| GPENS: Ground Plane Estimation and Navigation System for Autonomous Mining Trucks |
| Eren Aydemir1,2, Mustafa Unel1 |
1Faculty of Engineering and Natural Sciences, Sabanci University, Istanbul 34956, Turkey
2Sancaktepe Engineering Center, Ford Otosan, Sancaktepe, Istanbul 34885, Turkey |
|
|
|
Received: November 4, 2024; Revised: December 10, 2024 Accepted: December 19, 2024. Published online: January 24, 2025. |
|
|
|
| ABSTRACT |
|
In this paper, we introduce the Ground Plane Estimation and Navigation System (GPENS) designed for autonomous mining trucks operating in unstructured environments. GPENS employs an efficient method for modeling unstructured terrain, which enhances the segmentation of 3D point clouds and improves the accuracy of ground versus non-ground classification. The integrated system facilitates real-time processing from ground plane estimation through to road boundary detection, motion planning, and control, enabling the autonomous navigation and parking of mining trucks. Additionally, we present a unique dataset comprising annotated point clouds, collected from a real mining area using an actual mining truck. Our state-of-the-art algorithms demonstrate a performance increase over current ground plane estimation solutions, achieving a 2% improvement in precision. Utilizing GPENS, we successfully showcase a truck-trailer combination capable of both navigating and parking autonomously in the challenging conditions of a real-world mine. |
| Key Words:
Autonomous driving · Truck–trailer combination · Ground plane estimation · Traversability estimation · Rough terrain |
|
Corresponding Author.
Mustafa Unel , Email. 
Preview 
Full text via DOI 
Download Citation
Print
