MULTILAYER MICROSERVICE-AGENT ARCHITECTURE OF A PERSONALIZED LEARNING SYSTEM

Abstract

The goal of this work is to develop a multi-layered microservice-agent architecture for implementing personalized learning based on the use of behavioral data with an integrated computer vision module (SBMS–LMS), which ensures the adaptability of the learning process and the ability to automatically track various student activities. The methodological basis is a combination of the principles of multi-agent systems, microservice architecture, sensory behavioral analytics, and formal security verification in temporal logic. The architecture implements a seven-level model that includes a sensor layer (SBMS), analytical services (Learning Analytics, Adaptive Learning), agent orchestration (MAS), and integration with the widely used LMS Moodle via REST API technology. The system allows for the collection of behavioral signals (student attendance, attention in lectures, disciplinary violations, and use of control system bypasses), students activity and full performance analytics, the formation of recommendations for students, and real-time feedback, which allows students to influence the learning process and improve its quality. The security module guarantees authentication, encryption, and fully anonymization of users data in accordance with GDPR/FERPA standartisation policies. The scientific novelty lies in the creation of a unified, standardized cognitive-analytical circuit that combines agent-based decision-making logic, stream analytics of behavioral data, and the microservice capability to break down educational service functions into independent modules, which in turn increases fault tolerance and protection against full access to the system by a potential attacker. The work results obtained in the course of the research confirm in real practice the effectiveness of SBMS integration with LMS Moodle and the ability of the created system to dynamically adapt current learning trajectories, improving the quality of learning and improve personalization level of education.