Nowadays with production of low-cost GPS receivers and equipping various tool with them (from smartphones to in-car navigation systems), large amounts of spatial movement data are produced. These data, that is in the form of a series of spatial and temporal points as (x, y, t), is called trajectory. Among the numerous analyzes performed on these data, semantic enrichment of trajectories is a relatively new research topic, that has been extremely a subject of interest among researchers in recent years. Semantic enrichment of trajectories that is an effective method for management of movement data and performing semantic analyzes, including multi-step process. In this process, raw movement data collected by positioning system such as GPS, are processed and semantic trajectories are produced. Episodes identification in trajectory is an important step that aims to identify meaningful part in the trajectory. The proposed methods in episode identification mostly are based on applications. Therefore, by understanding the capabilities and limitations of each method, depending on the type of data and its application, semantic enrichment process can be done more efficiently in various applications.

This paper reviews the studies in the field of semantic enrichment has focused on the identification of episodes of trajectory. With the focus on stop-move model as the first conceptual model presented in the field of semantic enrichment, various methods of identifying episodes of trajectory is divided into two groups. The first group of methods is based on defined criteria in the stop-move model to identifying episodes of trajectory. The second category of methods uses other than stop-move model in process. The methods in which uses stop-model, divided into two categories: The first that uses merely raw data to identification and the second are based on combination raw data with application-based and geographical environment information. The method that uses raw data such as location or derived information from trajectory such as speed and acceleration, mostly are suitable when there is not enough information in the path of the moving object. On the other hand, when geographical and applications-based information are used, identification of the homogeneous parts in trajectories process is more efficient. In the methods that uses other than stop and move criteria, identification homogeneous segments, in spatial trajectories are based on various applications. One of the most important criteria is to identify homogenous sub-trajectories based on different mode of transportation. This approach that is used in processes of network constraint trajectories, the purpose of enrichment, is semantic analysis of the users' path use in applications such as traffic management and urban transport. The other criterion in identifying homogeneous parts in trajectories is target of moving object in path. In this approach, identification of the heterogeneous trajectory can be defined as hierarchical, from common state to a special purpose. This approach is the most common applications aimed at enriching the users' path trajectories in semantic application such as urban planning. Various methods have been introduced in this paper based on the proposed categorization and the performances of them in different application domains have been evaluated.