Review of smart wearables sensor-based techniques and machine learning for change of direction detection in sports

Change of Direction (COD) is a critical movement skill during athletic gameplay. In team sports like football, basketball, and rugby, the on-field performance is often assessed by the ability to change direction quickly. Optimizing COD reduces injury risk and improves team outcomes. This review examines algorithmic approaches for detecting COD using wearable sensor data. Each method's process, results, strengths, and limitations are summarized. An exploratory methodology was used to search databases like Google Scholar, PubMed, IEEE, and Science Direct. Findings show ongoing progress in COD detection, but also highlight gaps, such as non-standardized sensor placement, inconsistent sampling rates, and limited open datasets. These issues hinder the use of Machine Learning (ML) and Deep Learning (DL) models. To address this, we propose an AI-based framework to automate COD detection. The review also highlighted the lack of standardized protocols for sensor placement, sampling frequency during COD movement analysis, and the limited availability of open-access datasets for COD movement, which leads to the low utilization of machine learning (ML) and deep learning (DL) models in automating the process of COD detection. Based on the identified limitations, we proposed a framework that uses Artificial Intelligence (AI) to automate COD detection. This review aims to improve COD movement detection with wearable sensors. It will help with informed decisionmaking and lay the groundwork for future research.