matching. In order to robust localization and mapping, the camera pose estimation is computed by 3D geometry and environmental information.

1.Image resolution: 720p
2.Operating system: Linux Ubuntu 16.04
3.Frame rate: > 10fps
4.Absolute trajectory error: < 1m
5.Relative trajectory error: < 0.8m
6.Interface: CAN/Ethernet

Autonomous driving, Autonomous mobile robots

Automated valet parking, Logistic robots , Intelligent medical healthcare