Robot Path Planning Using Human Skill Waypoints in Obstacle Fields

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing path generation techniques for industrial robots in complex obstacle environments are inefficient, error-prone, and fail to capture human intuition and collision avoidance skills effectively.

Innovation Solution

A method using a 3D convolutional neural network to extract obstacle features, combined with an encoder/decoder neural network system trained on human-generated motion programs, generates a distribution of waypoints that are refined using RRT or optimization-based techniques to ensure a collision-free path.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If teach pendant is used for path generation, then robot can be instructed to make incremental moves, but the process becomes difficult, error-prone and time-consuming in complicated obstacle environments

Engineering Contradiction:

Improveease of path programmingVSAvoidtime-consuming programming

Core Design Contradiction:

Ease of operationVSLoss of time

Solution Approach 1:

The system captures human operator movements and automatically generates robot motion programs by recording and processing the operator's lead-through motions, converting manual demonstration into automated path programs without requiring manual programming of each incremental move

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces manual mechanical teaching operations with an automated system that uses sensors, processors, and algorithms to generate motion programs, substituting the mechanical teach pendant interface with an automated path generation system that processes obstacle data and computes optimal paths

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If lead-through process is used, then human expertise in path selection is captured, but it may be difficult or impossible to manipulate the entire robot to avoid collisions in complex obstacle environments

Engineering Contradiction:

Improvehuman expertise captureVSAvoiddifficulty of robot manipulation

Core Design Contradiction:

Adaptability or versatilityVSEase of operation

Solution Approach 1:

The system separates the path planning function from the robot manipulation function by using automated algorithms to compute collision-free paths based on captured human expertise and obstacle data, eliminating the need for manual manipulation of the entire robot while preserving the benefits of human path selection intuition

Inventive Principle:

Principle #1Segmentation

3Ease of operation

If human demonstration is used for path generation, then path selection is simplified, but positional accuracy for precise movement is insufficient and robot arm collision avoidance is not accounted for

Engineering Contradiction:

Improvepath generation simplicityVSAvoidpositional accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system incorporates feedback mechanisms that evaluate generated paths for collision avoidance and positional accuracy, using obstacle data and robot kinematics to verify and refine the paths generated from human demonstration, ensuring both simplicity and precision are achieved

Inventive Principle:

Principle #23Feedback

4Extent of automation

If automatic path generation computation is attempted, then path computation is automated, but existing techniques fail to capture human intuition and collision avoidance skills effectively

Engineering Contradiction:

Improveautomation of path computationVSAvoidhuman skill capture

Core Design Contradiction:

Extent of automationVSAdaptability or versatility

Solution Approach 1:

The patent introduces human lead-through demonstrations as an intermediary that bridges automated computation and human intuition, where human operators demonstrate desired paths that are then processed by automated algorithms to capture collision avoidance skills and generate adaptable motion programs for various obstacle environments

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS12576516B2Human skill based path generation

Publication Date: 2026.03.17 FANUC LTD

AI summary

A method for robot path planning using skills extracted from human-taught motion programs applied to a new obstacle environment. A three-dimensional convolutional neural network is used to extract features characterizing an obstacle environment, where the feature vector representation of the obstacles overcomes problems encountered when using point cloud obstacle data. The obstacle feature data and robot path start and goal points are provided to an encoder/decoder neural network system which is trained to extract skills from a database of human-generated motion programs. The encoder/decoder neural network system produces a distribution of waypoints for the current obstacle environment and start/goal points. The distribution of waypoints is used to perform a final collision-free path generation using either a rapidly-exploring random tree (RRT) technique or an optimization-based technique.