Real-Time Collision Avoidance for Programmable Machines from 3D Model Analysis

Brooker, Rob (2019) Real-Time Collision Avoidance for Programmable Machines from 3D Model Analysis. [USQ Project]

Text (Project)

Download (1MB) | Preview


Collision avoidance in programmable machines can reduce programming and setup time, and reduce the likelihood of needing to replace or repair parts during commissioning. While collision avoidance can be accomplished manually by a thorough analysis of the 3D model of the machine, and additional PLC code, this may protect the machine, but costs additional time, and is susceptible to human error.

The proposed system includes a computer program to export the 3D model of the machine, and a custom computer that is attached to the machine which manipulates the 3D model in real-time to detect approaching collisions. This computer signals the machine to stop when a collision is predicted. By performing interference detection on the 3D model of the machine, it effectively eliminates the possibility of human error, and saves the additional time that would otherwise be dedicated to the structural analysis and protection code mentioned above.

This project used a Raspberry Pi 3 Model B+ connected to the machine via a fieldbus link, to read axis positions and speeds from the machine PLC. It sends stop signals directly to servos when a collision is detected. From simulation testing it was determined that model complexity has a large effect on performance, but using a more powerful computer, and developing a better 3D model exporting algorithm could improve performance significantly.

Physical testing demonstrated accuracy and reliability, with reasonable response times. With limited optimization conducted for individual axes during testing, the performance of the test system showed great promise for further development, including an auto-tuning mode which will measure the dynamics of each axis to find the best response parameters for each. Work will continue on this system until a commercial product is realized.

Statistics for USQ ePrint 43111
Statistics for this ePrint Item
Item Type: USQ Project
Item Status: Live Archive
Faculty/School / Institute/Centre: Historic - Faculty of Health, Engineering and Sciences - School of Mechanical and Electrical Engineering (1 Jul 2013 - 31 Dec 2021)
Supervisors: Low, Tobias
Qualification: Bachelor of Engineering (Honours) (Mechatronics)
Date Deposited: 12 Aug 2021 01:00
Last Modified: 26 Jun 2023 05:52

Actions (login required)

View Item Archive Repository Staff Only