A Review of Machine Learning Methods for Aviation Turbulence Identification and Forecasting
Year: 2026
Type: article
Abstract: Aviation turbulence remains a major challenge to aviation safety, frequently causing passenger injuries. However, traditional turbulence detection and forecasting methods based on airborne equipment and operational numerical weather prediction (NWP) lack the temporal and spatial resolution and timeliness needed to acquire precise turbulence information. To overcome these limitations, machine learning (ML) has drawn extensive attention in aviation turbulence research. This paper primarily analyzes and summarizes the progress of ML methods for aviation turbulence identification and forecasting. First, after extracting the common data foundations, we explore the core concepts, advantages, and practical applications of various ML methods for turbulence identification. Furthermore, categorized by driving data, we analyze the developmental trajectory of turbulence forecasting from single classification algorithms to ensemble learning and ultimately to deep learning (DL). Finally, we discuss primary challenges such as class imbalance, weak model generalization, and insufficient interpretability, along with a summary of future trends, providing a reference for the precise identification and forecasting of aviation turbulence.
Source: IEEE Access
Institution: Civil Aviation University of China
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,075
OpenAlex ID: W7204129263
