LabPulse: An Intelligent Multi-Modal Framework for Virtual Laboratory, Automated Programming Assessment and Learning Analytics

Abstract

Laboratory assessment in programming-oriented engineering courses combines executable artifacts, conceptual understanding, code quality, academic integrity, and submission discipline. Existing virtual laboratories and automated programming assessment tools commonly emphasize execution correctness and immediate feedback, while laboratoryassessment research identifies a continuing need to connect learning objectives with richer assessment evidence. This paper presents LabPulse, a prototype academic virtual laboratory and automated evaluation framework derived from an implemented React/Vite frontend and Express-based backend. The system organizes subjects and experiments, executes student programs through language-specific runner paths, evaluates submissions using a 100-mark multi-factor model, generates heuristic code-quality feedback, evaluates typed viva responses, screens code similarity, and exposes analytics/proctoring/help-request interfaces.