Flexible robotic sorting solution for wood material biproducts: A conceptual study to evaluate the automated process of sorting wooden blocks
2022 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE credits
Student thesis
Abstract [en]
The reason for this thesis was that Södra Building Solutions wanted to investigate if it is possible to automate the palletizing process for their new factory to further improvise their production system. This process began with a study of constraints such as available space in the new factory being built, the level of automation expected, man-human interaction, possible variants, and procedures to handle a variety of products as detailed in respective chapters.
The virtual model is built considering that all the parts are handled automatically without the aid of the operator using an autonomous robotic system that integrates machine vision to guide this robotic sorting solution. This thesis arose from studying various available concepts of image processing techniques, and robotic solutions for a similar application. Based on the knowledge gained from the literature study, a virtual model is developed to produce a satisfying and optimised solution for sorting operations. This virtual model built on RobotStudio is simulated to see if there are any further considerations to be made to develop the final setup for the sorting operation. The test results are obtained by processing the images captured in the RobotStudio simulation setup. Once the MATLAB is stabilised and fine-tuned by adjusting parameters such as threshold, and blob values in the code to produce reliable results. These results are sent to RobotStudio using TCP/IP communication protocol, which is determined to be a basic and most convenient way to transfer data between MATLAB and RobotStudio tools in a virtual model. The simulation of the overall solution begins with the testing of pre-determined parameters such as sizes, positions, and patterns. Once it is found that the simulation model functions as expected, the parameters such as part size, shape, and pattern are randomly generated to validate the MATLAB code to produce varying and matching results. During image processing, it is determined lighting plays a very important role in the image processing process since the shadows in the captured image results in producing false edges. These false edges detected in the image result to deviate from the estimated part size to the actual. After the machine vision system has been stabilized to properly detect the components, the solution is further enhanced to incorporate path planning for the robot movement, ensuring that the work is completed with the shortest/optimistic travel distance in the shortest possible amount of time. The layout is organized in such a manner that the robot's work is minimized by setting up an optimized layout in such a way that the task is accomplishedefficiently. A complete solution for performing the sorting and palletizing operation has been outlined in this thesis, along with the implementation ideas and suggestions for developing this concept to a detailed solution, though the implementation and testing of the solution occurred outside the scope of this degree project. This thesis work would benefit anyone who employs machine vision for part detection, size estimation, pattern recognition, position, andorientation detection. From the results obtained by the virtual model, the accuracy and repeatability of the measured values are within the acceptable range. And with these results, the simulation model developed in RobotStudio demonstrates the pick and place operationin a proving manner. This thesis also provides evidence of a working concept in which the complete system works autonomously while integrating with other systems.
Place, publisher, year, edition, pages
2022. , p. 41
Keywords [en]
Image processing, Robotics, Optimization, Path Planning, Simulation
National Category
Robotics
Identifiers
URN: urn:nbn:se:hv:diva-19273Local ID: EXC915OAI: oai:DiVA.org:hv-19273DiVA, id: diva2:1701214
Subject / course
Robotics
Educational program
Master i robotik och automation
Supervisors
Examiners
2022-10-202022-10-052022-10-20Bibliographically approved