[RA-L '24] Learning self-supervised traversability with navigation experiences of mobile robots: A risk-aware self-training approach

Based on TJ's Thesis Research.

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.