International Journal of Engineering and Technology

This paper presents the research work that aims at designing a speech based interactive ubiquitous teaching robot for quality enhancement of Higher Education ensuring better performance by making possibility of interaction between the Robot Teacher and Student/ Learner. Students/ Learner can ask technical questions for which the robot replies. Here the reply by the Robot is limited by the answers available in the database. The proposed speech based interactive teaching system has two phases; they are training and testing. A feed forward neural network is used as a supervised classifier after it is trained by means of a back propagation (BP) algorithm using the features of sound signal from the Student/ Learner. The number of queries present in the database and non database queries recognized is presented by Receiver Operating Characteristics curve (ROC). The ROC curve lies above the diagonal of the ROC plot. This indicates that the proposed method of interaction of Robot with the learners is acceptable. The percentage of question recognition resulted in 74%.