Collision Detection of a HEXA Parallel Robot Based on Dynamic Model and a Multi-Dual Depth Camera System

Area of Science:

• Robotics
• Mechatronics
• Computer Vision

Background:

• Parallel robots like the Hexa robot offer unique kinematic advantages but require robust safety mechanisms.
• Obstacle collision detection is critical for autonomous operation and preventing damage in robotic systems.

Purpose of the Study:

• To introduce an integrated method for obstacle collision detection, isolation, and identification for the Hexa parallel robot.
• To leverage dynamic modeling and a computer vision system for real-time collision analysis.

Main Methods:

• Analysis of Hexa robot configuration, kinematics, and dynamics during trajectory execution.
• Development of a virtual force sensor using inverse dynamics and a low-pass filter for collision signal estimation.
• Implementation of a dual-depth camera vision system for obstacle isolation and contact point determination.
• Application of the recursive Newton-Euler algorithm to compute collision contact forces.

Main Results:

• The proposed method successfully detects and identifies collision events on the Hexa robot.
• Experimental validation demonstrated the accuracy of the computed contact forces compared to sensor data.
• The vision system effectively localized contact points on robot components.

Conclusions:

• The integrated dynamic modeling and computer vision approach provides an effective solution for Hexa parallel robot collision issues.
• The developed method enhances the safety and reliability of parallel robotic systems.
• This research contributes to the advancement of collision avoidance strategies in robotics.