Assembly Assistance System with Decision Trees and Ensemble Learning

Abstract

This paper presents different prediction methods based on decision tree and ensemble learning to suggest possible next assembly steps. The predictor is designed to be a component of a sensor-based assembly assistance system whose goal is to provide support via adaptive instructions, considering the assembly progress and, in the future, the estimation of user emotions during training. The assembly assistance station supports inexperienced manufacturing workers, but it can be useful in assisting experienced workers, too. The proposed predictors are evaluated on the data collected in experiments involving both trainees and manufacturing workers, as well as on a mixed dataset, and are compared with other existing predictors. The novelty of the paper is the decision tree-based prediction of the assembly states, in contrast with the previous algorithms which are stochastic-based or neural. The results show that ensemble learning with decision tree components is best suited for adaptive assembly support systems.

Robust Assembly Assistance Using Informed Tree Search with Markov Chains

“…The Markovian-based predictors suffer from not having all the scenarios available in the training data, resulting in a lower prediction rate but higher accuracy. This was also observed in a previous study [37], where the feature importance metrics could be observed from the algorithms. The main feature used for prediction was the state of the tablet, with a representation of over 98% of the splits inside the predictor, at least for the GBDT presented in the study.…”

“…Another part of the heuristic obtaining a considerable score is the ordering of the items in the queue as when no Markov score is present, the last item will be the following child, with the lowest state score described in the methodology used in [28]. Analyzing Figures 9 and 10, similar results to [37] were observed, namely, that for the algorithm, it is easier to learn the behavior of the manufacturing workers, which, based on their experience, are more conservative in their assembly patterns. The increase of 14% in the prediction power can also be attributed to the narrower range of unique assemblies available in the manufacturing dataset compared to the students' dataset, where several unique assemblies were met; whenever the algorithm predicted equal chances for both of them, the first one added in the queue would be picked.…”

“…LSTM is usable in unmet situations, with good coverage, but its drawback is the lower accuracy. Until the current research, the Gradient Boosted Decision Tree (GBDT), evaluated in [37], has proved to be the most performant in predicting and suggesting the next assembly steps. In [37], there are several advantages to using GBDT, mainly maintaining a high prediction rate of 100% and a coverage of 65%.…”

“…Until the current research, the Gradient Boosted Decision Tree (GBDT), evaluated in [37], has proved to be the most performant in predicting and suggesting the next assembly steps. In [37], there are several advantages to using GBDT, mainly maintaining a high prediction rate of 100% and a coverage of 65%. The disadvantage of using GBDT comes from the increased complexity of the algorithm, resulting in high implementation times and high data levels used as it presents an increased risk of overfitting.…”

A Combined Approach for Predicting Employees’ Productivity based on Ensemble Machine Learning Methods

“…Some studies evaluated worker 's performance of textile company by using ML and ensemble learning algorithm, such as study as Saad 2020. [23] which applied different Machine learning algorithms including, decision tree and bagging algorithm to achieve the highest accuracy. The CHAID model produced high-level specificity and sensitivity.…”