In the field of smart home, although many companies have long been involved, but one of the most promotional and reliable technology model is still being explored. Recently, Zhang Yongqian, vice president of the horizon, attended the AWE forum, around the "smart home era opportunities and challenges" published the horizon view, and introduced the horizon in this area of technological breakthroughs and solutions.

  China's smart home market is very large scale

China has 100 million middle-class families, the average annual income of 100 million households in the 300,000 to 700,000 yuan or so, they are smart home market, the main consumer groups, which shows the size of the smart home market. Of course, when it comes to smart home, the first impression of many people is electronic equipment, this market is indeed great, there are 1 trillion. But in fact the smart home market real "potential shares" is the family consumption and services - the annual market can reach 50 trillion, of which 10% through the online. If the business can be more convenient, faster and more accurate to help consumers for family consumption, to provide services, then it will be a great success.

  Smart home market three trends

Through the development of this market trends, we can find three more obvious changes that have occurred or will happen soon.The first is the physical entrance, the most important entrance is the mobile phone and PAD, etc., this entrance is very crowded, a mighty force crowded wooden bridge. A very obvious trend in the future is that the entire consumer and service access is distributed to the various intelligent devices. In the future at home, every hardware is a standalone APP. You find that there is no drink inside the refrigerator, you do not need to take out the phone, unlock, open APP and so on to buy, this user experience is very poor. You can tell the refrigerator directly, in the afternoon to send me a dozen beer over.  The second is the interactive way, the most traditional is the keyboard and touch screen. With the development of artificial intelligence, voice technology will soon spread to our lives. But we believe that in the near future there will be a more important, more heavyweight way, that is, through visual interaction. Because now we can see, whether through touch or through the voice of interaction, in fact, are the initiative of human decision-making, the machine passive acceptance process. While vision is active interaction.  Visual interaction brings a third change, that is, the machine take the initiative to make decisions - people do not need to pay attention to many family things, and even orders are not under the machine will be able to complete the service in the most appropriate time. For example, now there is a laundry service, we can imagine that the future inside the house on the balcony put a large basket, the dirty clothes thrown into the balcony has a camera to see the basket inside the clothes accumulated to To a certain extent, automatically contact the door of the laundry service, the whole do not have to worry about human. The future of consumption and services will be more active through the AI recommended model, which brings the market business opportunities will be far greater than the traditional model. Because the visual comparison of voice and touch is a higher level of communication, it can collect information much larger than touch and voice such a two-dimensional way to bring the information.

  Smart home market challenges

  After talking about the potential changes, the smart home market, although speaking for many years, but there is no real "outbreak" up, why? We think there are four main reasons.  The first is the use of scenes scattered, safe, entertainment, family services, health, education and so many scenes, while the home of intelligent equipment is too scattered, such as home appliances, toys, robots, cameras, etc., all of these scattered applications Scene and electronic equipment will lead to when we put them intelligent, it is difficult to quickly copy.  The second challenge is technology. Family scenes are complex, need to be able to handle voice and images, so that good perception, awareness, technical challenges until today is still great.  Third, data. Data collection and analysis is a difficult process, first of all, every scene inside the home is highly decentralized, the needs of each family is not the same. Second, only to get big data is not enough, because even if the home to install a lot of sensors, to collect a lot of data, but most of these data are junk data, if the data to upload to the cloud , Will take up huge bandwidth and cloud processing resources, do not have the operability. So when we collect the data, we have to think, not only to collect data, but also to collect the real valuable data. But this is not so easy to do.  The fourth challenge, the real user experience is hard. We serve the object is the people, for the experience of the high demand. We are now artificial intelligence technology has been a very rapid development, but still can not really make people feel the way of interaction and service is very natural.With these challenges, artificial intelligence in the popularity of intelligent life, we think it is not overnight things, it will take some time to accumulate slowly.How to deal with these challenges of smart home

  To deal with these challenges, the horizon has its own point of view.

  First, the application of the scene scattered problem, we think to seize the main contradiction, and the invariant, that is, "people", we are now smart home this area should be done through people. Such as security issues, now home to install a lot of camera head main purpose is to safety, and security in fact what is the core? The most important thing is to know whether there is no family, family or stranger, and this person's behavior , He is doing the behavior is safe or unsafe, such as children into the kitchen, will not point gas stove, these are unsafe behavior. All application scenarios are around people. Such as entertainment, family services, health, education, etc., not for example, we do horizon is from the face, to the human behavior analysis, to the surrounding environment, and the person's voice and equipment interaction, we The main point is that around the people, the technology to do a good job, support the huge scattered smart home needs.  The second is a technical challenge, the way people interact and the way they get information is actually a fusion of the way, separate voice and images, or any other sensor alone, can not cover all the application scenarios. So it is necessary to image + voice. In addition the vision is high-dimensional high value, 80% of the value of information from the image, only 20% from the voice, technically to complete the best experience of smart home layout, image and voice technology must be integrated, Do voice or image is absent.  The third data challenges. To get valuable data, how to get it? This is not easy to do, our ideas are relatively clear, is to do embedded artificial intelligence "end + cloud" approach. Side of it, and now we talk about the so-called artificial intelligence business companies Ye Hao, big companies Ye Hao, they are actually more in the cloud to do, but the end of the irreplaceable value. Mainly in several areas, the first, the end is not dependent on the network, whether at home, or where, we will find the home wireless network coverage is difficult to be perfect, if it is downstairs, or in the mobile In the scene, the full coverage of the network is a very difficult thing, especially when you involve some intelligent image processing, it is difficult to real-time image to the cloud and then come back, this time no real time. Second, the end of the data can be selected, only the valuable data to the cloud. For example, I have a balcony on the balcony of the laundry, I only focus on the clothes inside the time, this state I spread to the clouds. If this dress is not full, 99% of the time, these data are meaningless, and abandoned. Must adopt a side + cloud way, so that it is possible to really do the most efficient, and is the most cost-effective smart service.  With the end of the smart, the cloud can do a lot of things, one can use the data iteration model, continuous optimization of the user experience; the other is every hardware, the future and the form of the same phone, he is above You can download a lot of smart APP, the hardware may have been out of the original form, you buy back when it may be a home appliance, it may be a robot, but with the intelligent upgrade, he can have a lot of new properties and functions TheWhen we really have this big data, but also have a better understanding of people's preferences, this is not confined to the home, and when we go to any place can be upgraded through the cloud of local smart devices, so that they become like home The same best understand my environment. The future of the smart home market must be the end + cloud architecture.  Fourth, the real user experience, it is very difficult thing. For example, voice, not only to support the far field and near field, to identify some with the accent of Mandarin, and even dialect. But also consider the practical application, if many people speak at the same time, many of the home voice equipment recognition rate will drop more powerful. So when we do the technology, we must consider the actual use of the scene. Image is the same, such as face recognition, if you are very close, really identify very well, but if you are far away, if it is moving state, if it is the case of backlight, how to do? , Only the user experience can really be so good, consumers are really willing to use this technology, the market can really up. Another one, our technology more efforts to make the camera to do the active service, to understand the user, only to understand the user in order to enable users to get a good experience.