Bringing the real world into the undergraduate lab
By Tom Lee Ph.D., chief education officer, Quanser Inc., Markham ONEngineering Electronics
At the annual NIWeek Conference in Austin, the Canadian control systems hardware company Quanser and the technology giant National Instruments (NI) demonstrated a series of new products that streamlines the hardware and software tool chain for control systems applications in engineering education and research. Additionally, the companies demonstrated a teaching concept that bonds conventional lab approaches to an engineering application framework and increases the relevance and motivation factor for the lab. The companies predict that this comprehensive approach will modernize control systems education, reflecting the major changes in industry over the past few decades.
Conventional control education
Control theory has existed for almost a century, but within the past three decades the actual practice of control has evolved to be a complex integration of computing, mathematics, and machines. This is the good news and has enabled advances in society. Our cars now achieve extraordinary efficiency. Surgeons can operate inside a patient’s body with minimally invasive techniques. The rover Curiosity is essentially an autonomous SUV on the ultimate interplanetary road trip. Indeed, modern control systems have become one of the most important engineering techniques to emerge in the twentieth century.
In spite of these great advancements in control engineering, the essential content of many undergraduate control courses has remained largely unchanged. Differential equations, Laplace transforms, root locus, Bode, Routh-Hurwitz, Nyquist, and more still comprise the bulk of a typical undergraduate course. Most of us educated within the past 50 years should now be squirming in a fog of unpleasant memories. However, as important as these techniques are, in today’s context, they are only a part of the workflow and indeed at most institutions the contemporary approaches to control have yet to be treated with sufficient rigor or consistency.
Modern control engineering and HIL
Figure 1 illustrates a typical process for modern control systems design. It is often typical to use a “V” form to illustrate the fundamental distinction between the more analytical processes of modeling and design, and the more practical concerns of real-world performance and testing. Collectively, many call the integration of these two dimensions as Hardware-in-the-Loop (HIL) control systems design.
As the diagram suggests, the traditional curriculum which samples a bit of the modeling elements and a bit of the controller design element is insufficient for the challenges of real-world, complex control systems. The missing parts include very practical things like sensors, data acquisition, and real-time software, and very conceptual things like non-linearities and unanticipated system dynamics. Finally, how does all this relate to the end goals of designing a high performance product?
Quanser Automotive Simulation for the classroom
At NIWeek, Quanser demonstrated its prototype called the Quanser Driving Simulation (QDS). The goal was to adapt classic, hands-on control exercise to a realistic automotive application. Figure 2 illustrates the basic configuration which uses the industry NI CompactRIO hardware platform that incorporates both a real-time processor and FPGA allowing data acquisition and processing to take place on one embedded platform. Quanser has developed a custom module for the CompactRIO platform, called the Q1-cRIO Module, which provides direct connectivity to Quanser control plants. All of these hardware pieces are tied together with NI LabVIEW system design software. Quanser has expanded upon the inherent ability of LabVIEW to make control design accessible by creating the Quanser Rapid Control Protyping Toolkit Software Add-on for LabVIEW. The toolkit abstracts I/O configurations in LabVIEW by allowing the user to program a wide range of both NI and Quanser data acquisition hardware with custom VIs that simplify I/O programming. This allows students to focus on the algorithm they wish to perfect and implement instead of spending excessive time configuring I/O. Historically, the classic DC servomotor is the most popular choice and the QDS approach used such motors as the foundation.
The new idea is the tight integration of both hardware and software components that map the DC motor and the control algorithm to an automotive application. In QDS, two DC motors were configured: one representing the speed of the wheels for velocity control and one for the position of the wheels for position control. The system input is from car game controllers with appropriate mappings for steering, throttle, and brake.
As a user “drives”, signal levels are computed by the analytical control algorithm and then sent via the DAQ to the motors. Conversely, the system via sensors and encoders will also read the motor speed and position and send it back to the algorithm for compensation. So if there is a disturbance on the motors, the driver can experience the response of the car. Additionally, environmental factors such as road geometry can be defined. Further, students can perform more advanced experiments such as autonomous driving (similar in concept to the Google autonomous vehicle) and wrestle with complex factors such as driver models.
Today’s students, having grown up with new media and video games readily engage with the system as it literally feels like a high-performance video game. More importantly, though, it exposes a much more realistic and industrially relevant engineering work flow. Indeed this style of HIL-connected 3D visualization is a very common technique within the auto industry. In essence, QDS delivers a rich, high-fidelity, theoretically rigorous modern engineering design workflow in a way that is engaging and familiar to today’s students.
Case study: University of Toronto
In 2011, the department of Mechanical and Mechatronic Engineering was one of the first to adopt this approach. The system was introduced into an introductory control systems course for undergraduates and a graduate course in advanced control. The results were uniformly positive as students and faculty noted the increased engagement and motivation of the students. At a practical level, instructors found the concepts to connect well to traditional theory and the lectures making the transition relatively easy. For the 2012 academic year, Toronto has increased its deployment of the concept to more student sections.
Beyond control systems
Fundamentally, the QDS approach acknowledges the changing face of modern engineering as design teams become more interdisciplinary and the actual techniques have evolved to accommodate increasing complexity. Recently, based on its experiences with QDS, Quanser launched a new initiative that adapts the philosophy to other engineering education domains.
One of the more interesting initiatives is the transformation of the first year programming course that all engineering students take. This course, like so many, has remained largely unchanged over the past decades. Students, using C, Java, Python, or other popular languages, work through abstract concepts for data types, program flow, and algorithms, often without any real context.
Recently, at a meeting of prominent Canadian research professors in field robotics (i.e. autonomous ground, aerial, and aquatic robots) at McGill University, renowned professor of computer science and robotics, Dr. Gregory Dudek of McGill, suggested that the future of computing was in autonomous robotics. He commented that the world only needs so many programmers who can build compilers, but all innovative computing is now moving off the desktop and into mobile devices including autonomous mechatronic devices. Consequently, the modern engineer needs to be proficient at managing both the purely digital as well as the physical when dealing with computers.
Quanser’s response is the conversion of the QDS architecture to one that offers the s
ame richness and relevant context in a way that is suitable for first year and even high school students. These concepts are now being adapted to replace traditional curriculum in the first year programming course.
Case study: University of New Mexico
Working with the Department of Electrical and Computer Engineering at the University of New Mexico in Albuquerque, Quanser has developed a series of exercises that connect the fundamental programming concepts to the task of controlling Quanser devices such as its series of helicopter flight dynamics systems. Unlike a controls systems course, the control aspects are encapsulated in pre-built modules, and students are focusing on how you get data from a real system, how you manipulate the data in the computer, and how do you send signals back. Like QDS, techniques of visualization also play a significant part.
The software dimension starts with LabVIEW where programs are written in more intuitive block diagram oriented ways. It is somewhat like a dynamic version of pseudo code that is often taught in the traditional courses. Once the basic manipulation of the data is mastered, students will then be introduced to formal C. The goal is to develop equal proficiency in C programming, as with the previous course, but in parallel, develop knowledge of an industry-standard application tool (LabVIEW), and more relevant concepts in real data acquisition and manipulation.
Quanser’s prediction for the future is very optimistic. Like most things in technology, time has brought the basic tools and methodologies to a robust and usable state that even undergraduates can command these tools proficiently. The challenge of these new lab concepts is really to facilitate the transition from the old to the new. Quanser believes that the focus on what industry is currently doing in the “real world” and reconciling this to a strong theoretical foundation, and adding sufficient motivating fun, is the right combination for a more effective education experience.
The partnership of Quanser and National Instruments allows students to learn fundamental concepts in control and other engineering fields in an engaging and relevant manner. Students learn conventional theory in a more engaging way through the combination of Quanser’s hardware systems and industry-standard tools such as LabVIEW and NI CompactRIO. This innovative approach prepares the students with the appreciation of the theoretical foundations and actual, practical, implementation knowledge. Both Quanser and NI believe that this combination is critical for preparing future engineers to meet the complex challenges of modern engineering.
***For more information on the design and manufacture of advanced systems for real-time control design and implementation from Quanser Inc., go to http:www.xxx…