A Cloud-Based Machine Vision Approach for Utilization Prediction of Manual Machine Tools

Since the last decades of the 20th century, the manufacturing industry has been moving toward the development of fully automated equipment; however, a large number of machine tools are still manual and require the operators to stay close by while operating. This allows solutions to be developed that measure machine utilization by tracking and correlating the location of personnel and manual machines. The knowledge of machine utilization and the prediction of machine availability can be extremely advantageous in efficiently scheduling the work that needs to be done with the machines. The lack of this knowledge, on the other hand, can cause long wait times and inefficiencies. In this article, we have proposed and studied a cost affordable cloud-based machine vision approach to capture and predict the utilization of manual machine tools. A case study was performed in Georgia Tech's ME2110 lab, where approximately 400 students design and develop their projects every semester. Because this lab is open to all of these students with no predefined schedules, the statistical analysis on the historical equipment utilization could be one of the only methods of predicting machine availability. By analyzing the data of a security camera mounted in this lab, the location of students was tracked and correlated with the location of the machine tools to find out the utilization time of the machines. The autoregressive-moving-average (ARMA) method was then used to predict the machine utilization for days after. The evaluation results of this framework show that the error between the actual and predicted utilization was less than 20 %. Although the accuracy of this framework with the data collected in 27 days is high and can be used to increase the efficiency of the lab, the accuracy is expected to increase by capturing more data in a longer time period.