Matthew Doude

Papers

Total Citations

H-Index

About

Matthew Doude’s research lies at the intersection of autonomous ground vehicles, simulation environments, and machine learning, with a particular focus on off-road and unstructured terrain. His major contributions center on developing scalable training methods for neural networks used in vehicle autonomy. In his most-cited work, “Training of Neural Networks with Automated Labeling of Simulated Sensor Data” (2019, 12 citations), Doude introduced a novel approach that eliminates the costly, time-intensive process of manual data labeling by leveraging simulated sensor data to automatically generate ground-truth labels. This breakthrough significantly accelerates the training pipeline for convolutional neural networks (CNNs) in autonomous systems. Doude further advanced the field through his 2020 study “Exploring the Requirements and Capabilities of Off-Road Simulation in MAVS and GAZEBO,” which systematically evaluated simulation platforms for developing unmanned ground vehicles (UGVs) in challenging off-road conditions. His work directly addresses the critical need for realistic, repeatable testing environments that are difficult to recreate physically. By bridging the gap between simulation and real-world deployment, Doude’s research has provided foundational tools for researchers and engineers working to make autonomous vehicles safer and more capable in complex, off-road settings.