3D Camera Robot Control for Unexpected Moving Object Avoidance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for controlling robots in industrial environments to prevent collisions with moving objects are complex and inefficient, often requiring significant setup and limiting task flexibility due to the need for physical separation or complex sensor systems.

Innovation Solution

A system using a three-dimensional camera to identify and track moving connected objects, comparing frames to detect unexpected objects and control robot motion by slowing or stopping the robot based on their proximity to these objects, thereby simplifying collision avoidance without adding unnecessary complexity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If physical separation or complex sensor systems are used to prevent robot collisions, then safety is improved, but device complexity and setup effort increase significantly

Engineering Contradiction:
ImprovesafetyVSAvoidsetup effort
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent replaces complex mechanical sensor systems (LIDAR, pressure mats) and physical barriers (cages) with a vision-based system using standard cameras. The camera captures images that are processed to generate depth maps and identify moving connected objects, substituting sophisticated mechanical sensing with optical detection and computational processing.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system creates a virtual representation of the physical workspace by generating depth maps from camera images. These depth maps serve as a digital copy of the three-dimensional space, allowing the robot to perceive and respond to objects without requiring complex physical sensors. The depth map copying approach simplifies the sensing architecture while maintaining safety capabilities.

Inventive Principle:
Principle #26Copying

2Reliability

If physical separation or isolated locations are required for robot workspaces, then collision safety is improved, but productivity and task flexibility are reduced

Engineering Contradiction:
Improvecollision safetyVSAvoidtask flexibility
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent replaces physical isolation barriers (cages, exclusion zones) with a vision-based detection system. Standard cameras capture the workspace environment, and image processing algorithms identify moving connected objects in real-time, allowing robots to operate safely in shared spaces without physical separation.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system introduces an intermediary processing layer between the camera and robot controller. Image data is processed to generate depth maps and identify moving connected objects, creating a virtual safety mechanism that mediates between the robot and human workers, eliminating the need for physical barriers while maintaining safety.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If complex sensor systems like LIDAR or pressure mats are deployed, then detection accuracy is improved, but device complexity and cost increase

Engineering Contradiction:
Improveobject detection accuracyVSAvoidsensor system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent substitutes sophisticated mechanical sensors (LIDAR, pressure mats) with standard vision cameras. By processing camera images through depth map generation and moving connected object detection algorithms, the system achieves comparable detection accuracy using simpler, more cost-effective optical sensors.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system creates a digital depth map copy of the physical workspace from standard camera images. This computational approach to depth perception allows the robot to achieve three-dimensional object detection accuracy previously requiring complex sensors, using only standard vision hardware and image processing.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP3715065B1Controlling a robot in the presence of a moving object
Publication Date: 2022.01.26 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3715065B1 patent drawingFigure 1
  • EP3715065B1 patent drawingFigure 2
  • EP3715065B1 patent drawingFigure 3

AI summary

A method, system, and one or more computer-readable storage media for controlling a robot (102) in the presence of a moving object (112) are provided herein. The method includes capturing a number of frames from a three-dimensional camera system (104) and analyzing a frame to identify a connected object. The frame is compared to a previous frame to identify a moving connected object MCO. If an unexpected MCO is in the frame a determination is made if the unexpected MCO is in an actionable region. If so, the robot (102) is instructed to take an action.