As a command was issued, the first man-eating fly micro-attack drone in the swarm array of man-eating flies broke away from the cluster and immediately flew towards the arranged obstacle channel at high speed.
The entire obstacle course is very narrow. It is actually separated by wooden boards and acrylic boards. The narrowest part is only a few centimeters wide and requires continuous turns. Moreover, the light transmittance of the acrylic boards greatly increases the number of unmanned passages.
The detection difficulty of the sensors on the aircraft.
As the first man-eating fly micro-attack drone entered the obstacle test channel, the speed immediately dropped. However, the speed was still very fast. The entire drone body shuttled quickly in the small space, avoiding all directions.
Obstacles move forward in the narrow obstacle test channel according to the 'Z' shaped flight path.
However, this man-eating fly micro-attack drone did not fly for long in the obstacle test channel. Just when it passed a narrow acrylic obstacle, the tiny fan blades on the drone were hung on the acrylic plate.
The edge caused the drone to lose control, directly mounted on the wooden board on the other side, and crashed directly into the obstacle test channel.
Coco, has the flight data transmission been returned? Wu Hao asked.
A window appeared on the screen next to it, with Coco's avatar in it. It reported: "Sir, the flight data has been transmitted back to the database and shared with other drones through the cluster control system data link."
Wu Hao nodded when he heard this, smiled and waved his hand: "The second test begins!"
Okay. As Coco responded, the second man-eating fly micro-attack drone flew out of the cluster array and began to fly into the obstacle test channel. After obtaining the flight data of the first drone, this drone
The speed was obviously much faster during the obstacle crossing flight in the first section. Obviously, it referred to the flight data of the first drone.
This is also a major feature of the cluster control system, that is, resource sharing. All detection data can be shared, including flight routes, etc. The previous drone will transmit the relevant obstacles that have been detected to the swarm array.
, this also allows other drones to know what the obstacle sections here look like, so subsequent drones can follow up along the flight path of the drone in front.
Of course, it is not a complete follow-up and imitation. These flight routes are only used as a reference, which can greatly reduce the detection data processing capabilities of the drone, thus allowing it to fly faster in narrow channels.
This does not mean that it will completely give up detecting the surrounding environment data information, because in reality, no matter how short the flying distance of these drones is, the surrounding environment may change during this interval, resulting in the loss of flying companions.
The position and angle of the obstacle change.
At this time, if you follow the flight path of the plane in front, it is very likely to hit an obstacle and crash.
Especially this kind of micro drone is more susceptible to changes in the surrounding environment.
To give a very simple example, the branches and leaves of trees and plants sway with the wind. Large drones may not need to consider these, but micro drones must take these into account. It is possible that even a leaf can completely blow the entire drone.
The machine stopped.
Soon, the second drone passed over the wreckage of the first drone and began to fly forward, but compared to the previous flight speed, the current flight speed dropped significantly.
The drone flew forward about one meter, and because it could not turn in time, it was mounted directly on the obstacle, causing the entire drone to be crashed.
Then the third drone broke away from the swarm array and began to continue testing.
Doing this is not playing or wasting, but training the drone's intelligent flight control system. Through a large amount of obstacle-crossing flight data, the flight control and obstacle avoidance performance of the drone can be greatly improved.
Especially the obstacle avoidance and crossing performance in such a small space.
In fact, they had already accumulated a very rich amount of drone flight data before this. After all, they were the first to start out in drone technology and have developed along the way. It can be said that they have accumulated a very rich experience in this area.
However, these are all large, medium and small drones, and even the smallest drone they make is much larger than the current one.
Therefore, the flight control data accumulated by previous drones may not necessarily be suitable for this kind of micro-drone. Some of them can be used as reference, and the rest requires the drone to fly independently. Through the continuous accumulation of flight data, then unmanned aerial vehicles
The aircraft's intelligent flight control system can learn how to deal with the same or similar situations.
In fact, the entire testing process does not require Wu Hao's participation at all and can be left to other people. Of course, there is a more convenient and convenient way, which is to leave it to Coco and the intelligent simulation robot. It only needs to wait for the experimental data results.
That’s it.
However, there are some things that artificial intelligence cannot understand. Although they are smart, they cannot have as much divergent creative thinking as humans. Sometimes too accurate data, or too rational judgments are not conducive to innovation. In fact
Many innovative inventions are produced by inadvertent accidents or even mistakes. However, artificial intelligence does not care about these mistakes or instinctively filters these mistakes and pursues successful results.
They can be very rational, but they cannot be emotional. This is the biggest technical problem of artificial intelligence technology in the world at present. How to make them emotional is the key to allowing them to develop autonomous consciousness.
Wu Hao obviously broke through these and used the things in his brain to create Coco. However, limited by the current world's comprehensive technological level, he was unable to fully utilize the trends in his head, one by one with intelligence.
Make some progress.
For example, from the birth of Cocoa to the present, the performance of Cocoa has been continuously improved. Compared with the beginning, it is now smarter and closer to humans. Otherwise, Chen Keren would not be able to be so successful in the company.
long time.
Now, although he is directly involved in the entire experimental process, he does not actually need to do it himself. He only needs to give instructions and Coco will be there to complete them according to his instructions. Of course, some experiments still require him to do it himself.
.Although Coco is very smart and considerate, she cannot fully understand what he means.
And many things have two different feelings when they are said and done, just like the old saying "can only be understood but cannot be expressed". Some things can be thought of by the brain and can also be made, but how to express them accurately
Come out, this requires thinking.
This is why it is difficult to write papers on many results. You must express the results intuitively and accurately without allowing readers to misinterpret them.