[frontier of Kobayashi Qilun's UAV]...

01
06

[frontier of Kobayashi Qilun's UAV] Bee Research on Vol.37 Evolutionary UAV | DRONE

Reverse engineering of living things

AI (artificial intelligence) has been called a craze for a long time, and the technology that leads the current AI popularity (considered to be the third AI craze in history) is a technology called deep learning. As the basis of deep learning is the neural network model. Simply put, as its name (translated as "neural network" in Japanese) implies, refer to the function of the brain to make the computer "think". That is to say, after a long time of research, the development has achieved great results, and now human work has brought about threatened AI progress.

Like this, the function and shape development technology of reference biology is called "bionics (bioengineering)" and "bionics (bionics)" and so on. Creatures tend to achieve what they want in a more effective way than the means that humans have come up with, even before they think of the means.

Therefore, the idea is to steal the technology of these creatures through "reverse engineering". In addition to deep learning, there are all kinds of examples, such as being bitten by a mosquito does not hurt, but being bitten by an injection needle will feel pain So can the mosquito's blood-sucking mechanism produce "painless needles"-such a situation (in fact, some scholars and companies are studying such needles).

In the field of UAV, it is also learning from the creatures flying in the air, trying to create a more efficient body and an organism with special performance.

In famous places, it should be the development of drones that flap their wings and fly like birds. A variety of researchers and institutions have begun to develop "bird drones" for a variety of reasons. In China, for example, there are reports that drones that do not allow surveillance to stand out have developed a body with a bird posture (this is just a form of camouflage, perhaps not bionics).

In addition, at Stanford University, it does not simply imitate shapes, but tries to imitate the functions of birds (especially its wings) to develop flying drones. It is named "PigeonBot (Pigeon Robot)" and has achieved a flying body.

You can also see from the animation that PigeonBot has wings that use real feathers. The team studied the flight mechanism of pigeons and found that real pigeons control their feathers through the movements of their wrists and fingers, and PigeonBot can do the same. It is said that in flight experiments, even sharp turns can fly steadily. It is said that by studying the various structures of other wings, it is possible to achieve a more stable flight even in bad weather.

[小林啓倫のドローン最前線]Vol.37 ドローンを進化させるミツバチ研究 | DRONE

More "beehived" drones

Having said that, Drone originally means "male bee". So, this is not to say that some researchers are trying to evolve drones by studying bees. Once practical, not only the name of the drone, but also the entity will be "beehived".

This was done by a team from the University of Sheffield in the UK who received funding from the British Research and Innovation Institute (UKRI) to reverse engineer the honeybee brain.

The aim is to develop a UAV (control mechanism) that can fly around the outside world more efficiently and safely. Bees can fly around without GPS, navigate accurately within a few kilometers and return to their nests accurately. And there are few CPU, and there is no need for communication capabilities to connect to the cloud. Therefore, we need to refer to how their brains process information and make rapid learning possible.

The content of the study is like this. First, install an ultra-small transmitter on the back of a real bee to make it fly directly. Not in special circumstances, but outside for the farm to collect nectar. The researchers then explained that by mastering the flight path by the transmitter and analyzing it, you can see what neural processes bees use to process information.

In addition, in the next study, it is planned to obtain data of neural signals by piercing electrodes into the brains of bees and flying directly in their prepared virtual reality (VR) space. The team has successfully modeled about 25 per cent of the honeybee brain, with the goal of putting it into practical use within five years to provide companies with AI for honeybee-like flight. In fact, he has also developed a 20-centimeter drone and is said to be studying experiments in cities.

Home page of the official website of Opteran Technologies

A university-derived company called Opteran Technologies has been established, which is not a model that ends with basic research alone, and it may be expected that the research results can be put into practice unexpectedly quickly.

Is it a bird or an insect, or some other flying creature? In any case, the drone in a few years' time may have become a more biologically active machine.