This invention describes a method for robots to grasp objects. The robot uses sensors to identify objects in its environment and creates a digital blueprint, or "virtual model," for a specific object, including its fixed characteristics like shape and weight. It then estimates the object's current position and orientation, and uses both the blueprint and this real-time estimate to figure out the best way to reach out and pick up the object. The claims narrow this to estimating an object's weight based on its detected class.
Why it matters: Filed before recent leaps in AI for perception and planning. Modern foundation models can now significantly improve object detection, modeling, and grasp trajectory generation.
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