Abstract:

Many important applications in engineering and applied sciences require fitting a set of experimental data to mathematical models represented by parametric functions. The obtained mathematical models can then be programmed for simulation or design purposes. This problem is referred to as the least-squares curve fitting where parametric functions are constructed to optimally fit data according to the least-squares criterion. Most undergraduate engineering programs cover the topic of curve fitting at the sophomore level. The use of a technical computing program such MATLAB facilitates the calculations of the model function parameters. However, most programs teach students how to perform only simple curve fitting problems where the parametric model function is represented by the equation of a straight line. While this topic, known as linear regression, has several applications and can easily be taught at the undergraduate level, students often find difficulties understanding the application of curve fitting to nonlinear problems.

This paper presents the methodology we have used to introduce students to nonlinear data fitting in an engineering sophomore-level programming course. The course introduces electrical engineering students to MATLAB programming. The course has a lecture component and a computer laboratory component where students work on laboratory problems with the aim of developing the necessary MATLAB programming skills that they can use in other electrical engineering courses such as electric circuits, signals and systems and automatic control. An important focus of the course is on the implementation of numerical methods.

This paper presents the step-by-step approach introduced in the classroom to help students develop the necessary skills to solve nonlinear curve fitting problems. Examples of laboratory exercises developed for the course are presented and discussed. Finally, the paper summarizes the learned lessons and discusses the effectiveness of our educational approach.

Keywords:

Engineering education, Numerical methods, Technical computing.