17950021. ROBOTIC ASSEMBLY INSTRUCTION GENERATION FROM A VIDEO simplified abstract (Accenture Global Solutions Limited)

ROBOTIC ASSEMBLY INSTRUCTION GENERATION FROM A VIDEO

Organization Name

Accenture Global Solutions Limited

Inventor(s)

Kumar Abhinav of Hazaribag (IN)

Alpana Dubey of Bangalore (IN)

Shubhashis Sengupta of Bangalore (IN)

Suma Mani Kuriakose of Mumbai (IN)

Priyanshu Abhijit Barua of Pune (IN)

Piyush Goenka of Bangalore (IN)

ROBOTIC ASSEMBLY INSTRUCTION GENERATION FROM A VIDEO - A simplified explanation of the abstract

This abstract first appeared for US patent application 17950021 titled 'ROBOTIC ASSEMBLY INSTRUCTION GENERATION FROM A VIDEO

Simplified Explanation

The abstract describes a patent application for a system where a robot host receives a video of an assembly, analyzes spatio-temporal features to identify actions, maps actions to sub-objects to create an assembly plan, and generates instructions for robotic machines based on this plan.

• The system uses a video to analyze assembly actions and generate an assembly plan.
• It combines output from a point cloud model and a color embedding model to generate coordinates for sub-objects.
• Object segmentation is performed to estimate grip points and widths for sub-objects.

Potential Applications

This technology could be applied in industries such as manufacturing, logistics, and construction for automated assembly processes.

Problems Solved

This technology streamlines the assembly process by automating the generation of assembly plans and instructions based on video analysis.

Benefits

The benefits of this technology include increased efficiency, accuracy, and consistency in assembly processes, as well as reducing the need for manual labor.

Potential Commercial Applications

One potential commercial application of this technology could be in the development of robotic assembly systems for various industries.

Possible Prior Art

One possible prior art for this technology could be existing robotic systems that use computer vision for object recognition and manipulation tasks.

How does the system handle variations in assembly tasks and objects?

The system's ability to adapt to different assembly tasks and objects is not explicitly mentioned in the abstract. It would be interesting to know how flexible the system is in accommodating various scenarios.

What is the accuracy and reliability of the system in generating assembly plans and instructions?

The abstract does not provide information on the accuracy and reliability of the system in generating assembly plans and instructions. Understanding the system's performance metrics would be crucial for assessing its practicality and effectiveness.

Original Abstract Submitted

In some implementations, a robot host may receive a video associated with assembly using a plurality of sub-objects. The robot host may determine spatio-temporal features based on the video and may identify a plurality of actions represented in the video based on the spatio-temporal features. The robot host may map the plurality of actions to the plurality of sub-objects to generate an assembly plan and may combine output from a point cloud model and output from a color embedding model to generate a plurality of sets of coordinates corresponding to the plurality of sub-objects. The robot host may perform object segmentation to estimate a plurality of grip points and a plurality of widths corresponding to the plurality of sub-objects. Accordingly, the robot host may generate instructions, for robotic machines, based on the assembly plan, the plurality of sets of coordinates, the plurality of grip points, and the plurality of widths.