Amazon patents robot arm that adjusts grip if objects swing
This patent describes a method where a robotic arm uses a machine learning model to analyze sensor data at two different times. By comparing how the object appears in these snapshots, the system calculates a 'swinging score' and adjusts the robot's movement plan to handle the object more stably.
In the filing’s own words
Techniques described herein relate to swing anomaly detection for objects transferred via robotic manipulation. For example, a computer system can generate, based on first sensor data of a robotic manipulator manipulating an object and by using a machine learning model, a first segmentation of the object represented by the first sensor data. The first sensor data can be received at a first time. The computer system can generate, based on second sensor data of the robotic manipulator manipulating the object and by using the machine learning model, a second segmentation of the object represented by the second sensor data. The computer system can generate, by comparing the first segmentation and the second segmentation, a swinging score indicating swinging of the object by the robotic manipulator. A motion plan for the robotic manipulator can be updated based on the swinging score.
What exactly they claim (claim 1)
1. A computer-implemented method, comprising: generating, based at least in part on first sensor data corresponding to a robotic manipulator manipulating an object and by at least using a machine learning model, a first segmentation of the object represented by the first sensor data, the first sensor data being received at a first time; generating, based at least in part on second sensor data corresponding to the robotic manipulator manipulating the object and by at least using the machine learning model, a second segmentation of the object represented by the second sensor data, the second sensor data being received at a second time subsequent to the first time; generating, based at least in part on an area of intersection between the first segmentation and the second segmentation, a swinging score indicating a degree of swinging of the object by the robotic manipulator; updating a motion plan for the robotic manipulator manipulating the object based at least in part on the swinging score; and causing the robotic manipulator to execute the updated motion plan.
Facts
- Company
- Amazon (Amazon Technologies, Inc.)
- Official title
- Swing anomaly detection for robotic manipulation of objects
- Patent number
- US12734687B1
- Application number
- 18978595
- Filed
- Dec 12, 2024
- Granted
- Sep 15, 2026
- Inventors
- 5
- Claims
- 20
Classified under
B25J 9/163 Operations & transportB25J 9/1664 Operations & transportB25J 9/1674 Operations & transportB25J 9/1697 Operations & transportMore from Amazon in Robots & automation
- Amazon patents software that designs custom robots for specific tasks Sep 8, 2026
- Amazon patents robots that move items to custom box makers Sep 8, 2026
- Amazon patents robots to move boxes between warehouse floors Sep 8, 2026
- Amazon patents robot system to slide storage pods for item retrieval Sep 15, 2026
- Amazon patents robot gripper that lifts items from crowded bins Sep 22, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.