If a young person wants to
be a complete baseball player, he must be able to field, throw, run the bases,
hit with power, and all these skills must be applied in an actual baseball
game. To achieve this goal, he learns all these skills at the same time,
improving gradually in each one while playing actual games and, over time,
develops into a complete baseball player. The result is more than just the sum
of the skills learned, but a sense of confidence and savvy that makes him a
In the area
of mechatronics, the necessary skill set includes modeling and analysis of
multidisciplinary dynamic systems, analog and digital control systems, and
sensors and actuators with the necessary electronics. Theory and practice must
be in balance when mastering these skills. If ?"playing a game" means putting
these together to create a system to solve a problem, then that rarely happens
in engineering education, and if it does, it happens for only a few students
who aggressively seek out that integrated, total experience. We devote separate
courses to each skill and somehow think that learning each skill very well will
somehow magically enable the student to graduate and critically think,
integrate it all, and solve a real-world problem. In the baseball analogy, this
would be utter madness, yet in engineering education, it is routine.
I am speaking
from first-hand experience, as I am presently teaching a course to 60
second-year engineers called Electromechanical Engineering Systems, with 16
personal contact hours each week - 10 in studios and six in classes. There are
no teaching assistants, just graders. Everything is done in the context of
real-world engineering practice and problem solving.
works as follows, with mathematics and physics learned and applied as needed: A physical engineering system
(electrical, mechanical or electromechanical) is chosen that must behave
dynamically in a specified way. The system is first physically modeled with
simplifying assumptions and then mathematically modeled by applying the laws of
nature and appropriate component constitutive equations to the physical model.
We start with a system whose model is first-order and study it from both
time-domain and frequency-domain perspectives. Putting the mathematical model
in a standard form (i.e., time constant, steady-state gain) allows an engineer
to relate performance (e.g., speed of response, steady-state error, relative
stability) to the hardware parameters in the physical model. As is often the
case, the system cannot meet performance specifications operating open loop. A
feedback control system is then designed and implemented. Closed-loop PI
control of a first-order model results in a closed-loop differential equation
that is second-order with a numerator zero. Second-order dynamic systems are introduced
naturally, as part of the process, along with the effect that a real zero has
on ideal second-order behavior. Again, time-domain and frequency-
domain perspectives are emphasized.
Once PI control gains are selected
by a combination of pole-placement and simulation iteration, it is time to
build the system. First an analog op-amp system is built with a difference
amplifier and PI controller. Loading effects must be addressed, as must the
limit on the control effort due to op-amp implementation. Measurements are
compared to model predictions and model adjustments are made. Digital control
with the Arduino microcontroller, inexpensive and open-source, is then used
with the MatLab/Simulink Real-Time Workshop providing automatic code
generation. Issues such as pulse-width modulation and low-pass filtering (which
introduces a real pole), saturation, and A/D and D/A resolution all can be
addressed in simulation and then easily in hardware implementation. Loading
issues are again addressed with buffer op-amps.
In this scenario, the students are
"playing the game" from the start. In past columns, I have decried engineering
silos and engineer comfort zones, both in industry and academia, as the two
biggest obstacles to innovation. Add this educational deficiency to that list.
Let's get our heads out of the sand!
Click here to visit the Mechatronics Zone
for the latest mechatronics news, trends, technologies and applications.
John, sounds like your class is just what is called for in terms of bringing a more practical approach to a modern engineering education. I hear over and over again from Design News engineering readers and the vendor community about how systems engineering principles are more important than ever. While it's obviously far from a new concept or discipline, it does appear organizations are still stymied by trying to collaborate between the separate disciplines of mechnical, electrical, and software. Perhaps arming the next-generation of engineers with hands-on experience and in-the-trenches best practices for cross-disciplineary work will finally remove the barriers to early collaboration and foster less iteration in design cycles.
Are they robots or androids? We're not exactly sure. Each talking, gesturing Geminoid looks exactly like a real individual, starting with their creator, professor Hiroshi Ishiguro of Osaka University in Japan.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.