This invention describes a computer system that helps robots, especially large industrial machines, navigate safely through tight spaces without human help. It works by accessing information about obstacle locations, a desired path, and a collection of pre-defined movements, each with an associated cost. The system uses these costs, often represented in a cost map of the area, to choose the safest and most efficient way for the robot to move.
Why it matters: Filed before widespread advancements in AI-driven predictive modeling. The system's 'black box forward model' and real-time obstacle perception could now leverage more sophisticated and efficient machine learning techniques matured since 2021.
AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.
Reinvent this with AI