Simulation models in electronic learning systems developed for the control of interactive audiovisual students

Our work focuses on one of the main actors in distance education who is the Tutor who intervenes to help the learner in critical situations of decision-making in his learning path. Our goal is to integrate an emotion recognition system into the tutor module of the e-Iearning platform to automate the task of tutoring. MOODLE platform is the online learning environment used. This platform is called “open and remote drive systems”, it has a user-friendly interface easy to manipulate by actors using it. Moodle is organized in blocks and has a modular architecture that makes it easy to add new blocks. Our goal is to contribute to the creation of an automatic tutor who can make the right decisions to guide the learner during a learning situation. These decisions are based on emotions extracted using an audiovisual Emotion Recognition System (ERS). For the design of this ERS system, we used two modeling approaches. The first approach is the workflow to model the decision-making block and the second is the connectionist approach used to model the emotion recognition block using a neural network. The system uses several algorithms for the analysis of the audiovisual information of which we cite PCA and MFCC.