Xiaomi is a Chinese consumer electronics company that was famous/notorious (at least a few years ago when this was still a big deal) for its iPhone clones. Now, Xiaomi is proud to present another product line that has great potential: connected robotic toys. How does that work for us Lego Ev3 enthusiasts? Let’s find out.
I do not want to say that the Mitu robot is a clone of Lego EV3. Let’s say it is “inspired” by the latter and is designed to be somewhat compatible with it. This is not a problem for me as I think it is actually the only way to do business: what could be a better practice than to make your product compatible with the major player of the field and to offer a little improvements in key aspects?
The key advantage of the Mitu platform is its cost. Currently the Mitu set is priced at 499RMB, which is about 72USD at current exchange rate. The Lego EV3 home edition, on the other hand, sells for 345USD. No surprise here. Just like the iPhone: we don’t need to be better; just more affordable. And this strategy will always work. Innovation means huge research investment (both time and money) and it involves big risk. Cloning is safe; and it is still profitable.
So much for the business side of the product. This is not the real concern of this article, which is the following: is the Mitu set an alternative to Lego EV3?
To answer this question fairly we need to look at the technical specifications of the product, both what is available now and its future potential. We also need to look at the available curricula, which means: is this suitable for classroom adoption?
What is included?
Piece-wise, I think Xiaomi did a really good job. There have always been Lego clones manufactured in China, and I have tried a few of them. IMHO, this is the best quality clones I have seen so far. It even surpasses the quality of non-Chinese Lego-compatible brick system such as Mega Bloks (Canadian company). Apart from some small problems (sometimes axles are a little bit too hard to push and some pieces such as the tread link are not strong enough), the assembling experience is very similar to that of authentic Lego technic sets. Finally, there are more pieces (close to 1000) included in this set than what the EV3 set offers (500?)
Compare to the EV3 system, Mitu currently doesn’t any sensor, except the built-in gyroscope. As a result one cannot build anything with autonomous behavior. Also, there are only TWO wheels included in the set so you can’t build a car with it — this is a serious mistake.
Not including the other sensors (infra, touch, color, ultrasonic, etc.) could be a business strategy: to lower the research investment of the first stage of this product. But the truth is, with the addition of those sensors, the cost of the package will rise dramatically, and quickly to the point that it will no longer feel much cheaper than the Lego EV3 set. Or maybe the sensors are there already, but in order to lower the package to a certain price point, the company chose to hold onto them.
The Mitu set doesn’t come with a remote controller. I feel this is a bad decision. Although any serious engagement with the set would require a tablet or smart phone. Having a physical remote controlling device would considerably enhance the appeal of the system (and the remote in EV3 set dubbed as infrared beacon).
The central piece
The central piece of the set, the intelligent brick, is lighter than the EV3 counterpart. The good news: it has rechargable battery built-in (saves 50 dollars!). It has four USB-C type ports (smaller footprint and wider compatibility in the future). It has much better bluetooth connectivity (definitely 4.0). This said, with the next windstorm refresh, these pros may no longer be true.
As you can see clearly, the Mitu intelligent brick doesn’t have a display or any navigational buttons (except the power on/off). So it cannot do what EV3 can do in terms of display black and white patterns. I agree with this design decision: the display functionality is not really useful and it adds too much to power consumption and system complexity. In the future, one could make a device that is dedicated to display and have much better specifications. The power button on Mitu is extremely sensitive: it is really easy to turn it on. And it is not difficult to turn it off by holding it for a second or two. The whole process is lighting fast compared to EV3, where the booting process feels like waiting for an old PC to start up.
Now for the APP. This is not an exhaustive review of the app as there is not really much to look at now. Also annoying is the fact that there is not even a iPad version of the app. Running it on iPad means using the iPhone version.
Apart from basic setup and config, building instruction displays, the main functionality of the app is that it offers three modes to operate your model, once it is built: the route mode, the controller mode, the programming mode. The route mode is something we see often in lower end connected toys such as sphero or dash/dot. It allows you to draw a path and then your robot will try to follow that path.
In the controller mode you can move and turn (if your model allows it). There is also a gyro mode which leverages iPhone/iPad’s built-in gyroscope to steer. Both work as they should.
Finally, the block programming interface. This resembles Blockly but the interface can use some fine-tunings.
I use Chinese but English language interface is available if you can put up with badly translated nonsense (can the company afford to hire someone who speaks English?).
Models and Curricula
Currently there are four official models: robot, T-rex, airplane and motobike. Apart from the first, you need to download building instructions from the app. In my case, the instructions are not even there and I have to look for them elsewhere! Compared to the EV3 app, however, Mitu is SERIOUSLY lacking as these models are NOT designed for programming actions! I have not built all of them but from what is looks, the airplane is basically a display model and the motorbike cannot turn! Again, this goes back to the initial decision as to whether include a medium motor, which can serve steering purposes.
What I mean is that the designing of this product should follow, instead of being followed by, the design of programming curricula. Instead of starting from a primary model (such as the self-balancing robot) and then try to come up with alternative models one can build — which is the design process of Lego Technic sets — the designer should have come up with ways to program the block, ways to teach robotics first, and then integrate those lessons into a bigger project that is the model.
Basically, this means that Xiaomi should have hired someone who teaches robotics (such as EV3) and then recommend the kind of pieces, sensors this set should include. And the programming curricula will come naturally out of this process. Instead, it is designed as a typical Lego Technic set that emphasizes different models to build, how realistic it looks, etc. But the situation is not hopeless: if somebody is committed to developing such a curriculum, the Mitu system then can unleash its true potential. Without this, it can never become an alternative to the Lego Mindstorm series.