Session: 03-12-01: Digital Manufacturing Process Simulation and Validation

Paper Number: 113990

113990 - Soft Sensor Digital Twin Implementation of a Pick-and-Place Operation 

In advanced manufacturing today, there is often a reliance on robotic arms and actuators to complete assembly processes. In the increasing frequency of the use of these arms in manufacturing settings, there arise more situations in which interaction between the robotic system and humans becomes necessary. In these situations, the safety of the operator must be taken into consideration, and specific practices and procedures must be adopted to avoid any accidents that may result in injury or death to the operator. Some corporations are exploring industry 4.0 trend to implement next generation automation systems. The use of emerging technologies can enhance the safety for workers who may come into proximity with these robots. Some researchers have been exploring safety aspects using smart sensors. However, few of them have been exploring how smart sensors could be implemented using a manufacturing digital twin. To improve the safety of human-machine interaction, this research project focuses on the implementation of a smart sensor attached to the end effector of the robotic arm that can detect collisions and respond adequately. This research explores digital twins using Siemens’ Process Simulate Tecnomatix allows for the monitoring of the robotic system from a remote location, as well as provides a way to simulate the collision of the end effector and the response of the robot.

The first area of consideration was the creation of a novel modular sensor that would be able to enclose the area around the end effector of the robotic arm, allowing for ample detection of the space that the arm would travel to. Advanced materials were taken into consideration, and soft, flexible resistance sensors were chosen to be the material that would provide a sensor reading upon collision (deformation). SolidWorks was used to model this sensor and determine its viability for this project. 

To model the scene of the robotic arm, Tecnomatix was used. A model of a Kawasaki robotic arm was placed into the simulation and the 3D model of the sensor was attached to the end-effector. Various pick and place operation sequences were then simulated for this robotic arm, and human models were introduced into the scene as well. These simulations allowed for the testing of functionality of the sensor in a virtual environment with controlled situations, as well as providing for a basis for the digital twin with the physical system to be implemented with.

Connectivity between the physical sensor and the Tecnomatix implementation was provided by PLCSIM Advanced, a software tool that facilitates communication between the physical PLC systems and the virtual system. The data garnered from this sensor was used to train a machine learning algorithm to differentiate between collisions and situations that look like collisions in the sensor readings but aren’t truly collisions. This could be in the case of sudden stops, vibrations in the robotic arm, and others. In this case, supervised learning was implemented as the optimal machine learning algorithm. This diferentiation between true and false collisions allowed for the system to only alert/stop when a real problem occurred. The digital twin of the physical prototype was able to be monitored in a remote location with an IIoT PLC. This allows for a safety operator to monitor the functionality of multiple robots at one location, not just at the specific robot location.

With the advanced manufacturing capabilities and technologies representing industry 4.0, a safer way of manufacturing with robotic arms and humans can be achieved. For example, safety protocols using conventional robots against collaborative robots. The results of this study show the possibility of using industry 4.0 technology to improve safety conditions in factory settings and will contribute to the advancement of human-robot interaction and the integration of IIoT and digital twin technology in the manufacturing industry. This paper is exploring some of these aspects as research novelty contribution. 

Presenting Author: Brandon Schrader Kennesaw State University

Presenting Author Biography: Brandon Schrader is a graduate student at Kennesaw State University pursuing a Masters of Science in Intelligent Robotic Systems. He is currently a Graduate Research Assistant for Dr. David A. Guerra-Zubiaga researching the use of PLM and Digital Twins to improve advanced manufacturing. Brandon currently holds a Bachelor of Science in Mechanical Engineering from Florida State University ('22).

Authors:

Brandon Schrader Kennesaw State University
David A. Guerra-Zubiaga Kennesaw State University
Grayson Mcmichael DataSeers