3D Robot Interference Detection Using Dynamic Point Cloud Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing proximity sensors using electrostatic capacitance methods in collaborative robots often mistakenly detect robot accessories as interfering objects, leading to restricted movement and reduced production efficiency.
Innovation Solution
A robot system equipped with a three-dimensional sensor that generates point cloud data, distinguishes between robot components and interfering objects by generating a robot model, and calculates distances to detect approaching objects accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a proximity sensor using electrostatic capacitance method is used to detect interfering objects, then the robot can detect approaching objects for safety, but the sensor may mistakenly detect robot accessories and parts as interfering objects, causing restricted movement and reduced production efficiency
Solution Approach 1:
The patent segments the detection space into robot-occupied regions and free space by generating a robot model from point cloud data. The sensor controller divides the detected objects into robot parts (based on the model) and actual interfering objects, allowing accurate distinction between them. This segmentation enables the system to ignore robot accessories while detecting real obstacles.
Solution Approach 2:
The patent creates a virtual copy of the robot's three-dimensional structure by generating a robot model from point cloud data. This digital model serves as a reference to identify and filter out points corresponding to robot parts from the sensor data, preventing false detection of robot accessories as interfering objects.
2Reliability
If the robot controller memorizes positions of accessories and parts to prevent false detection, then false detection is reduced, but the robot's range of movement becomes restricted and production efficiency drops
Solution Approach 1:
The patent implements a dynamic robot model that updates in real-time based on the robot's current joint angles and link parameters. As the robot moves, the model automatically adjusts to reflect the new positions of robot parts, allowing the robot to operate throughout its full range of motion without restricted zones. The detection system dynamically adapts to the robot's changing configuration.
Solution Approach 2:
The system changes the parameter representation from static memorized positions to dynamic parameters including joint angles, link parameters, and real-time point cloud data. This allows the robot to maintain accurate self-identification across all positions and orientations, enabling full adaptability without sacrificing detection reliability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The object of the present invention is to distinguish between a robot and an interfering object and to detect an approach of the interfering object. A controller (4) of a robot system includes a link parameter storage part (62) for storing link parameter values of the robot, a point cloud data acquisition part (63) for acquiring point cloud data, which is a set of points indicating a position of a surface of an object placed in the space, from the three-dimensional sensor, a joint angle acquisition part (64) for acquiring a current joint angle of each joint of the robot arm, a robot model generator (65) for generating a robot model that shows posture that is similar to current posture of the robot based on the link parameter and the joint angle, an interfering object data setting part (66) for removing at least points indicating the robot from the point cloud data by using the robot model and setting the remaining point cloud data as interfering object data, which is a set of points indicating a position of an interfering object, and an approach determination part (67) for calculating a distance between the interfering object and the robot by using the interfering object data and the robot model and determining whether the interfering object is approaching the robot or not based on the distance between the interfering object and the robot.