Intelligent tutoring system design for engineering science courses

Abstract

Recent changes in computer technology have made the development of an intelligent tutoring system (ITS) a much more attractive proposition. Reviewing the current educational engineering software, two major deficiencies surfaced: (1) a method for integration into the classroom was not identified and (2) an instructional model was not used in the development process. This indicated a clear need for a science-based design approach founded upon the cognitive sciences and educational research. In taking the science-based approach, the organization of knowledge and use of different learning styles are identified as two important factors that affect the students' learning and hence the design of the ITS. Input from the faculty to identify key problem areas in classroom learning of engineering science courses was obtained through a survey. These results were used in developing an instructional model for the tutoring module. Fluid mechanics, solid mechanics, and vibration courses were examined for commonalities and differences in approach and a knowledge structure developed for each. These were then generalized to form a basis for the knowledge base. A simple yet effective student model is proposed. This strategy forms a basis for the proposed ITS design. The system design was implemented using LUCID's Symbolic Programming Environment which is an integrated package including LISP, flavors, and a window toolkit. The framework implementation was demonstrated on two fluid mechanics examples and one solid mechanics example. The issues addressed include supplementary needs of the engineering science classroom, hardware and software concerns, identification of key attributes adequate to define a problem, choice of knowledge representation scheme(s), development of an educationally-based tutoring strategy, and implementation issues such as the formalization of the problem solving process. In conclusion, the framework is a viable means of providing supplementary tutoring for the engineering science courses. However, more development work and testing is necessary to produce a classroom product.