Perception and mapping for unmanned ground vehicles in unstructured environments

This research work presents a comprehensive study on enhancing the perception and mapping capabilities of Unmanned Ground Vehicles (UGVs) operating in unstructured environments. The key perception challenges addressed include effective ground segmentation, negative obstacle detection, and terrain traversability mapping, all crucial for enabling safe UGV navigation in complex terrains. First, this research work explores and implements state-of-the-art ground segmentation methods. Then, this work conducts theoretical analyses on LiDAR-based negative obstacle detection and proposes a novel detection and mapping method. Additionally, a multi-modal terrain traversability mapping method is explored and implemented in this work. This multi-modal method integrates elevation data and visual information for more accurate terrain traversability mapping. The proposed techniques are tested on several autonomous platforms, demonstrating potential improvements in perception and mapping. Overall, the preliminary results are promising and provide great insights into future research to enhance UGV’s perception capabilities in hazardous environments.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Mechanical Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Cheng, Weihao

Contributors dc:contributor

• Norris, William R
• Hovakimyan, Naira

• Autonomous Vehicles
• Perception
• Negative Obstacle Detection
• Sensor Fusion
• Lidar

Statement dc:rights

• Copyright 2024 Weihao Cheng

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/127459

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F127459