patent.fyi
Applications  /  US 2026/0355955
United States Patent Application Publication

US 2026/0355955 A1

Published
Nov 12, 2026
Filed
Apr 21, 2026
Earliest date on the face
May 8, 2025
Classification
B25J 9/1661
First published here
Nov 19, 2026

Generated automatically. Not reviewed by a person. Not legal advice or an opinion.

Section 1 · machine-generated

What if?

Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.

No entries.

Section 2 · machine-generated

Variations

Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.

Other field of use

Symbolic planning for library returns

A tabletop gantry uses a barcode reader and a gripper to move returned books from an intake tray to sorting bins. A domain generator converts textual descriptions of scan, pick and place actions into preconditions and effects. State generators parse text that lists the books in the intake tray and specifies the destination bin for each book. A symbolic planner orders the moves, and the gantry controller checks the barcode and bin occupancy before each placement.

Component substitution

Planning with breadth first search

A breadth first search engine replaces the symbolic planner for a benchtop fixture that moves parts among a feeder, a clamp and an output tray. The three generators convert text into action preconditions and effects, initial facts and target facts. The engine starts at the initial state, adds each reachable successor to a first in, first out queue and records the action and parent state for each previously unseen state. When a state satisfies the target facts, the engine traces its parent links, reverses the action sequence and sends it to the fixture controller.

Parameter or scale variation

Staged planning across forty eight bins

A transfer cell has 48 bins arranged in four banks of 12, a six-position staging tray and two manipulators that share access to the tray. A domain generator converts text into pick and place actions with reachability, occupancy and tray reservation preconditions. The state generators parse text listing the occupied locations and the destination bin for each part, and the symbolic planner produces batches containing up to six cross-bank transfers. The controller releases each reserved tray position after a presence sensor confirms that the receiving manipulator has lifted the part.

Combination with existing systems

Planning with a machine tending controller

A symbolic planning service connects to a programmable controller that coordinates a manipulator and a milling machine. The domain generator maps textual descriptions of loading, clamping, machining and unloading to actions backed by controller routines. The state generators parse the current cell status and a work order that specifies how many finished parts belong in the output rack. The planning service sends actions one at a time and waits for completion signals, while the programmable controller enforces the door, spindle and clamp interlocks.

Automated or data-driven variant

Camera generated initial state descriptions

An overhead camera observes a parts tray and an assembly fixture at fixed, calibrated locations. Image processing software reads fiducial markers on the parts and fixture and writes a textual description of part identities, locations and empty slots. The initial state generator parses this text, while separate generators convert the textual action descriptions and assembly instructions into domain and target models. After each placement, the software captures another frame and requests a new plan if the observed state differs from the predicted state.

Cycle time estimates from execution logs

A controller logs the start and completion time of every pick, transfer and place action along with the part class and travel distance. After each shift, software calculates the median duration for each action, part class and distance band, using the nominal duration from the textual domain description for combinations with fewer than five observations. The domain generator attaches the resulting durations as action costs, and the state generators parse textual descriptions of the current and required bin contents. A cost-based symbolic planner selects a sequence with the lowest total estimated duration, and the controller executes it using completion acknowledgments to advance between actions.

Cite aspatent.fyi, US 2026/0355955 A1, claim 1, first published 2026-11-19T17:33:15-05:00, https://patent.fyi/US20260355955/, SHA-256 98813f5734f30fe26ea010dc15982c5cb608a903e6045d2749b79da9060b05fb
First published
Permanent addresshttps://patent.fyi/US20260355955/
Content fingerprint (SHA-256)98813f5734f30fe26ea010dc15982c5cb608a903e6045d2749b79da9060b05fb
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi