Loading [a11y]/accessibility-menu.js
Real-time autonomous ground vehicle navigation in heterogeneous environments using a 3D LiDAR | IEEE Conference Publication | IEEE Xplore

Real-time autonomous ground vehicle navigation in heterogeneous environments using a 3D LiDAR


Abstract:

The ability to drive autonomously in heterogeneous environments without GPS, pattern identification (e.g. road following), or artificial landmarks is key to field robotic...Show More

Abstract:

The ability to drive autonomously in heterogeneous environments without GPS, pattern identification (e.g. road following), or artificial landmarks is key to field robotics. To address this challenge, we present a complete waypoint navigation framework for unmanned ground vehicles. A Velodyne PUCK VLP-16 LiDAR and an IMU are mounted on an autonomous, full size utility vehicle and used for localization within a previously created base map. We redesign a six degrees of freedom LiDAR SLAM algorithm to achieve 3D localization on the base map, as well as real-time vehicle navigation. We fuse the low-frequency, high precision SLAM updates with high-frequency, odometric local state estimates from the vehicle. The navigation costmap consists of a 2D occupancy grid which is computed from the 3D base map. Relying on this setup, the vehicle is capable of navigating through a complex site completely autonomously. The test site has densely and sparsely built areas, bushland, industrial activities, pedestrians, and other manned or unmanned vehicles. Extensive testing was done using both current and outdated base maps for comparisons, and a high precision RTK-GPS was used for ground truth. So far, more than 60 km of completely autonomous driving has been performed without a single system or navigation failure.
Date of Conference: 24-28 September 2017
Date Added to IEEE Xplore: 14 December 2017
ISBN Information:
Electronic ISSN: 2153-0866
Conference Location: Vancouver, BC, Canada

References

References is not available for this document.