An integration framework for UGV outdoor navigation system based on LiDAR and vision data

Abstract

In this paper, we present an architecture to fuse different data from onboard sensors (LiDAR and Camera) for real-time navigation of Unmanned Ground Vehicle (UGV). The proposed system architecture comprises three main interacting modules; line and obstacle detection module, path planning and tracking module, and a real-time motor driving processing module. While these modules were developed individually on different platforms, however, the integration framework that we developed allowed these modules to coexist and interact seamlessly. The UGV prototype was put to the test in a variety of outdoor environments, both cluttered and uncluttered, to navigate safely within limited area bounded by lines. While the work on the proposed architecture is ongoing, this paper shows initial results of system platform development, modules integration framework, and real-time control accuracy and performance.