Interactive DSP Teaching

FADZIL, Ahmad M. H.

Universiti Teknologi PETRONAS, Seri Iskandar, 31750 Tronoh, Perak. fadzmo@petronas.com.my

1 Introduction

UTP has adopted a multiple educational delivery system that incorporates both formal and informal mechanisms [6]. The formal education delivery comprises of lectures/instructions, practice/project-based work, CD-ROM interactives, and industrial training. The lecture-based delivery is principally centred on the teaching aspects whilst the rest of the delivery are student-centred which encourage students to be responsible for their own learning. Examples of informal educational delivery are peer learning, on-line discussions, email, cyber-surfing, and self-learning modules. These are also student-centred modes of delivery. The virtues of the various delivery methods are not in doubt but the timely use of the methods can inevitably enhance overall effectiveness in delivery.

There is now a shift of emphasis towards learning by the students themselves. The reason is simply that students are motivated by being actively involved in the learning process, and are fascinated by the process of discovery itself. With advent of information technology (IT), informal educational delivery becomes more feasible. The rate of absorption or learning is further increased when the process is conducted within groups. Although one can deliver more via formal lectures, learning by the recipient may be somewhat limited as very little interaction occurs. A lot of studying or scholarly pursuit is needed after attending lectures in order to ensure maximum absorption of concepts. However, there is no guarantee that students will put in the time and effort needed. Practice- and/or project-based delivery on the other hand, ensures learning will take place and thereby is suitable for the semester-based curricula at UTP. Figure 1 illustrates the UTP educational delivery model.

Figure 1. UTP Educational Delivery System

It can be seen that the educational delivery model entails a superior learning environment with resources in place. These resources include, hardware and software (IT-based) tools, and adequate lecturers and facilitators.

The UTP engineering curricula is a 5-year programme with an 8-month industrial training stint at PETRONAS operating units (plants) or other related companies (see Appendix 1) [1]. UTP students will be required to conduct multi-disciplinary and inter-disciplinary engineering team projects in their first, third and final years of study. In the curricula, the science and mathematics courses would have been completed by the second year of study. Thus, for EE students, the SS course is offered early in the third year, followed by the DSP and Control courses in the fourth year.

The SS course introduces fundamental signal processing concepts and thus is de facto pre-requisite course for the DSP and Control courses. Appendix 2 outlines the syllabi for the SS and DSP courses at UTP. Notice that the courses are given a 3-credit-hour rating each. To achieve the 3-credit-hour rating, the course is split into two portions; two 1-hour lectures followed by 1.5-2 hours of practice or project laboratory work. This is in line with the stated UTP educational delivery model.

2 Learning and design tools for signal processing

In order to facilitate the signal processing learning environment at UTP as mentioned in the preceding section, two support tools are used, namely MATLAB and TMS320C5x DSP starter kit (DSK).

MATLAB is a useful tool to facilitate the computation, visualization and programming of signals processing processes in a user-friendly computing environment [3]. The problems associated with signal processing processes can be expressed in the usual mathematical notation and solutions displayed in varying forms. MATLAB has evolved over many years and has become a standard instructional tool for engineering, mathematics and science courses as well as research tool in university environments.

With a family of application specific solutions called toolboxes, MATLAB enables students to learn and apply specialized technology topics. The toolboxes are comprehensive collections of MATLAB functions (M-files library) that extend the MATLAB environment to solve particular classes of problems. The MATLAB functions in the toolboxes are prepared algorithms that can be immediately used in solving complex engineering problems. Some of the toolboxes available are signal processing, control systems, neural networks, fuzzy logic, wavelets and many others.

Figure 2. Typical MATLAB Codes

In particular, the Signal Processing Toolbox provides a range of signal processing functions, from waveform generation to filter design and implementation, parametric modeling and spectral analysis [2]. The basic functions in the toolbox such as basic waveform generation, filter implementation and analyses, impulse and frequency response, zero-pole analysis, linear system models are extremely useful in the SS courses delivery. The other functions such as discrete Fourier transform, IIR and FIR filter design and implementation are particularly useful in the DSP course delivery.

At UTP Engineering Workstation Laboratory (see Figure 2), there are 40 MATLAB licenses running on Windows NT networked workstations and another 11 MATLAB standalone licenses on Windows 95 PCs. The 40 MATLAB licenses are for student use and the other 11 are for the academic staff use at their desktops. The MATLAB licenses come complete with most toolboxes available such as signal processing, control, telecommunications, etc. The MATLAB software is used by both engineering and math academic staff in their respective course delivery.

Figure 3. UTP Engineering Workstation Lab

The C5x DSK is a low-cost, stand-alone DSP application board that allows students to experiment with signal processing concepts in real-time and engage in DSP project or development work [4, 5]. The DSK system comes with the DSK assembler and debugger software interfaces that help the user to develop, test, and refine DSK assembly language programs. Thus, with the DSK, students can create their own application software to run on the DSK board, build new applications or expand the system

Figure 4. Developing Code for DSK

Figure 4 illustrates the software development flow for the DSK system. The assembler translates DSK assembly language source files into machine language object (executable) files for the TMS320C5x family of processors. The executable file is downloaded to run on the TMS320C50 processor of the DSK board. The goal of the development process is to produce code that can be executed on the DSK system as designed. The debugger can be invoked to correct and refine the code.

Figure 5. TMS320C5x DSK Block Diagram

As shown in Figure 5, the DSK board quite simply consists of a C50 DSP, TLC32040 Analogue Interface Circuit (AIC), a 32Kbytes boot PROM, I/O expansion bus, standard RCA connectors, and XDS510 emulator connector. The kernel program is contained in the 32Kbytes, 8-bit PROM and is used only for DSK boot loading. No external memory is available on board as the 10K on-chip RAM of the C50 suffices for most DSP application programs. PC communications is via the RS232 communications port. The AIC enables voice quality data analog acquisition (14 bits of dynamic range) and interfaces to the C50 via its serial port. Two RCA connectors provide analog input and output on the board.

At UTP Electronic Engineering Laboratory, there are currently 17 DSK systems for use by students during laboratory sessions and for project work.

3 Effective learning

To create an effective learning environment for SS and DSP courses, the following issues were addressed:

1. Appropriate use of development tools for different modes of delivery,
2. Appropriate sequencing of topics in the delivery of SS and DSP courses, and
3. Pairing of SS and DSP courses to enhance learning of signal processing.

Appropriate use of development tools for different modes of delivery.

In the preceding two sections we have discussed the UTP educational delivery model and the development tools available at UTP. To achieve the 3-credit-hour rating, each course is split into two portions; two 1-hour lectures followed by 1.5-2 hours of practice or project laboratory work. It is therefore important to adopt a strategy that maximises learning when choosing the tools for the second portion of the course.

Figure 6. Appropriate use of development tools for different modes of delivery

The features of MATLAB software such as graphics for visualisation and ease of mathematical programming, make it suitable to be use in practice laboratory work. Here, the students use MATLAB codes to reinforce the understanding of signal processing concepts and apply concepts for analyses. Prepared MATLAB codes help to reinforce the theory taught in lectures. In the next stage, problem-solving assignments are given, requiring students to apply the concepts in a simulated environment of MATLAB.

On the other hand, the features of the DSK board such as, the code development environment and available interfaces to the real world enable students working in teams to carry out real-time DSP applications or project work. Here, a longer time frame is given to students to synthesize concepts to complete the project work, for example, a three-week period.

Appropriate sequencing of topics in the delivery of SS and DSP courses.

Traditionally, the SS course deals with signal processing concepts associated with continuous-time linear time-invariant (LTI) signals and systems. Whilst, the DSP course would deal with concepts normally associated with discrete-time LTI signals and systems. The justification for the above was quite acceptable as the following three reasons suggest:

1. lag time in understanding of concepts,
2. confusion between continuous-time and discrete-time scenarios, and
3. unavailability of simulation and visualisation tools.

With the advent of mathematical simulation and visualisation tools such as MATLAB, the learning of SS and DSP becomes easier and more interesting. The dryness of mathematical rigour is replaced by colourful and animated graphics that simulate the mathematical process. MATLAB is found to be much suited to DSP learning since it handles the mathematics in the form of matrices that is, it handles sequence(s) of numbers.

In MATLAB, signals are actually sampled versions of its continuous-time counterpart. Consequently, it is difficult to use MATLAB as a visualisation tool for signal processing concepts relating to continuous-time signals and systems normally covered in the SS course. In order to overcome this setback, it is important to review the sequencing of signal processing topics for the SS and DSP courses.

Figure 7. Appropriate sequencing of topics in the delivery of SS and DSP courses.

At UTP, the SS course covers both continuous-time and discrete-time LTI signals and systems. Although it was mentioned earlier that this would cause confusion in amongst students, steps are taken to ensure students plot continuous-time and discrete-time signals differently. As an example, one should use the 'stem' function to plot discrete-time signals and the 'plot' function to plot continuous-time signals. This way, students are able to distinguish different types of signals not just mathematically but also visually. Using MATLAB will also reduce the lag time in understanding of concepts. Hence, by proper use of the visualisation features of MATLAB, the learning of both continuous-time and discrete-time signals in the SS course can be achieved without the aforementioned problems.

In addition, it is now possible to introduce to students the DSK development tool at an early stage of their undergraduate study. At UTP, the students are introduced to the DSK system in the last quarter of the SS course. The students are given a simple group project such as tone generation, to be completed within 3 weeks. At the end of the SS course, the students have been exposed to DSP software development and real-time signal processing.

Pairing of SS and DSP courses to enhance learning of signal processing.

With the appropriate sequencing of topics for the SS and DSP courses, it is now possible to pair the course objectives. This, in effect, will enhance the overall learning of signal processing.

As discussed in the preceding section, the development tools such as MATLAB and DSK system can be utilised at an early stage of the undergraduate study. The experience of using the learning tools during the SS course will definitely prove useful for the following DSP course. By pairing course objectives, we mean that the design and project works during the DSP course should not be limited to the topics in the course but include the topics previous covered in the SS course.

This is very true when using the DSK board. Students are only introduced to the DSK in the last quarter of the SS course and thus, are only able to carry out a simple real-time signal-processing project due to limited time. However, students get to use the DSK boards early in the DSP course. More projects can be carried out including topics in the SS course. Moreover, early use of DSK will enable students to achieve higher levels in project work in the DSP course.

Figure 8. Pairing of SS and DSP courses to enhance learning of signal processing.

4 Conclusions

The multiple educational delivery system that UTP adopts entails a superior learning environment. In implementing the system for SS and DSP courses, we have looked into the use of certain development tools such as MATLAB and the TI DSP Starter Kit (DSK).

Preliminary findings have shown that the use of development tools such as MATLAB and DSK has increased motivation in student learning, understanding and subsequently awareness of the applications of the signal processing concepts.

The effectiveness of these learning tools can be greatly enhanced by:

1. Appropriate use of development tools for different modes of delivery,
2. Appropriate sequencing of topics in the delivery of SS and DSP courses, and
3. Pairing of SS and DSP courses to enhance learning of signal processing.

As a consequent, the utilisation of the tools has greatly increased. It is also expected that the learning tools will be used at a higher level as the students progress in their study programme.

5 Acknowledgements

The author expresses his gratitude to Texas Instruments Malaysia for the support in the teaching of signal processing at UTP, in particular, for the donation of the DSK boards. The author would also like to thank his colleagues, En.Nordin, Ir. Idris and En V.V Yap for their support and assistance in the implementation of aforementioned educational delivery model in the case of the SS and DSP courses.

References