Objectives:
Vehicle movement datasets, primarily obtained from GNSS systems, are characterized by inaccuracies that negatively affect the reliability of analyses. To address this challenge, this paper provides an overview of map-matching techniques and trajectory prediction algorithms. The purpose of map-matching is to precisely reconstruct a vehicle's actual route, which serves as a foundation for further operations. Based on the trajectory reconstructed in this way, prediction algorithms can forecast subsequent movement, enabling the monitoring of transport, the detection of anomalies, and the support of intelligent traffic management through the forecasting of future traffic volume.

Methods:
The article employs a systematic literature review and comparative analysis. It provides an extensive overview of methods for performing map-matching tasks, as well as techniques used for predicting vehicle movement trajectories.

Results:
The results section discusses the detailed evolution of these techniques. A cross-section of solutions is presented, ranging from basic geometric and probabilistic models to advanced, contemporary approaches that utilize deep artificial neural networks.

Conclusions:
The summary points to the concept of unifying both processes within a single architecture. This approach is identified as a potential means of significantly increasing the accuracy and reliability of geolocation data processing systems in future applications.