Privacy-preserving generation and publication of synthetic trajectory microdata: A comprehensive survey

Abstract

The generation of trajectory data has increased dramatically with the advent and widespread use of GPS-enabled devices. This rich source of data provides invaluable insights for various applications such as traffic optimization, urban planning, crowd management, and public safety. However, the increasing demand for the publication and sharing of trajectory data for big data analytics raises significant privacy concerns due to the sensitive nature of the location information embedded in the trajectory data. Privacy-preserving trajectory publishing (PPTP) has been an active research area to address these concerns, and synthetic trajectory generation has emerged as a promising direction within PPTP. This survey paper provides a comprehensive overview of PPTP with a focus on synthetic trajectory generation methods, which have been insufficiently covered in previous surveys. Our contributions include a comparison of existing PPTP techniques based on their applicability and effectiveness for data analysis tasks. We then review and discuss the existing work on synthetic trajectory generation in the context of PPTP. Specifically, we classify the existing studies into two main categories, algorithm-based and deep learning-based approaches, and within each category, we perform a comparative analysis of the studied methods, focusing on their different characteristics. Finally, in order to encourage further research in this area, we identify and highlight a number of promising directions for future investigation that deserve to be explored in greater depth.