METHODS FOR MONITORING AND WARNING SYSTEMS TO SUPPORT SHIP ARRIVALS BASED ON DEEP NEURAL LEARNING NETWORKS

Abstract

The invention relates to a monitoring and warning system to assist ships in berthing using two BAS units integrated with laser distance sensors. The two BAS units are positioned at two different locations along the berth area to measure the distance from the berth edge to the ship's hull in real time. Data from the two BAS units are fed to a central processor using an STM32 for verification, standardization, and calculation of parameters including representative distance, approach speed, and relative approach angle. Each unit has an HMI screen for on-site display, a communication unit and antenna for transmitting data to an online monitoring platform, a waterproof and dustproof enclosure, and a water level for installation calibration. The processed data is standardized into a JSON package, transmitted to the web, stored per berthing session, and used as input data for an LSTM model. The LSTM model predicts the berthing status at the next time step, thereby supporting early warning when there is a tendency towards unsafe approach.Thanks to the combination of two BAS measurement points, STM32 embedded processing, on-site display, web data transmission, and time series forecasting, the invention enhances the ability to observe, visualize, store, and support decision-making during ship berthing.