"Humanoid Robot", new industry or old thinking?

Overview of this article: At the Xiaomi autumn new product launch conference on August 11, a full -sized human bionic robot Cyberone was officially unveiled. Xiaomi's "people" plan detonated the capital city humanoid robot. One of the directions of uncertainty, as Musk said, "To make humanoid robots reach the same popularity as mobile phones, its market share will be difficult to estimate."

However, the symbol of the mature smart terminal industry is to form a perfect ecology of hardware, software, content and services. After the concept first, there is still a long way to land and commercialize the landing and commercialization of humanoid robots: bright prospects, and at the same time difficulty.

1. High R & D cost and capital investment are indispensable

2. The demand scenario is wide, the service industry or the incision point

3. Difficult commercial landing, the three major modules determine the commercial prospects

Humanoid robots are not new concepts, and industry giants are constantly entering this field. As early as 1927, the US West House created the world's first humanoid robot "Televox". In the past 100 years, companies have been trying to innovate technologies and improve their functions in order to finally create a "human stand -in" one day.

Therefore, although the Xiaomi robot has no mass production plan, the "making people" plan has detonated the capital market, and many concept stocks continue to blow well.

Whether the future of humanoid robots still stagnate at the conceptual level or truly enter the market and become the "organic unit" in production and life, which is worth seeing everyone.

High R & D cost and capital investment are indispensable

On August 11, Xiaomi released the humanoid robot "Cyberone". Earlier, Tesla CEO Musk also announced that it will release the humanoid robot prototype "Optimus" on September 30.

Affected by the news of the two major technology companies, the concept of humanoid robots quickly burst into heat.

A crowd of new energy vehicle companies, such as Weilai Automobile and Xiaopeng Automobile, began to lay out; Timsi, who focuses on the core technology of natural language processing, announced the study of exploring the application scenarios and related technologies of humanoid robots; Dyson also said in the next ten years It will invest 50 billion humanoid robots.

More importantly, the core parts of the robot account for 70%of the cost of the whole machine. If the humanoid robot can really achieve mass production, it will indeed open the global trillion -level blue ocean market.

The current track of humanoid robots is very lively. New energy vehicle companies, Internet, and home appliances are in charge.

But can reality really think so beautiful? You know, the research and development and implementation of humanoid robots are not only relying on conceptual piles and capital pursuits. Enterprises that deeply cultivate this track have to solve the problems of technology, capital, and commercialization one by one.

The first difficulty is technical difficulty. During the research and development of autonomous vehicles, Tesla has a large number of human -computer interaction, environmental perception and construction machinery, which can be reused on robots. Technical difficulties.

The second difficulty is that it is difficult to land. Humanoid robots are not like new energy vehicles. At present, I have not seen a humanoid robot that can really enter smart factories and enter millions of households for reference. Even Musk's "Optimus" (Optimus) It may be difficult to produce in the laboratory.

The third difficulty is that investment is difficult. Humanoid robots are a complex technology that requires long -term investment, which tests the accumulation of corporate capital. The acquisition of external investment is not easy. Although the concept is hot, in 2021, a total of 238 investment and financing incidents occurred in my country's robotics industry. Industrial robots are still the category with the highest amount of financing. many.

Although the mass production and landing difficulties of humanoid robots, the movements of the giants seem to indicate that the firming of machine substitutions and huge applications in the industry and huge demand.

The "Tesla Effect" will bring new vitality to the domestic humanoid robot industry chain, which not only means an incremental component order, but also the entry of more talents and funds, and the approach of the industry's inflection point.

The demand scenario is wide, the service industry or the incision point

At present, the development of artificial intelligence does not reach the real "general intelligence", and the hardware capacity of the robot has not yet reached the critical point. The more clear the demand, the more vertical the scene. Essence

However, the obsession and exploration of humanoid robots is not empty talk. Humanoid robots have the advantage of surpassing traditional products: suitable for a wide range of demand scenarios.

On the one hand, compared with the alien robot, the biggest advantage of humanoid robots is that all tools that can be used by humans do not need to design and create related facilities for the robot, just throw it into life; on the other hand, humanoid robots It also has multifunctional characteristics that can adapt to different application scenarios.

Existing technology is breaking through the difficulty of the diversification of the scene, integrating the scene to the virtual platform, and realizing service training in multiple scenarios.

Dazhen Robotic Company created the "Yuan Universe" of the robot - "Haiyuan World", allowing the majority of developers to build a robot digital twin on the virtual platform Learn to achieve large -scale low -cost and high -efficiency training.

This allows robots to serve all walks of life. Through large -scale virtual training, they can complete closed -loop learning and intelligent evolution in a large -scale virtual training.

After completing the study robot, it is also marching towards the service field in the real scene.

After completing the work and other tasks before the factory, some humanoid robots have begun to be commercialized. They will be used to reach customers, mainly used to welcomes, guidance and consultation.

Walker, a large -scale humanoid robot independently developed by Youbi Technology, has been upgraded for four major iterations after five years, and can complete home tasks such as massage, twisting bottle caps, and tea pouring water. Scene applications such as scientific research and development are also the world's first large -scale humanoid robot to realize commercial applications.

Dyson, who is "out of the circle" with home appliances, hopes to launch a humanoid robot that can do housework in the next 10 years. At present, Dyson has released the robotic arm that can pick up the bleach and pinch the plate, and try to create the next more efficient alternative home appliances with experience and techniques accumulated in the family service field with products accumulated in the family service field with products such as scanning robots, hair dryers and vacuum cleaners. Essence

Difficulty in commercial landing

Three major modules determine commercial prospects

Even if it is partially realized, the "humanoid robot" is still very far away from most of us.

The cost of high -production robots in humanoid robots makes it difficult for it to really enter the practical scene.

Based on cost -effective considerations, Xiaomi's cost of up to 700,000 yuan Cyberone has made it clear that it will not mass production.

Tesla's "Optimus" prototype is slightly smaller than Cyberone, and the price is lower than Cyberone. It is expected to be about $ 25,000 and is expected to be put into production in 2023. However, according to the previously expected price of $ 25,000 (about 160,000 yuan), there is a certain distance from the purchasing power of ordinary families.

Even the humanoid robot that has been mass -produced is not optimistic.

The most typical example is the power of Boston. The more advanced technology, the less business value is. At present, the company's valuation is only 1.1 billion US dollars, and the valuation has decreased by 66%compared with 6 years ago. French Aldebaran Robotics released only about 10,000 NAO sales that can play football, dancing, and boxing. It is more used for college scientific research and teaching competitions instead of implementing life scenarios. ASIMO, which can jump and use the stairs, has officially announced his retirement recently, ending a 22 -year performance service career.

Regardless of whether it is high or commercialized, its common difficulty comes from the development of artificial intelligence and does not reach the real "general intelligence", so the demand is not clear and the scene is not vertical. The key transition is very difficult.

To this end, while Xiaomi released Cyberone, it proposed that the three core technologies of "Environmental Period+Movement Control+Human Machine Interaction" can bring motivation to achieve the commercial prospects of humanoid robots.

First of all, machine vision is the core, and environmental perception technology continues to develop technical support for the commercial use of robots.

Xiaomi Cyberone is equipped with a self-developed MI-Sense deep visual module, combined with the AI ​​interactive algorithm, so that it not only has a complete three-dimensional space perception ability, but also achieves character identification, gesture recognition, and expression recognition. Cyberone has a visible basis and further has space perception and cognitive ability.

Secondly, the exercise control module, algorithm flexibility+hardware sensitivity determines the commercial effect of robotics.

Tesla's humanoid robot movement control performance is more prominent. Optimus's body consists of lightweight materials. It has human -level hands and feet. To achieve smooth and agile walking, the arms and hands depend on 12 electromechanical actuators, which can perform human refined operations.

Finally, the human -computer interaction module, natural language processing technology processing is the core of supporting the development of human -computer interaction technology.

The instructions of industrial robots are mainly through programming, which are applied in electrical electronics equipment, automobile manufacturing, logistics and transportation industries.

Humanoid robots have to identify natural language to achieve intelligent service functions such as housekeeping services and accompanying entertainment. Takhis, who focuses on the core technology of natural language processing, has also recently announced that it will also study exploring the application scenarios and related technologies of humanoid robots.

The market space of the humanoid robot has a wide space, but the technical difficulty leads to the difficulty of commercialization, and the high cost causes the penetration and acceptance of the product to break through.

Nevertheless, the two major technology companies have become a new vane: we are getting more and more close to the market turning point of the market that are getting close to the market. More companies and products that have been deeply cultivated in the robot industry for a long time also have the opportunity to use this Dongfeng to usher in the process of leapfrog development and industrialization and commercialization.

- END -