Algorithm for automatic analysis and evaluation of communication skills of participants in VR public speaking exercises

This research develops and validates an algorithm for automated analysis of communication skills in public speaking within a Virtual Reality (VR) environment, addressing limitations in traditional training methods. The study introduces an algorithm to analyze comprehensive data, including speech patterns, non-verbal cues, and tailored feedback, particularly for users with disabilities. The method involves multiple steps: processing input data, analyzing gestures, spatial movements, voice tone and timbre, and speech rate. The algorithm was tested against expert evaluations using a dataset of 20 VR public speaking recordings. This VR-based approach provides an immersive, adaptive training environment, crucial for high-stakes fields such as emergency services, where effective communication can impact outcomes in lifethreatening situations.