Autonomous Vehicles Recognise Terrains
A new algorithm has been developed by the California Institute of Technology (Caltech) that allows autonomous vehicle systems to recognise where they are simply by looking at the terrain around them. This new algorithm works regardless of seasonal changes to the terrain. The present visual terrain-relative navigation (VTRN) system depends on close matches to the images in its database to identify terrains. Anything that alters or obscured the terrain, such as snow cover or fallen leaves, causes the images do not to match up and fouls up the system. With the use of deep learning and artificial intelligence (AI), Caltech scientists have been able to remove seasonal content that hinders current VTRN systems. It uses a self-supervised learning technique and the AI looks for patterns in images by teasing out details and features that would likely be missed by humans.