SYnet: 4D Convolutional Neural Network-Based Maintenance Lift Vehicle Demand Prediction Model for a Large Shipyard

Abstract

Large workshops such as shipyards commonly employ various types of vehicles for specific tasks. Important among them is the maintenance lift vehicle utilized for welding or painting on the upper decks of large ships. The use of this vehicle at a construction site can enhance the overall process efficiency. Hence, we propose a 4D CNN-based method for predicting the regional demand for maintenance lift vehicles in shipyards. In addition, a method for data preprocessing that reflects the shipyard’s strict limitations and requirements is proposed. This method collects real-time data and implements a maintenance lift vehicle demand forecasting system with a dashboard that shipyard workers can use for daily dispatch. The pipeline of the corresponding prediction system is described in detail to demonstrate the utility of our system.