SCOTT, Liliane G. A.1 & OLIVEIRA, Jose’ Carlos R.2
1 Av. D. Jose’ Gaspar, 500 – 30535-610, Belo Horizonte, MG, Brasil, Department of Electrical Engineering, Catholic University of Minas Gerais, liliane@dee.pucminas.br
2 CP 209, 30161-970, Belo Horizonte, MG, Brasil, Department of Electronic Engineering, Federal University of Minas Gerais, jcarlosr@novell.cpdee.ufmg.br, www.cpdee.ufmg.br/~gaeep
Abstract: This work proposes to apply the learning methodology ”Motivation-by-Challenge”, presented by professor J. Jim Zhu in 1994, adapting it to the teaching of Control of Linear Systems in the curricula of Electrical and Control & Automation Engineering in Brazil. In this methodology, the students are exposed to a challenge since the beginning of the course, receiving a control problem to solve vis-a-vis of a set of performance specifications. In theory classes, a model of a specific real process is furnished, and the student is driven to use the tools constituted by the Root Locus and the Frequency Response Methods, and also by simulations. After making the system analysis, in open and in closed loop, the student will be able to design a controller, to improve its dynamic and static characteristics. Further investigation is encouraged, as a talk to a specialist in the field, or a research in papers and at the Internet, to give more realism to the theoretical problem. In the Laboratory, small scale plants are presented or they are proposed to be constructed by the students. In this case, the tools are data acquisition, modeling and parameters identification, model validation by simulation, controller design and, finally, a refined controller tuning. As in theory, the challenge remains, until a reasonable solution is obtained.
In this methodology, the students share the learning task with the instructor, step by step, being an active part of this process. Differently from the conventional methodology, where students and instructor are impelled to do many ready and repetitive tasks, they have an innovative participation in the learning process, sharing responsibility and the obtained results. The planning of the course and the evaluation of the students must also be thought from this methodology. One of the problems of this approach is a bigger amount of work attributed to the instructor, forcing him to choose and prepare new challenges each semester. On the other hand, it results in a more effective learning, traduced by the enthusiasm shown by all the people involved in this activity.
Keywords: Motivation-by-Challenge Methodology, Teaching Control of Linear Systems, Education in Electrical and in Control & Automation Engineering.
The increasing importance of the field of Control and Automation has been shown by the great number of meetings and conferences recently promoted in different countries, by various publications including papers and text books and, principally, by the impact of this field in industrial applications and in other sectors which affect direct or indirectly our lives.
The teaching of Control, in this scenario, is an important task to be carried on, and it applies to a variety of academic courses in Electrical Engineering, Control and Automation Engineering, Mechanical Engineering, Chemical Engineering, Aeronautical Engineering, Transport Engineering and others. A survey of Control Education in the world was presented by Zhu [1996].
Although sophisticated equipment and techniques are being more and more allocated at the engineering curricula, like personal computers, data shows and the Internet, the current applied methodology has been the same conventional one over the years, e.g., practically all the responsibility of the learning process is attributed to the instructor, and the students stay merely as passive agents.
This paper describes the adoption of the methodology ”Motivation-by-Challenge” in the curricula of Electrical and Control & Automation Engineering, and more specifically in the course of ”Control of Linear Systems”, in opposition to the conventional methodology, detailing its application in theory classes and at the laboratory, and also discussing some directives concerning the planning of the course and the evaluation of the students’ performance.
This new methodology has been presented in the literature by Zhu, J.J. [1994], and its essentials are such that the students have an active participation in the learning process, being motivated by a challenge proposed to them. It was conceived in opposition to the so called conventional methodology, where the instructor is the main actor, orienting his (or hers) effort to systematically disseminate knowledge, and being the main responsible for the learning activity.
Before describing this new methodology, it is important to remember the main characteristics of the conventional one, which has been adopted since a long time ago, not only to teach Control, but in almost all the fields of knowledge. Basically, the teacher gives lectures to the audience, when the program is presented in a logical sequence, followed periodically by chosen examples to reinforce concepts, and by exercises to be solved by the students. To close these cycle, the students are submitted to some tests during the course. After this procedure, it is expected that the students have assimilated the program, and that they will be able to apply the acquired knowledge in their future professional activities.
This conventional methodology is based on a top-down approach, and has shown to be efficient in the training of the students in analytical abilities. Besides that, it works well when teaching simultaneously to a large number of students, because they generally stay as passive agents, trying to assimilate what the teacher presents in class. In the best case, a student works hard, studying the received material, without critically questioning it, and not having the opportunity of contributing with his ideas or personal experience. He (she) is evaluated by his grades in written tests or homework. On the other extreme, the course does not motivates the student, becoming very difficult: in this case, there is a weighty ritual to be followed, until the diploma can be obtained. Focusing the teacher’s side, in the same way that he has learnt the subject, he is going to teach, without taking the risk of changes, as he does not know exactly what would be the new results. There is a natural tendency of accommodation, repeating old procedures, which transmit a false feeling of safety.
On the other hand, in the global economy of today, it is more relevant to help the students in the ability of searching for and/or developing new solutions. Of course, it is also important to give them the basic material in the field, so the top-down procedure should not be completely discarded. But even if the school could transmit all the accumulated knowledge (this is not possible!), that would become rapidly obsolete, in function of the astonished speed in which knowledge grows.
The methodology Motivation-by-Challenge is structured in some basic principles, detailed in Scott [1998]. The first one is that the more important task of the instructor is to assist the student during the search for the solution of the proposed challenge. It is very critical to know how to balance this participation, interfering whenever the students exhaust their capability, but without giving them a closed solution. The challenge would loose its motivation, both in the case of being solved by the instructor, or in the hypothesis where the students could feel themselves not capable of facing it, giving up in the middle of the way. Each challenge must be dimensioned for the actual potential of the group at that point of their studies, and it must also be compatible with the period of time allocated to the course. Corrections of route are permitted, during the search for the solution, giving the teacher an opportunity to adjust the size of the challenge, or the grade of difficulty for that specific class of students, if necessary.
A second point is that the challenge(s) must be proposed at the beginning of the course. The students must be informed, at that time, that they do not already have all the knowledge necessary to arrive to a solution for the challenge, but that the tools will be presented to them during the development of the course. This policy has two advantages, permitting a global vision of what the student can expect from the course, and also motivating him (or her) to see the theoretical part (sometimes hard in the case of Control) as a valuable tool. In this manner, the work will be performed with objectiveness, and its target will be reached with efficacy.
Figure 1, from [Zhu, 1994], shows a block diagram which explains the methodology. The reference to this feedback system is a real problem proposed to a unique student or to a group, who must try to find a good solution for it. While the solution is being pursued and has not yet been reached, there exists an error, e.g., a challenge. This challenge is the ”force” that motivates the students to learn and to use the tools, represented by theory, strongly based on Mathematics in the case of Control. Laboratory activities are also encouraged, but they are sometimes suppressed in some curricula , or they constitute a separate course. When the solution is reached, the challenge disappears, and the learning process must be completed with a generalization of the knowledge acquired during all the process.
Figure 1. The Motivation-by-Challenge Methodology
The application of the Motivation-by-Challenge in theory classes implies that a challenge be launched at the beginning of the course. A simpler way of doing that is choosing control engineering problems in the literature (text books, papers, technical magazines), normally modeled by transfer functions or state variables, but always referred to the real world. From this point of view, the parameters must be extracted from real plants, and it is not a good practice to propose ”academic” examples, for instance, of the type where all poles are integer and well conditioned numbers (0, -1, -2, etc.). If possible, the proposition should be accompanied by schematic diagrams and sometimes by photos [Dorf, 1998]. To give more realism to these challenges, the students must be encouraged to search for more information about the plant in the Internet, or to talk to specialists in other departments of the university.
Another possibility of electing control engineering problems is from the experimental research of the control group. Interactions with industry will also provide good examples to give to the students: in this case, technical visits to real plants may be programmed along the course, acting as an additional motivation.
A second step is done through the choice of a set of performance specifications for the plant to be controlled. This part is critical, because it must take into account the physical limitations of the real process under study, and because it will establish the main directives for the conduction of the challenge. The response of the closed-loop controlled system must satisfy time domain performance specifications like rising time, peak time, overshoot, settling time and steady-state error, as well as frequency response specifications like open-loop phase and gain margins, and closed-loop bandwidth.
The development of the challenge, in a first course of Control of Linear Systems, will lead the students to use of the Root Locus and the Frequency Response Methods, presented in class. These theoretical tools, and also a good simulation software [MATLAB/ Simulink, 1994], will help the students in the analysis of the stability and closed loop response of the control system. In view of the previous set of performance specifications, normally not satisfied simply by a gain adjustment, the student will be challenged to choose and design an specific controller (or compensator), to improve the system dynamics and/or the static characteristics. Using one of the above methods, the controller parameters will be determined by a trial-and-error procedure.
The auxiliary tools normally used to help the students are constituted by text books, lists of solved exercises, lecture notes and, of course, an extra class teacher’s assistance. These tools are also valuable with the new methodology but, with the spread of personal computers, one can make an intensive use of them. With this in mind, we have developed special tutorials to construct and use Root Locus and Frequency Response plots with MATLAB [Oliveira & Scott, 1997]. Alternatively, one can find excellent material in the Internet, like the Control Tutorial developed by the Universities of Michigan and Carnegie Mellon in USA [Michigan University and MATHWORKS, 1999]. In this same line, complete courses may be offered and followed at distance, using the Internet, like the ”C Language Programming Course” organized by a group at UFMG, Brazil [ Mesquita, 1999], but this approach is not the purpose of this paper , and will not be detailed here.
Another powerful tool is the Intranet, e.g., the use of Internet inside the organization: this mechanism permits the teacher to pass additional material to the students, with the advantage that they can access it in an asynchronous way, whenever they need or they have time for it. The communication between the students and the teacher may be done with the aid of the e-mail and an electronic discussion list. The personal contact must not be discarded, of course, but these are additional channels to improve the learning process.
In the conventional methodology, the control problem would be finished with the design of an analog controller, by means of operational amplifiers and passive components. In the presented approach, based on the Motivation-by-Challenge methodology, however, there is a final and very important task to be executed, e.g., the validation of the project. Even in a theoretical point of view, the compensated system must be studied in a more realistic way, and this part can be performed with the aid of simulation. Real effects as saturation or hysteresis, not covered by the Linear Theory, can be easily introduced in the simulated validation model, and this procedure may conduct to a change in the tuning of the controller. Another effect that can be added to the validation model is some non-dominating pole, not considered in the model used for the design, which may prevent the physical variables of the simulated plant from instantaneous and non-realistic variations in time. Finally, a critical view of the whole control system, based on information collected by the students, may show, for instance, that some variables are outside their normal values, so the controller design must, once more, be reexamined.
Considering that a good solution has been found for the system, e.g., that the performance specifications have been satisfied, the challenge has arrived to its end. To finish the learning process, the students may be invited to proceed to a generalization of the main control aspects found along the work they have done. This will give them a synthesis of the course.
The students must write about and also present orally the intermediate steps and the final results of the challenges. In this perspective, at UFMG they have being using e-mail to send files to the teacher, containing the partial and the final reports of the challenges, at predetermined dates.
The Motivation-by-Challenge methodology is very suitable to be applied in the laboratory. By the way, in our experience in teaching Electrical Engineering, during a few years we have being applying this methodology just in experimental classes, not only in Control, but also in Power Electronics. The application of the methodology is based on the control of small scale (but real) plants, which must have the following characteristics:
The plants can be chosen among those found in the international market (specially conceived for teaching Control) or they can be constructed locally in the laboratory. In the undergraduate control laboratory at Federal University of Minas Gerais (UFMG), Brazil, these two options have been considered but, in our teaching experience, we have preferred the second. We are developing small plants, because this creates an additional motivation for the students, it is more economical, promotes a closer interaction among our laboratories, permits to change easily some parameters and, finally, we do not loose our capability of maintaining and up-grading the plants without depending on a furnisher.
In earlier days, we have taught with (and also written) Laboratory Guides, which are another form of top-down approach, with some advantages and the problems already pointed to the conventional methodology. Now, from the point of view of the Motivation-by-Challenge, the students are divided into groups, each one receiving (or choosing) a challenge, e.g., a plant to control. The output variable depends on the type of sensors available in the lab or in the market, and a set of performance specifications are established for the plant.
Differently from the top-down method, a predetermined task is not necessarily attributed to each laboratory session, but the group has to spend the time in accordance to a planning, made by them. The teacher must pay attention to the evolution of each group, to avoid the risk of having the groups arriving at the end of the course late in the schedule, implying that they do not have reached the target. A manner of avoiding that is to divide the challenge into parts, and to predetermine dates for the students to make little oral presentations. At the middle of the semester, they must present also a written report, which will serve, for them and for the instructor, as a measurement of what they have done, and what are the tasks remaining. This partial report will serve also a draft of the final one, made when the challenge has been reached.
The steps for the development of the challenge are: familiarity with the plant, open-loop operation and data acquisition of its response, modeling and parameters identification, model validation by simulation, controller design and, finally, a refined controller tuning.
When the students started constructing a plant, it is clear that they do not had time to explore it profoundly, from the point of view of control, in function of the amount of tasks involved. The better that they have reached in that case was a closed-loop control with a proportional controller, because the students had spent much time in developing the other previous steps. Alternatively, for plants already developed, the teacher can propose a variety of challenges. In the case of working with a unique plant for each group, the students may be asked to improve its open-loop dynamical and/or static performance, or to design more than one controller for closed-loop operation, comparing the respective system performances. Other forms of challenges could lead the teacher to attribute the same plant to various groups, each one presenting a different control solution. Finally, a third and more advanced option could be, for example, the insertion of a second loop (a speed loop inside the position loop, in a servo), or the compensation of non-linearities or disturbs.
In the undergraduate control laboratory at UFMG, we have constructed and controlled, until now, the following plants:
The infrastructure necessary for the application of the Motivation-by-Challenge methodology is not different from that of an ordinary laboratory:
The same plants that are controlled in an analog form (continuous in time), as we have described above, may be used, with the aid of the microcomputer, to provide another set of challenges in Digital Control.
The Laboratory Course finishes at the end of the semester, with each group of students presenting a final written report, which will serve as a documentation of the work and an evaluation of if and how they have reached the challenge. An oral and individual presentation, with the plant being running and controlled, is also a good instrument to complete the evaluation.
The following items must be considered in planning the course: its objective, the learning methodology, the program, the chronology and the form of evaluation, all of them coherently with the adopted methodology.
When establishing the objective, one must have in mind what are the abilities which are expected to be developed in the students. Besides the improvement of technical skills, the students need to find satisfaction, and to grow in self-confidence and initiative. Although the course may be oriented to a medium profile of people, it shall also introduce the liking for research and development, and the motivation for advanced studies.
The proposed methodology has been already discussed along this text.
The program is the same presented by the conventional approach: the Laplace Transforms, the Transfer Functions Models, the Root Locus and the Frequency Response Methods, for the analysis and the design of closed-loop controllers. The difference is in the way the course is conceived and developed: the ”challenge” versus the ”ready solutions”, real world problems versus prepared examples. In the top-down learning method, all the subject is assimilated by the students in a fragmentary way: they finish the course not knowing exactly how to join all the parts of the presented knowledge, for the solution of real problems. On the opposite side, the Motivation-by-Challenge considers the above program items as tools (important, but just tools) to solve real control problems.
The chronology must be suited in accordance to the challenges and their solutions.
The evaluation of the students is a form of feedback. In theory classes, the conventional tests only certificate that the tools have been assimilated. The final result, the effective learning, will be confirmed by the quality of the solutions proposed and developed for the challenges.
These tests shall be maintained, but with a smaller emphasis. At the laboratory, the evaluation is conducted continuously in time, because the instructor has a little number of students in each session, which permits him (or her) to accompany closer how they face and develop the challenges. A partial and a full report are recommended, as previously discussed, together with oral presentations.
Until now, some practical results can be attributed to the adoption of the Motivation-by-Challenge. Satisfaction and self-confidence have been clearly demonstrated by the students. They are learning how to take the initiative in facing the challenges. Some of them are discovering their liking for research. Others are being motivated for advanced studies in Control (optimal, robust, adaptive, non-linear). A few students have been engaged in the development of autonomous projects in control, as an undergraduate course [Awwal, 1997]. And finally, some challenges have attained a level of being accepted and presented in Brazilian Conferences [Rodrigues et al., 1998].
On the other hand, the teachers and the technical staff are discovering a new motivation for their work, learning to share and to valorize each one’s abilities.
The departments involved in this experience, as well as the curricula of Electrical and of Control & Automation Engineering, are been renewed with the creation and the improvement of plants for their undergraduate control laboratories. Subjects for Master in Science thesis at UFMG are also positive results of this policy [Scott, 1998].
The difficulties arising from this methodology are evident. More time and creativeness are required from teachers and the technical staff. Financial resources are a critical point. A better and more profound knowledge in other Engineering areas is necessary. The teacher must transmit a spirit of leadership and self-confidence to his students. The course does not become easier, but it demands more dedication of the students in extra-class activities. And some challenges does not reach a reasonable result.
The methodology Motivation-by-Challenge permits an active participation of the students in the learning process, in theory classes and in the laboratory. A control problem is proposed as a challenge, which finishes when a good solution is found.
After some semesters adopting this new methodology in the courses of Control of Linear Systems at Federal University and at Catholic University, both in Minas Gerais, Brazil, more satisfactory learning results have been obtained, demonstrated by a greater involvement of students and some teachers. The students have shown more interest in the course, going beyond the initial established goals.
The evaluation process must be deeply analyzed, from the point of view of the principles introduced by this methodology, incorporating the experience accumulated with its application.
The great challenge presented to the teachers and to the Department is to create new real control problems, in theory as well as in the laboratory. Trained and motivated people, and material resources, are also essential.
Although we do not have yet statistical data collected from the application of this methodology, we have preliminary results showing that it is improving the quality of the learning process, when compared to the conventional approach. The basic key for the success of this idea is to believe and to invest in it.
ASTROM, K. J. Education in Automatic Control at Lund Institute of Tecnology. In: Proceedings of the 1991 American Control Conference. USA, 1991, pp. 306-311.
AWWAL, M., A. Aperfeicoamento de Prototipo do Laboratorio de Ensino de Controle e Instrumentacao (Improvement of a Prototype for the Laboratory of Control and Instrumentation). In: Technical Report. Department of Electronic Engineering, School of Engineering, Federal University of Minas Gerais. Belo Horizonte, Brazil, 1997, 49pp, (in Portuguese).
DORF, R. C., BISCHOP, R. H. Modern Control Systems. Menlo Park, California, USA, Addison-Wesley, 8th edition, 1998. 855 pp.
KHEIR, N. A. , ASTROM, K. J., AUSLANDER, D., CHEOK, K. C., FRANKLIN, G. F., MASTEN, M. & RABINS, M. Control Systems Engineering Education. In: IFAC Automatica. Great Britain, 1996, vol. 32, N. 2, pp. 147-166.
MESQUITA, R. Curso de Programacao com a Linguagem C (Course of Programming in the C Language). In: listproc@cpdee.ufmg.br. Belo Horizonte, Brazil, 1999 (in Portuguese).
MICHIGAN UNIVERSITY AND MATHWORKS. Control Tutorial for Matlab. In: http://hpme12.me.cmu.edu/matlab/htm1/. Carnegie Mellon University, USA, 1999.
OLIVEIRA, J. C. R., SCOTT, L. G. A. Tutoriais para Utilizacao do Software Matlab em Controle (Tutorials for the Utilization of Matlab in Control). In: Technical Report. Department of Electronic Engineering, School of Engineering, Federal University of Minas Gerais, Belo Horizonte, Brazil, 1997, 7 pp (in Portuguese).
RODRIGUES, D. D., BERING, H. S., OLIVEIRA, J. C. R., JOTA, F. G. Uma Metodologia Pratica e Simples para Analise e Projeto de Controladores PID (A Practical and Simple Methodology for the Analysis and Design of PID Controllers). In: Proceedings of II Congresso Mineiro de Automacao e V Simposio de Instrumentacao ISA-GRINST. Belo Horizonte, Brazil, 1998, pp. 251-259 (in Portuguese).
SCOTT, L. G. A . Aplicacao da Metodologia ”Motivacao pelo Desafio” ao Ensino de Controle de Sistemas Lineares (Application of the Methodology ”Motivation-by-Challenge” in Control of Linear Systems). In: Master of Science Thesis. Graduation Program in Electrical Engineering, Federal University of Minas Gerais, Belo Horizonte, Brazil, 1998, 113pp (in Portuguese).
THE MATHWORKS, INC. Matlab / Simulink Version 4.2 c.1. USA, 1994.
ZHU, J. J. Control Education: A World Showcase. In: IEEE Control Systems Magazine. USA, 1996, vol.16, pp.8-78.
ZHU, J. J. Motivation – By – Challenge For Teaching Control Systems. In: IEEE Control Systems Magazine. USA, 1994, vol. 14, pp. 64-68.