Model Fuzzy Logic Untuk Memprediksi Resiko Kecelakaan Kerja Berbasis Smart Construction

Abstract

A fuzzy logic model can be applied in predicting the risk of work accidents, where an intelligent framework that integrates the Internet of Things (IoT). Fuzzy Logic Model, Work accident risk, and Safety Recommendation Engine (SRE) for predicting the risk of work accidents in a smart construction environment. Existing construction safety systems generally use a threshold-based approach and are not yet able to provide transparency in decision-making. To overcome these limitations, this study uses parameters such as working temperature, noise level, worker fatigue, compliance with the use of personal protective equipment (PPE), and distance to hazardous areas. A total of 500 synthetic work safety scenarios were built based on construction project conditions validated by OHS experts. The evaluation was carried out by comparing the proposed framework with Decision Tree, Random Forest, XGBoost, and Fuzzy Mamdani. The results of the study show that this model design has better classification performance and is able to explain the factors causing risk through the Risk Explanation Index (REI). In addition, the Safety Recommendation Engine successfully produces mitigation recommendations according to the dominant factors causing risk. This research contributes to the development of Artificial Intelligence Models in work safety management and supports the implementation of Smart Construction.

Keywords: Artificial Intelligence, Decision Tree, Fuzzy Logic, Internet Of Things, Smart Construction

GS Cited Analysis