Learning robot-specific traversability from a short manual drive — Zero annotations, beyond driven trajectories. [RA-L '24]
-
Updated
Sep 4, 2026 - C++
Learning robot-specific traversability from a short manual drive — Zero annotations, beyond driven trajectories. [RA-L '24]
Modular ROS 2 autonomy project for the Clearpath Husky UGV, covering simulation, sensor integration, localization, perception, terrain-aware navigation, and skid-steer control.
A deep learning–based system for Terrain Relative Navigation (TRN) that detects lunar craters using a YOLOv8 neural network and estimates crater depth and inter-crater distances for spacecraft localization.
Autonomous hexapod rover using Python Raspberry Pi control, Arduino/C++ sensor firmware, Buehler-clock gait control, and simulation-backed terrain navigation.
To associate your repository with the terrain-navigation topic, visit your repo's landing page and select "manage topics."