Determination of Task Plans for Robotic Devices

This invention describes a way for robots to figure out what to do. It takes human instructions and translates them into a precise, formal language that a computer can understand, then uses a learning system to create many possible action plans. The robot also builds a statistical picture of its surroundings from its sensors and compares its potential plans to this real-world model, using statistical methods like Bayesian inference, to pick the most likely successful plan.

Why it matters: Given the filing date of 2025, the core challenge of robustly integrating formal methods with probabilistic machine learning for real-time robotic task planning likely remains. A single year is insufficient for fundamental breakthroughs in such a complex, interdisciplinary field.

Status: Under examinationA pending application. It has not been granted, and it has not been abandoned.
How hard to build: SpecializedComplex software system, ML, formal logic

3 AI reinvention directions for this patent

See all three

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

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10