Auto-Encoder-Enabled Anomaly Detection in Acceleration Data: Use Case Study in Container Handling Operations

The sudden increase in containerization volumes around the globe has increased overall number of cargo losses, infrastructure damage, and human errors. Most critical losses occur during handling procedures performed by port cranes while sliding containers to inner bays ship along vertical cell guides, damaging main metal frames causing structure deform lose its integrity stability. Strong physical impacts may at any given moment, thus in-time information is ensure clarity processes without halting operations. This problem not been addressed fully recent literature, either researchers engineering community or logistics companies’ representatives. In this paper, we have analyzed conventional means used detect these found that they are outdated, having no real-time assessment capability, only post-factum visual evaluation results. More reliable could benefit many actors transportation chain, making more efficient, safer, reliable. proposed solution incorporates monitoring hardware unit analytics mechanism, namely auto-encoder technology, uses acceleration parameter identify sensor data anomalies informs end-user if occurred procedures. analytical method compared with detection methodology (IDM), result indicates well capable detecting events analyzing curves reshaped signals, same as IDM, improving speed short-term periods. We managed detect–predict between 9 18 impacts, depending on axis container sway. An experimental study suggests programmed correctly, (AE) can be deviations time-series different scenarios.