Methods of telematics data analysis for decision support systems for optimal unmanned vehicle control

The purpose of the research is purpose of the study. The analysis of modern solutions aimed at supporting decision-making on optimal driving within the framework of the Connected Car concept, as well as systematize the main methods of using telematics data and architecture of similar systems. Methods . Methods. The study is based on an analysis of domestic and foreign publications, patents and practical implementations in the field of Connected Cars, as well as examples of the introduction of telematics platforms in the automotive industry. Classical statistical methods, machine learning algorithms, and big data streaming tools are considered. Special attention is paid to scalability, standardization and quality of telematics information. Results . It has been established that most modern systems rely on basic statistical and machine learning methods (classification, clustering, regression models) to analyze large amounts of data on vehicle movement. However, unified approaches to the integration of these methods into the integrated architecture of decision support systems have not yet been formed. Hybrid approaches combining statistical methods, ML algorithms, and Big Data technologies demonstrate the greatest effectiveness. Their widespread adoption is hampered by the lack of uniform standards for telematics data exchange, the difficulties of reliable data storage, and the need to filter noise and omissions. Based on the review, the advantages and disadvantages of various methods are identified, as well as the requirements for the architecture of the DSS for the Connected Car are formulated. Conclusion . The review confirms the high demand for flexible, scalable solutions capable of processing telematics data in real time and taking into account individual driving characteristics. The further development of such systems is closely related to the unification of telematics information formats, increased security (both in terms of data protection and in the field of traffic), as well as expanding the range of analyzed sources (road infrastructure, weather conditions, smart city ecosystems, etc.) to improve the accuracy of recommendations and optimize driving.