Mingmin Xiao Xiao came from Aofei Temple. Nonono, the closest distance you are to autonomous driving may be your own sweeping robot. In addition to scanning rooms and building self-built maps, it can understand the "terrain" of your home even more than you:

Mingmin Xiao Xiao from Aofei Temple
qubit Report | Official Account QbitAI

What? Your impression of the sweeping robot is still on the "dung-feeding wall-coating" as a zi?

Nonono, now the closest distance to autonomous driving may be your own sweeping robot.

In addition to scanning rooms and building self-built maps, it can understand the "terrain" of your home more than you:

can also scan directly through the camera to obtain a 3D map with depth information, just like Apple iPad Pro :

not just a 3D map that scans shapes and depths, but also a scene understanding map with its own tags:

That's right, this is indeed the look of the modern generation of sweeping robots. Not long ago, the latest sweeping robot Dibao X1 released by Ecovacs , all of the previous articles have been realized. How smart is

?

You just need to say "OK YIKO, go to the sofa to clean", and the sweeping robot can accurately complete the master's task:

△Voice assistant YIKO

This makes people curious, from artificial intelligence to "real intelligence" , from The semi-automatic "tool man" evolved into a "steward". What has the sweeping robot experienced?

sweeping robots do not bump randomly, why is it so difficult?

In fact, a sweeping robot needs to consider many issues if it wants to complete a cleaning work.

What things need to be cleaned? What should I avoid? How far away should you avoid it? How to clean the ground around the items after avoiding them? What to do if you encounter a walking person or pet...

The "simple sweep the floor" you see, the sweeping and obstacle avoidance decisions that the sweeping robot needs to make is actually as high as more than 100,000 times .

In order to better cope with complex cleaning situations, the " eyes " and " brain " of the sweeping robot have achieved super evolution.

Now, sweeping robots can not only "see" them, but also recognize them around 2-3cm away from them, and then complete a handsome avoidance.

Behind these small actions, it uses LDS lidar, 3D structured light, dToF, AI algorithm and other technologies that are also equipped in aerospace, autonomous driving and smartphones.

: Taking the operation of "avoid slippers", it is not easy to make the sweeping robot do it well.

. Just identifying objects requires the combination of 3D structured light, dToF, and AI algorithm technologies.

3D structured light technology allows sweeping robots to "see" objects and "understand" the environment, that is, through sensors, they collect information such as the depth and position of the object.

uses a dot matrix projector built into the sensor, and uses it to emit thousands of tracking points to the object to be identified, and then uses the infrared camera to take a snap.

uses the data collected by infrared cameras and combined with built-in algorithms to quickly obtain 3D information of objects.

Our commonly used face recognition is to apply this technology. If you have an Apple phone, you can see the dot matrix projector and infrared lens required for 3D structured light from its small bangs .

At this point, the sweeping robot only does the preliminary work.

is not enough to obtain three-dimensional information on roadblocks. Today's sweeping robots must understand what roadblocks are.

This requires the AI algorithm to appear, allowing the sweeping robot to "understand" objects and realize classified cleaning.

Through the target recognition algorithm, the sweeping robot can classify and judge common objects in the home environment.

For example, small debris and carpets are items that need to be cleaned, while slippers, socks, wires, etc. are objects that need to be avoided.

looks very simple, but in fact, it requires very high computing power of the robot.

, like the latest Dibao X1 released by Ecovacs, uses a special AI chip for horizon at the intelligent driving level.

You should know that the computing power of this chip can reach 5TOPS. In the official promotion of Horizon, its capabilities can reach the L2+ level of autonomous driving.

So, with the self-driving technology, the sweeping robot is really not going to "spray the wall" or drag your socks to run around the house.

After you can see and recognize objects, the most critical step is: avoid .

From the action of avoiding slippers, it can be seen that the sweeping robot does not avoid it after hitting the slippers, but when it is a few centimeters away from the slippers, it determines the obstacle in front, which is very silky throughout the process.

This reflects the magic of 3D structured light combined with dToF technology.

D structured light can also obtain the depth information of the object while obtaining the 3D information of the object, but the distance can be achieved relatively limited; dToF can supplement the depth information of the long distance.

It transmits and receives multiple optical signals to the object within a single frame measurement time, records the flight time of each signal back and forth, and makes histograms, using the flight time with the highest frequency (ToF) for use in the most common (ToF) Calculate the distance between the target.

△Photo source: Wikipedia

This is also why the sweeping robot will brake and readjust the route in time after reaching the safe distance of the obstacle.

, and with the support of AI recognition, the sweeping robot can also make different avoidance strategies for different types of items.

For example, slippers, socks, and chairs that belong to safety items will approach the object as much as possible and will adjust the route only when the distance is about 3cm; for more dangerous wires, the safety distance will increase to about 5cm. Scan

, but not completely?

Do you think this solves the problem?

In addition to the obstacle avoidance technology of the big show operation, if the sweeping robot wants to complete a perfect cleaning task, it must have a full understanding of the home environment and a clear route plan.

△Snakeskin movement vs planning route

This is what we need to look at the LDS laser navigation and dToF technologies that help it get rid of "snakeskin movement".

In the small disc on the top of the sweeping robot, there is a rotatable lidar hidden.

It can rotate and emit laser light at high speed, and judge the distance between itself and the target object based on the time when the laser light is reflected back after it is irradiated on the target object, so as to determine the position and achieve positioning.

, like the random sweeping robots that had not been navigation before, often go wherever they go, without any planning at all. The final result of

is: it was scanned but not completely scanned, and the owner had to scan it again by himself.

and dToF technology can further improve the navigation accuracy.

's biggest advantage is that is almost uninterferenced by light , which can ensure that the sweeping robot still maintains super high accuracy when working in a dim environment.

In fact, the current sweeping robots will scan the home environment in a comprehensive manner before working, and build an 3D map .

still has to rely on the strong combination of LDS lidar and dToF.

combined with SLAM algorithm, the sweeping robot can start from any point in a completely unknown home and 360° comprehensively identify home environment information.

can also build an accurate 3D floor plan, identify complex floor plans and plan cleaning routes. In the subsequent daily use of

, it will further refine the built floor plan, re-correct the deviations, and ultimately allow the robot to completely "clear" the terrain of the home.

In addition, LDS laser navigation can also allow robot to self-locate in real time and record it.On the one hand,

allows the robot to clean according to the planned route, and on the other hand, it can continue to work after charging to work in the place where it has not been cleaned, avoiding the missed and repeated cleaning of the robot before. .

also breaks these "stereotypes" impressions

BUT. As we all know, the inhuman "artistic ability" of sweeping robots is not only running around.

swept halfway through the dust box, and went on strike on the spot; the floor, carpet and cement floor were dragged together, and the replacement was accelerated; the virtual wall was accidentally forgotten to set up, and went out to the community to take a few turns...

But now, these have become " Traditional impressions”.

In response to the situation where personalized cleaning cannot be achieved without virtual walls, the latest sweeping robot can meet the cleaning needs of any location through AI semantic segmentation, 3D target positioning and recognition, and combined with path planning technology.

In order to make the interaction of demand more direct, AI intelligent voice interaction function is now added, so that it can be assigned to work by saying a word. Especially for the elderly, the sweeping robot is much easier to use after it has voice control.

For example, if you don’t need a sweeping robot to clean the whole house, if you want it to clean up the dog food that is poured under the sofa or spilled oil in the kitchen, you just need to call on a voice assistant named YIKO , "OK YIKO, go to the sofa and sweep it. The robot can understand the instructions of "sofa" and "sweep" and arrive at the designated location to work accurately.

In response to the "automatic strike", the latest sweeping robot base station has integrated functions such as automatic dust collection, automatic mop washing, sterilization and drying, and truly realizes the effect of "completely liberating both hands" in the entire cleaning process;

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : There is also the situation of mopping the floor without any difference. The latest sweeping robot has equipped AI automatic recognition to accurately identify the material of the floor and choose the most suitable cleaning method, such as mopping the floor on wooden floors and ceramic tiles, while cement floors and carpets. Then use sweeping instead... all the flaws that the first generation of sweeping robot players can think of, the latest Ecovacs Land Treasure X1 has basically been solved. The technological explosion accumulated under

Back to the industry itself, sweeping robots seem to be just in time to catch up with the era of technological explosion.

is just like the smartphones that emerged back then. The technologies that can be equipped with are inexhaustible. Whoever can carry more cutting-edge technologies at the same time will have a greater advantage in the market.

But did the technology really explode without warning?

is not the case.

As a type of technological product, the development of sweeping robots has gone through two decades and 2 iterations.

has changed significantly from the "seeking freshness" in the 1.0 era to the technological breakthrough in the 2.0 era, and then to the real intelligence in the 3.0 era. The sweeping robots in each era have undergone very significant changes compared to the previous generation.

1.0 era refers to around 2001 to 2010. At that time, the sweeping robot was not intelligent. The built-in algorithm could only meet the most basic sweeping requirements. The switch could only be carried out through physical buttons, and basically no other functions were available.

△It seems that although the sweeping robot in the 1.0 era, although technologies such as inertial navigation and ultrasonic navigation appeared later, their disadvantages were also obvious. The hardware ( gyroscope , accelerometer, etc.) that relies on inertial navigation is not accurate enough and still relies on collision algorithms; ultrasonic navigation is too disturbed by noise and is too heavy, and the sweeping robot is in the stage of technical accumulation as a whole.

There is indeed a group of consumers who are willing to take the lead in this stage. Supporters are often the few who believe in technology to change the world, and mass consumers may not pay.

2.0 era appeared around 2010, during which AI algorithms mainly based on deep learning began to emerge, and sweeping robots also began to usher in technological breakthroughs.

Around 2010, the cost of lidar technology plummeted. The positioning and navigation technology of sweeping robots with lidar as the core began to appear, solving various problems in inertial navigation and ultrasonic navigation. Active navigation planning has a mature solution. ;

015, the industry began to show the " sweeping and mopping integrated " technology that combines cleaning and mopping into one, subverting people's impression that sweeping robots can only sweep the floor. Nowadays, this function has become the wealth password for market share 92% ;

In early 2017, visual algorithms with AI technology as the core and interactive methods began to enter the sweeping robot industry, further providing the ability to identify objects of sweeping robots. It has truly shortened the communication distance between machines and people... In this stage of

, sweeping robots have certain practicality in intelligence, functions and interaction, but most of them still remain in the state of semi-automatic . Although buyers have The number of people who stay at home still cannot completely clean up.

Now, with the launch of Dibao X1 by Ecovacs on September 15, the sweeping robot has also ushered in the 3.0 era , which is the ideal form of "future robots" that people call.

is closer to maturity in terms of intelligence, interaction and function. The robot's vision also has 3D depth and object information. The interactive ability and technology are deeply integrated, allowing the machine to have "thoughts" and fully automatic completion of the entire cleaning process. .

During this stage, the sweeping robot truly emerged from the closed loop of " technology to define the requirements ", which is as convenient as a smartphone.

As the gradual iteration of sweeping robots, Ecovacs has been in the middle of the wave.

As early as 2009, Ecovacs had entered the field of sweeping robots as a pioneer in China. At that time, the industry was still in the 1.0 stage, but Ecovacs had accumulated dozens of technical and design patents.

In 2013, As the first person in China, Ecovacs applied LDS-SLAM to the field of sweeping robots, leading the industry into the 2.0 stage and embracing technological upgrades.

In 2015, it was the first to launch the integrated sweeping and mopping technology to promote sweeping robots to enter the era of integrated sweeping and mopping; in 2017, it was the first to launch a 3D-drive three-dimensional cleaning system... The "industry first" experience like

, Kovo There are many more: layout VSLAM and LDS-SLAM dual-tech routes ; launch AIVI for obstacle avoidance; apply TrueDetect 3D structured light obstacle avoidance technology to sweeping robots...

from L DS-SLAM, sweeping and dragging Integrating AI obstacle avoidance and interactive algorithms, the new technologies that Ecovacs constantly break through have given more practicality to sweeping robots and truly realized a certain degree of automation. This is also the symbol of the era of sweeping robots 2.0.

After that, Ecovacs did not stop moving forward. In 2020, R&D investment alone reached .338 billion ; in the first half of 2021 alone, .202 billion has been invested in R&D in the first half of 2021 alone. .

And as mentioned before, the latest Dibao X1 produced by Ecovacs has opened a new year for home service robots.

If the 2.0 sweeping robot can quickly plan and complete the whole house cleaning, Dibao X1, as a sweeping robot, is more like an intelligent butler:

has been upgraded to a positioning and navigation plan for 3D map construction, which is stuck in lidar construction with the 2.0 era. The 2D effect in the figure is already very different, and the robot truly "sees" the world in the eyes of humans.

uses AI voice interaction technology to sweeping robots, which truly enables the robot to "understand human words" and perform tasks that humans need. This is the first in the industry;

's cleaning function of sweeping robots is also simple "Sweeping and mopping" has ultimately evolved into a robot that integrates automatic hydration, mop washing, sterilization, drying, dust collection, etc.

is right. From the perspective of ordinary people, Dibao X1 is closer to our cognition " real lazy artifact ".

is different from previous generations of products, and the era of sweeping robots 3.0 has arrived.

Review of the process of sweeping robots, from the 1.0 era that is regarded as a "sweet rib" to the comprehensive technological breakthroughs ushered in the 2.0 era, it has evolved to the "comprehensive intelligence" of the 3.0 era, and it is also the evolution of intelligence, functions and interactive technologies. a continuous accumulation result.

Among them, Ecovacs is not a player who just catches the nodes of the times, but a pioneer in the process of daring to take risks, daring to overcome obstacles, and setting off technological innovation behind it.

These changes are not achieved overnight, and require the evolution of quantitative change to qualitative change. Although some people question and mock them in this process, those who insist on deepening their cultivation of technology and constantly pursue the ultimate in products will eventually become pioneers in the new era.

How convenient and comfortable our home cleaning will become in the future, we have to rely on the efforts of these technological pioneers to explore.

—End —

Quantum bits QbitAI · Toutiao Sign

Follow us and learn about cutting-edge technology dynamics