As an important landing direction for artificial intelligence, autonomous driving was once a hot pursuit of technology companies. But after the tide receded, some companies became leaders and continued to spend money on research and development, while some players have quit.
However, the money invested in R & D is substantial. How to deal with the “legacy” of technology? 360 gave an answer. If autonomous driving can’t be done, you can also be a robot cleaning.
Technology is interlinked, but it is not a dimensionality reduction strike yet
In 2017, 360 released the first sweeping robot. At that time, the publicity was mainly based on “automatic driving technology to reduce dimensionality”. At the press conference at that time, 360 also demonstrated some intelligent driving technologies. Results.
However, this year, 360 has no longer demonstrated the progress of autonomous driving. Although it is still the technical team that is doing sweeping robots, during the interview process, the 360 team admitted that the autonomous driving business has indeed given up. No longer continue to do, the original R & D personnel began to gradually enter and form a new department.
The core technologies of sweeping robots are mainly divided into sweeping and dragging, obstacle avoidance, modeling, human-computer interaction, etc. The autonomous driving technology heritage can help, mainly obstacle avoidance and construction.
Early-sweeping robots mainly use the random collision mode for cleaning, which is not only inefficient, but also may damage the furniture walls. With the development of technology, SLAM (Simultaneous Localization and Mapping) positioning and map construction technology has given sweeping robots the ability to recognize the environment.
According to the type of sensor, SLAM technology is divided into two categories: SLAM technology based on LDS laser ranging sensor, and SLAM based on machine vision.
The 360 team chose the laser ranging sensor, which is similar to the principle of the self-driving LiDAR sensor. It measures the time of flight of the laser beam in the air (transmitted from the transmission window, irradiated to the target, and then reflected back to the reception window. Time), multiplied by the speed of light, to get the distance from the rangefinder to the target, which is why the 360 team connected the automatic driving and the sweeping robot.
Applied to sweeping robots, it is to build a map of the house through lidar scanning, so as to calculate the layout of the whole house, the location of furniture, etc., so as to plan the route in advance and clean more effectively. The robot can also take the initiative (or Manually identify the forbidden area for cleaning.
This is a relatively advanced technology in the sweeping robot industry. Of course, it means higher prices. Similar products are priced at more than 2,000 yuan. The latest product of the 360 team, X90, is also in this range.
Domestic sweeping robots are a fiercely competitive field. It is not impossible to achieve breakthroughs with “autonomous driving technology”. The top players in China are Cobos, iRobot, Stone and Xiaomi, especially traditional robot manufacturers Cobos has an offline share of over 70%.
For its own advantages, the 360 robot team considers it as “technology and product capabilities.” 360 said that the current standard features of high-end products, such as setting a virtual wall (that is, automatically avoiding during the cleaning process), mobile app control ” The functions such as “Which to Scan” are all introduced earlier by 360, and the noise control can also be achieved by more than 3 decibels lower than the competition.
But the 360 team also admits that it is probably the third echelon of the domestic sweeping robot industry, with a share of about five or six, and it has a certain gap with the head companies in products, channels and brands.
In terms of technology, if the 360 team stated that it would not include labor costs, the R & D investment last year was at the tens of millions level. In the interview, the 360 team believes that the current crux of the industry is mainly not smart enough. I believe everyone has seen the paragraph of “sweeping robots spread pet poops.” At present, this technical difficulty has not been perfectly resolved, and radar ranging is mainly Responsible for positioning and ranging, the ability to recognize such small obstacles is limited.
Last year ’s CES Asia, 360 demonstrated a technology called “human-like eyes”, claiming to be able to avoid pet feces perfectly, but the 360 sweeping robot team said that the technology is still in the testing stage at the earliest. The product.
The data shows that the scale of global sweeping robots is US $ 1.7 billion, the penetration rate of the US market is 10%, and China is about 4%. There is theoretically room for growth, but the growth rate in 2019 has slowed significantly. AVC data shows that in 2018, China’s sweeping robot market sales were 6.13 million units, an increase of 45.2% year-on-year, while from January to October 2019, the sweeping robot market sales were only 4.19 million units, an increase of 6% year-on-year .
In the eyes of many consumers, sweeping robots are products that “have better products and don’t have them.” Like dishwashers, smart toilet lids, and dryers, the entire industry is faced with how to continue to grow.