Dumbbell as a tool of the mind

Preamble

On Friday, March 31, we and my comrades are conducting intensive classes for robotics teachers, heads of CMITs and in general everyone who is interested in the applied aspects of pedagogy and the equally applied issues of robotics. The seminar will focus on teaching and mentoring in the field of IoT and robotics in general, and on its specific form such as the “project form”, in particular.

The concept of "project" has been included in the vocabulary of teachers for a long time and is used, often quite thoughtlessly. What is hidden behind this concept, what problems are there in teaching a design form? We will try to answer these questions during the intensive course.



To illustrate the theses, IoT-project “Smart Dumbbell” was chosen, about which under the cut.

Ambula

A variety of sports and near-sports gadgets penetrate our lives more often and take root in it more seriously. You can trivially schedule sports training using the GymBoom mobile application (or any other, a million of them), you can take steps-calories-kilometers with a special tracker, you can even straighten your posture by attaching a tricky controller to your belt.

On the one hand, the trend of “smart sports equipment” is quite old and completely formed: it was not born today, it will not die tomorrow, but how it will develop depends largely on you and me. On the other hand, the trend is wayward, and in its aerodynamic quality it is very similar to a proud penguin bird: until you kick, it’s not that it will fly, but will not move at all.



Bearing this in mind, some time ago, at our next school, as part of a group of young (and not very young) researchers, we developed the Smart Dumbbell project .

The tasks were as follows:

• Creating an IoT project that fully implements the functionality of Intel Genuino,
• Creating a training course of interest to illustrate the applied aspects of the project form of education: the project tree, various research areas for different microgroups of participants, and much more.

Dumbbells get smarter

About the trend has already been said above, although, frankly, our intentions were a little more modest than the shock of the mountains and the upheavals in the industry. The overall goal was to create a smart dumbbell: it differs from a regular dumb iron by the ability to control the number of repetitions in the approach and the correct path of execution for a number of selected exercises (this is a minimum task, note in brackets).

First things first. We had two bags of varied Arduino powder, seventy-five three-volt nutrition tablets, five trendy Intel Genuino, a half-spider mp3 player, and a mountain of enthusiasm, creative ideas and all that. And also a liter of coffee, a liter of coffee, a liter of coffee ... sorry, jammed. The only thing that bothered me was the accelerometer and gyroscope. In the world there is no one more helpless, irresponsible and immoral than a programmer trying to master working with such things according to the documentation. And I knew that pretty soon we would plunge into it.

about the project

Okay, okay, joking. Everything was a little less enchanting than in the original source, a quote from which I processed with a crowbar and file. However, the “quote” turned out to be marvelously accurate.

The new Genuino board from Intel has a number of interesting features that allow it to be used as the “brain” for a smart sports equipment. First of all, it is an accelerometer and a gyroscope - with their help you can determine the position of the device in space. Secondly, it is the Bluetooth Low Energy module that implements the IoT part of the functionality: connecting to a mobile “server” and exchanging data in low power mode.

Each research task — and the “Smart Dumbbell” is just that — has been living its own life at some point. Another problem solved within the framework of the general problem reveals a new branch of research - purely “Civilization” by Sid Meyer. Let's try to illustrate it with a living example.

Historically, the first problem on the dumbbell horizon was the problem of counting the number of repetitions (for example, in the exercise “bending the arm with a dumbbell”). The cycle of "lifting-lowering" of the dumbbell gives one repetition, therefore, cyclically changing parameters need to be fixed: these can be the start-end angles characterizing the position of the dumbbell in space or the speed-acceleration bunch demonstrating its change.

Having decided to push off from the corners, we come to the natural conclusion that data from the accelerometer will be required. Or from the gyro? Or from both at once? Hmm ...

Here we need to make a methodological remark (in serious terms): a person with a technical background, but little knowledge of the specifics of using all these gyroscopes, accelerometers, as a rule is sure that the accelerometer returns acceleration - well, judging by the name, it can still to return? The documentation “helps” in this regard, I can’t resist a quote:


It became immediately clear, right? On the other hand, well, acceleration can even be converted to mg! ..

We won’t bother you with an overly long story about how the views of researchers were transformed during the war with the accelerometer (and your own stupidity), we just mention in brackets that the accelerometer returns coordinate. But the coordinate of which - we will bypass the default figure: we must leave some kind of riddle to the thoughtful reader.

Just in case: "guessing the terrible secrets of the accelerometer" was by no means our only adventure. The most epic episode was when, with an almost finished project, successfully counting the number of repetitions, the most intelligent researchers managed to burn Genuino off - it turned out to be much simpler than it might seem at first glance. Everything would be fine, but it was almost the only Genuino in Russia at that time ...

And yet, “how long, how short”, as they say in these epics, we got a prototype that successfully counts the number of repetitions. Is the problem solved? Long live the new task: can we set and control the trajectory so as to signal an error when deviating from the standard? From one basic task we get two:

• Monitoring the current angular coordinate and checking its compliance with the limit values ​​(solved);
• Full restoration of the projectile trajectory (in the process of solving).

I would like to postpone a detailed story about the development of the project tree to a methodological seminar; the current story, however, would be incomplete without mentioning another distinctive feature of Genuino - working with Bluetooth Low Energy (BLE).

A few words about BLE

The task of interacting with a mobile application using the BLE protocol arose during the natural development of the dumbbell as an IoT project. The dumbbell should connect to the mobile "server", accept the user-selected exercise, the number of repetitions in the approach. During the approach, the dumbbell sends out the total number of actually made repetitions and the number of errors - when the dumbbell left the optimal path during the exercise.

Technical details of the implementation are of little interest. The only non-trivial point here is, perhaps, changes in the concept of permissions associated with the release of Android 6.0.
The methodological details are much more interesting. BLE as an idea is very different from the "usual" exchange via Bluetooth, which (externally) is nothing more than working with a serial port. Everyone who worked with the HC-05 knows how intuitive the Bluetooth sharing functionality is.

With BLE, everything is completely different: the concept of services and characteristics (simple in itself) will require, for example, a teacher (managing a dumbbell project) noticeable efforts to overcome the inertia of students' thinking. Here the forefront is the ability to briefly and easily expound the main provisions of the theory and present examples that implement the desired functionality, but not overloaded with details.

By the way, about this, in particular, we will talk at a methodical intensity.

From prototype design to prototype design

I would also like to mention such a non-trivial thing as prototyping and design. In educational robotics, design is most often spit on - it works, okay. Moving from pure uncomplicated functionality with wires sticking out from everywhere to a finished product that is pleasant in every way, we are faced with the following difficult questions:

• What restrictions are imposed on the physical parameters of the case (strength and temperature characteristics, ultimate dimensions, etc.)?
• How to place components in the space inside the case?
• How to provide access to components in debug testing mode?
• What are the requirements for the appearance of the device and how do these requirements relate to the above issues?

Illustrative example: in the original TK on a dumbbell, such a thing as ... a switch was ignored. When designing, accordingly, this moment was missed: simply because the “from pure functionality” approach did not provide for any switch - pulled out the power cord and hello.

As a result, a restart of the sketch requires a series of difficult manipulations with the body (you need to unscrew the bolts that secure the central part of the dumbbell handle - the pad) and the power wires. Plus, the dumbbell constantly consumes current from the battery, which leads to a completely unreasonable decrease in "motor resource".

To eliminate this design flaw, the circuit breaker had to be “embedded” in an already finished case.

How to do without such unpleasant pads, we will also tell at an intensity.

The planned event can be of interest to both teachers of robotics, curators of project activities in the field of IoT, and generally to everyone who is interested in non-trivial tasks in general. The workshop includes several discussion topics. Each topic will have a theory (a story about how exactly the ships plow the expanses of the Bolshoi Theater) and practice (a smart dumbbell in da action - right on the spot). In the plans:

• A discussion of what the design form is and how it differs from the usual classroom system.
• How to understand what to teach in the framework of an existing project task.
• The problem of project development: what to do when everything is already done?
• Practical skills of working with the project “Smart dumbbell” - do it yourself: we come up with new branches of the project and implement them as far as possible right on the spot.

If you feel like participating, register here .