At present, on the technology side, the intelligence of service robots has reached the commercial level, and API and modularization accelerate product development iterations. On the application side, the service robot has built a "six-dimensional model." In the future, these two aspects are still the key to the layout of service robots. 
In 2015, the global market for professional service robots was US$4.6 billion, and the market for personal/home service robots was US$2.2 billion. It is expected to see sustained and rapid growth in 2016-2019. IFR divides service robots into professional service robots and personal/home service robots. According to IFR statistics, the CAGR of the global professional service robot market was 5.1% in 2010-2015, and the CAGR of personal/home service robot market was 33.6% in the same period. IFR predicts that in 2016-2019, the global professional service robot market will accumulatively sell 23.1 billion U.S. dollars, and the personal/home service robot market will be 24.2 billion U.S. dollars. Based on this, CAGR for professional service robot market was 9.4% in 2016-2019, and CAGR for personal/home service robot market was 45.0%. Technology side: Robot intelligence reaches commercial level, API and modular acceleration accelerate product development 1 Deep learning greatly advances the intelligence of service robot Deep learning is an important milestone in the development of artificial intelligence, and it promotes the rapid development of voice interaction and image recognition. From the Dartmouth Conference in 1956, the development of artificial intelligence has gone through three stages. (1) The first stage: the reasoning period (1956-1960s). The subject of this stage is to assign logical reasoning capabilities to computer systems, the main achievement being the development of an automatic theorem proving system, etc.; (2) the second stage: the knowledge period (1970s-1980s). At this stage, the researchers are committed to summarizing human knowledge and handing it over to the computer system. The main achievement is the development of an expert system; (3) The third stage: the learning period (1990s to date). Researchers are trying to make computers learn their own knowledge. An important breakthrough at this stage is the introduction and development of deep learning. Hinton proposed a deep learning algorithm in 2006. In 2012, Hinton's team took advantage of the deep learning algorithm to shine on ImageNet. Subsequent deep learning gained widespread attention in the industry, while attracting huge amounts of investment from IT giants, which in turn drove the rapid development of speech recognition, image recognition, and natural language processing. Voice interaction is becoming one of the mainstream technologies for man-machine interaction of service robots. The way of human-machine voice interaction is more similar to the mode of communication between people. At present, the service robot products mostly adopt the voice interaction mode. Domestic service robots mainly use voice interaction modules provided by enterprises such as HKUST, Baidu, Sibi Chi, and Cloud Sound. Apple, Google, Microsoft, Amazon, Facebook, IBM and other IT giants have already deployed their own speech recognition technology. Thanks to deep learning technology, speech recognition has made great strides and is now approaching or even exceeding humans in some areas. Since 1970, speech recognition technology has developed rapidly, and the number of machine recognition words has grown exponentially. At present, the word recognition accuracy of all mainstream speech recognition platforms has exceeded 90%. In October 2016, Microsoft speech recognition technology achieved an error rate of 5.9% in the industry-standard Switchboard speech recognition task test, which has reached the level of a professional stenographer. The image recognition ability has reached the human level. Image recognition capability is also an important part of the field of human-machine interaction for service robots. In 2012, Hinton's team used GPUs to train neural network models, dramatically reducing the error rate from 26% to 15%. Since then, the error rate has been declining year by year. By 2015, the error rate of the Microsoft Asia Research Institute computer group has been as low as 3.6%, and the human eye recognition error rate was 5.1%. 2 Development and iteration speeds of APIs and modular upgrade service robotics APIs and modular upgrade service robots development and iteration speeds. The intelligent modules of service robots include sensory interaction, cognitive decision-making, positioning and navigation, and motion control. The technical span is extremely wide. Service robot start-ups can generally study in depth one module at a time, and even many startups only do platform integration. In this context, the opening of the API interface of artificial intelligence technology and the modularization of core components will undoubtedly greatly reduce the difficulty of product development and iteration for service robot manufacturers. At present, a number of outstanding companies have emerged in the core technology of service robots, and the service robot integration industry is being promoted to mature, enabling service robot integrators to devote their energy to the exploration of industrial applications and business models. The artificial intelligence API enables service robot manufacturers to use external forces to achieve product intelligence. There are a large number of open artificial intelligence APIs (Application Programming Interfaces) that cover (1) machine learning and prediction, (2) image recognition, (3) text analysis, NLP, and sentiment analysis, (4) Machine translation, (5) speech recognition. The Internet giant also began to join the camp of open artificial intelligence API. In 2015-2016, Google launched Google Cloud Vision API and Google Cloud Speech API, which can be applied to software or devices. Devices connected to these two APIs can upload pictures and voices to the Google Cloud. After processing in the cloud, they will return images and voice content that need to be recognized. Microsoft's open API includes expression recognition, speech recognition, video recognition, and more. The opening of artificial intelligence API interfaces has greatly reduced the difficulty for service robot manufacturers to realize product intelligence. The modularization of the core components accelerates the production R&D process of service robot manufacturers. The modularization of the core components facilitates the service robot manufacturers to acquire the functions required by the robot products through outsourcing. Taking the navigation mobile module as an example, Sci-Tech launched the SLAMWARE, a modular robot autonomous positioning and navigation system, which is only coin-sized and integrates laser radar-based synchronous positioning (SLAM) and matching path planning capabilities. The host system can use the communication interface provided by SLAMWARE, combined with the high-precision robot position information acquired by SDK in real time and the environmental map data constructed by SLAMWARE. Application side: Two types of target users, three types of products, building a "six-dimensional model" 1 "six-dimensional model of service robots" Building a "six-dimensional model" based on the role of target users and products. Various types of service robots are extremely different from each other. For example, Da Vinci surgical robots, Cobos sweeping robots, and UAV drones are service robots, but the logic behind them is obviously different. We categorized the application of the service robot from two perspectives: the target user type and the product role type, and proposed the "six-dimensional model" of the service robot for the first time. The advantages of this classification are: (1) Avoiding general discussion of all service robots under one framework. For example, we will point out that rising labor costs can only promote the development of “replacement human†service robot products; and (2) avoid being trapped in the list of poor companies listed by the company. IFR divides service robots into 2 categories, 13 sub-categories, and hundreds of specific products. The analysis of the main lines cannot be done in an exhaustive manner. Classification point one: According to the target customer type, it is divided into to B and to C two types of products. The same type of service robot, the main to B market or to C market, determines its different development model. For example, to C's aerial drones and to B's vicious drones have significant differences in price positioning, sales channels, and business models. (1) To B's service robot: Mainly refers to professional-grade service robots. The customers faced are corporations, governments, and institutions. Types include military, agriculture, animal husbandry, logistics, transportation, medical, construction, cleaning, testing, underwater, and rescue. (2) To C's service robots: Mainly referring to home and personal service robots. The customers faced are individuals or families. The types include household chores, entertainment, education, nursing care, and security. At present, we also observe the trend of mutual integration between products of To B and companies of C. For example, after Dadong occupied the aerial drone market of To C, it also launched the agricultural plant protection machine to B. The second perspective of classification: According to the type of product role, it is divided into alternative humans, auxiliary humans, new fields of creation, and three types of products. (1) Replacement of humans: In order to reduce labor costs or improve work safety, service robots have been used in many fields to replace human work, such as hotel delivery robots and explosion-proof robots; (2) Assisting humans: in some areas, services Robots are not intended to replace humans, but to improve the quality of work. Typical examples are Da Vinci Surgical Robots and IBM Waston. (3) Creation of new areas: The development of service robots has also created many new areas of trials. Most of these areas are currently At the stage of market exploration and verification, such as intelligent virtual assistants, emotional robots, and programmed toy robots. 2 Two types of target customers: to B products and to C products (1) to C products To C's products focus on price, ease of use, and product ecology. To C's products are directly facing the end-consumer. In the early stage, a large amount of market cultivation and promotion expenses were required. Therefore, compared to the to B products, the cycle of income and profit is generally longer. The advantage of to C's products is its ability to gradually develop its own brand and product ecosystem. We believe that the to C service robotics company that meets the following three criteria deserves our attention: (1) Low-priced but cost-effective products; (2) Products that are easy to use and can be quickly learned and used by consumer-level users (3) Establish product ecosystem around core products. Concern 1: Cheap and cost-effective products. Opening the to C market requires two things: (1) product intelligence can satisfy the customer's rational expectations, which is mainly achieved through the integration of more and higher quality intelligent modules; (2) the product price is low enough to satisfy the C-side Customer's affordability. Based on these two points, manufacturers of to C need to strike a balance between technology pursuit and cost control. Currently, the to C service robot companies that have emerged in China, such as Dajiang, Coworth, and Youbixi, all open up to the C market through low-cost and cost-effective products, and achieve large-scale sales. Pepper robots are used in Nestlé stores. Taking overseas markets with high consumer price tolerance as a breakthrough is a way for C companies to “curve the countryâ€. The domestic market's C-side customer's price bearing capacity is weaker than the foreign market, and the same cost of service robotic products is easier to open the market in foreign countries. In addition to the high price bearing capacity, overseas C-side customers have a good history of DIY and robot use, while domestic robot consumer groups are still under cultivation. Companies such as Dajiang and Youdian must first open overseas markets and establish industry status, and then turn to open up the domestic market. In 2015, Alpha 1S domestic sales accounted for approximately 15%, and DJI domestic sales accounted for approximately 20%. Concern 2: Easily accessible products that can be quickly learned and used by consumer-level users. To C's service robot integrates more complex intelligent modules than ordinary consumer electronics products, and it also increases its difficulty of use. The C-side customer itself does not have professional capabilities. Ease of use is a prerequisite for opening the to C market. Typical examples of successful operations include: (1) You must choose to develop a 3D visual programming system. Even child users can use the graphical interface to arrange the Alpha1S. Its ease of use is significantly ahead of competing products that use programming languages ​​for action planning; (2) DJI allows consumers to take high-quality picture content by increasing the stability of PTZ and other technologies. Focus 3: Build an ecosystem around core products. To C's service robot products have the characteristics of low industry threshold and short product iteration cycle. A star product can easily attract a large number of imitators. It is difficult to maintain a leading edge in the industry only by relying on a star product. Enterprises need to build their own ecosystem to maintain their leading position. (1) Dajiang UAV: ​​In terms of products, Dajiang established and improved its drone product line in response to different consumer needs and industrial-level needs. In terms of users, Dajiang established a friendly ecosystem, through sharing and competitions. , The community accumulates data to accurately characterize user characteristics; (2) Ubiquitous: Ubiquitous must choose multiple product lines including Alpha 1S, Alpha 2, and Disney Cooperative IP toy robots and building block robots to form home applications. The data is closed-loop. (2) to B Products to B products focus on product efficacy, stability/security/durability, and application scenario selection. To B's service robot products, the customer groups faced are enterprises, governments, institutions, etc. Compared with C-side customers, the characteristics of B-side customers are more rational, and they are more from the cost and utility of purchasing service robot products. Make a decision. We believe that the to B service robotics company that meets the following three criteria deserves our attention: (1) products can effectively solve customers' pain points; (2) products have high stability, safety, and durability; (3) products The application scenario has a deep understanding. Concern 1: The product can effectively solve customers' pain points. The B-side customer's purchase decision is more rational, and it pays more attention to whether the service robot product can effectively solve its business pain points, and will not easily pay for unpractical functions. The NAO and Pepper of Aldebaran Robotics, a French subsidiary of Softbank, are excellent humanoid robots abroad. Due to the higher unit price, the two products are currently used in the To B market. (1) NAO robots are currently mainly used in schools, Research institutions, government related project departments, etc. For example, the NAO robot was selected by the European L2TOR project to use the voice interaction function of the NAO robot to help European immigrant children learn the language of the new country; (2) The Pepper robot is used in the commercial field, for example, Nestlé launches the Pepper robot at the appliance store in Japan. To sell its coffee machine. However, at present, there is still a gap between the development of Pepper and the original expectation. Apart from attracting customers through novelty, how to effectively help merchants to serve customers is still a problem that they need to solve. B-end customers are more tolerant of prices, and the unit price is higher than to C products. According to the IFR report, in the service robot market in 2014, sales of to B products accounted for only 0.5%, but their sales accounted for 63.1% of sales. The B-side customers are more rational than the C-side customers, and they are more concerned with the cost-benefit comparison of purchasing service robot products. Especially in the field of medical service robots, Da Vinci surgical robots have a price of tens of millions of dollars. Concern 2: The product has high stability, safety, and durability. (1) Stability: B-side customers have low tolerance to service robot failures. For example, a temporary failure of a hospital delivery robot may cause a medical accident; (2) Security: It is used in crowded areas such as shopping malls, hotels, and conference venues. Service robot products emphasize safety more; (3) Durability: Products entering the B-side market have strong capital attributes, and their service life, depreciation speed, and maintenance cost are all important indicators for B-side customers. At present, most of the to B service robots are in the initial stage of market development. It is necessary to perfect the product's stability, safety, and durability by continuously conducting field trials and product update iterations. Taking service robots used in shopping malls at home as an example, problems that need to be resolved include human-machine dialogue capabilities in high-noise environments, reaction capabilities in human interference, and emergency capabilities such as weak WIFI signals. Focus 3: Have a deep understanding of the product application scenario. Service robot = intelligent module integration + specific functional components. Under the same intelligent module integration, what kind of application scenario is selected and matched with the corresponding functional components is the key to the development of the market for toroidal robot products. Also with indoor navigation as the core intelligent module, Aethon and Savioke chose different application scenarios. (1) Aethon's TUG robots mainly work in hospitals and provide services such as food delivery, medical waste delivery, and communication between medical staff and patients in streaming media; (2) Savioke's Relay robots work mainly in hotels and can be used as replacements. Hotel staff deliver hair dryers, towels, takeaways and other items for the rooms. Domestic cloud-robot robots are also service robots designed for hotel delivery. The selection of application scenarios and the complex approval process of domestic medical equipment are related to the reasons. 3 The role of three types of products: replacing humans, supporting humans, and opening up new areas (1) Replacement of humans Service robots that replace humans mainly focus on whether they can reduce labor costs or improve work safety, as well as the extension of application scenarios. Due to the rising labor costs and the aging of the population, a large number of companies are currently trying to replace human work by serving robots. Typical examples are storage robots. In addition, service robots have also achieved human replacement in parts of high-risk work. Because such robots are replacing human work in the original field, their prices need to be based on human costs. Their pricing cannot generally be higher than the artificial compensation costs that they can replace. At present, the prices of mainstream welcome, shopping guide, and consulting robots in the market are all between 50,000 and 200,000 yuan, which is not as significant as the labor cost. However, if such service robots can achieve mass production and sales, prices are expected to decline, and the competitiveness of their replacement laborers will also increase. There are indeed service robot products that are much higher than labor costs, such as welcome robots worth more than 300,000 yuan. We believe that merchants purchase such products not only to replace manual labor, but also to use robots to attract customer attention. These service robot companies that meet the following two criteria deserve our attention: (1) Effectively reducing labor costs or effectively improving work safety; (2) Products have the ability to extend application scenarios. Concern 1: Effectively reduce labor costs. Service robots are widely used in areas where there is a need to reduce labor costs. (1) A typical example is Amazon's KIVA warehouse robot. In 2012, Amazon acquired KIVA for US$750 million and obtained its logistics automation technology, which can help Amazon save US$900 million in labor costs annually. According to MWPVL, a logistics consulting company, after using KIVA, Amazon will save 21.3 cents per item, which is 48% of the cost; and (2) A typical example of replacing human work in the to C area is Coworth. Irobot's sweeping robot. Effectively improve work safety. Examples include (1) large-scale deployment of service robots in the military field to replace humans for high-risk work. For example, a military mine-clearing robot equipped with a mine detector and a mine failure device can speed up the speed of mine clearance and greatly reduce the number of casualties; (2) Pac Technology's glass exterior cleaning robot ispider can be independent of the building's external walls. Cleansing is replacing the high-risk external wall cleaning of Spiderman. Concern 2: The product has the ability to extend the application scenario. After successful sweeping robots, a repetitive housework replacement market was verified. On this basis, the service robot manufacturers further tap into the relevant market potential, mainly from two directions, (1) near the market expansion: the same housework labor substitution market, manufacturers developed automatic cleaning glass robots and automatic weeding robots, examples include Cobos WINBOT Glass Cleaning Robot and JOHN DEERE's Automated Weeding Robot, (2) Expansion of target customers: Cobos sweeping robot was originally aimed at the to C market, and in 2016 the company launched to B for solar photovoltaic power station. Solar panel cleaning robot Rui Bao RAYBOT. (2) Assisting Humans Service robots assisting humans pay attention to technological advantages and the ability of technology to land in application scenarios. Service robots and humans can be cooperative relationships, and can improve human existing work efficiency by effectively utilizing technologies such as artificial intelligence, motion control, and human-computer interaction. The main driving force for the development of "assisting human" service robots is related technological progress. In this field, enterprises with technological advantages are generally occupying the leading position in the industry. Typical examples are Da Vinci surgical robots and IBM Waston. The prices of such service robots are mainly based on their value for improving human work. Take the Da Vinci surgical robot as an example. Using its assisted surgery, it is generally necessary to add 30,000 yuan to the cost of traditional surgery. The incremental value of the 30,000 yuan reflects the improvement of Da Vinci surgical robots' surgery. At present, Da Vinci surgical robots are sold at a price of 100-1.5 million US dollars in foreign countries, and are sold in the country for 3 million US dollars. We believe that such service robot companies that meet the following two criteria deserve our attention: (1) they have a leading edge in technology; and (2) they have the ability to use technology to land in application scenarios. Concern 1: The product has technical advantages. A typical example is a surgical robot whose requirements for mechanical motion control are extremely high. The Da Vinci surgical robot consists of a surgeon's console, a bedside robotic arm system, and an imaging system. With the Da Vinci robot, doctors can obtain 10 times more magnification of the three-dimensional vision, a more comfortable surgical environment, and higher-precision surgical procedures to improve the quality of surgery. Other star medical robots include Auris Surgical, an ophthalmic surgical robot, ARTAS, a hair-transplant robot, and Medrobotics, a flexible medical robot. Concern 2: Have the ability to use technology in the application scenario. A typical example is IBM Waston. The purpose of IBM's development of artificial intelligence technology is not to use machines instead of humans, but to extract core technologies from artificial intelligence, such as deep learning and pattern recognition, and then use these core technologies to help humans overcome their own deficiencies. Currently, IBM Waston is widely used in the medical field. In addition, IBM Waston also tried in many areas, including: (1) Analysis of team data: The NBA Raptors teamed up with IBM to use data analysis to select players, formulate game countermeasures, and temporarily adjust match plans. Changed the situation where previous coaches took brains to make decisions; (2) Smart wear to monitor health status: Waston provided users with exercise suggestions by analyzing user data; (3) Cooperated with robot body manufacturers: IBM implanted Watson in robot products. , To enable human-computer interaction; (4) To help the government manage haze: Waston assists the Beijing Municipal Government in combating haze by analyzing data such as air flow, sewage, social network commentary, and photographs; (5) assisting the government in implementing grading clinics : Waston recognizes public information, hospital information, patient information, etc., provides decision-makers with timely reports, and evaluates the effectiveness and problems of implementing policies. (3) Creating new fields Service robots used in the field of innovation need to pay attention to whether there are real needs. With the development of service robotics technology, manufacturers are not satisfied with replacing or assisting human work only in the original field, and they are also starting to tap new demand areas. Currently, they are mostly in the stage of market exploration and verification, such as intelligent virtual assistants, emotional robots, and programming. Toy robots and so on. Such products generally enter a new market area and need to follow and verify whether the new demand areas created by them are real demand. The success of Amazon Echo smart speakers has attracted a large number of imitators. Echo is a smart speaker introduced by Amazon in 2014. Amazon pre-installed the personal virtual assistant Alexa (Siri-like) in Echo. Users can interact with Echo to implement news broadcasts, weather inquiries, task reminders, music playback, and smart home functions. Echo has the potential to become a smart home of the future. The control platform. Echo is priced at $180 and cumulative sales have already exceeded 3 million units. Echo's success factors include: (1) The market demand is real: Echo knocks on the market through the form of home speakers, which is a huge and just-needed market; (2) The technical solution effectively solves the user's pain points: The original speakers need to be booted and selected. Music, tuning, and other steps, and Echo can immediately start working according to the user's voice instructions after hearing the wakeup word “Alexaâ€; (3) Amazon uses Echo as a strategic product to invest in R&D. Echo is more than just a speaker. It is also the home entrance of the Internet of Things. Amazon has established a R&D team with more than 1,000 people and has established its advantages in technology to its competitors. For example, Echo's average response speed has been within 1.5 seconds. . Echo's success has enabled competitors to see the potential of the smart speaker market. In 2015, HKUST and JD.com established a joint venture to establish Linglong Technology and launched the audio system. In 2016, Google launched Google Home. Compared to Echo, Google Home features include (1) deep integration with various services provided by Google; (2) support for simultaneous use of multiple rooms and multiple devices; (3) use only 2 microphone combinations, and Echo is 6+1 Microphone array; (4) More customized appearance. MIT launched emotional robot JIBO. JIBO is a family social robot developed by a team at the MIT Media Lab. According to the development team, JIBO has emotion-aware functions that can recognize and interact with user emotions. JIBO will focus on the mid- to high-end family market. At the same time, some B-side real estate, hotels, etc. are also potential users. In August 2015, JIBO introduced investors from China, Japan, and South Korea to pave the way for entering the Asian market. Among them, Eastern Power Company invested 2 million US dollars accounted for 2.34% of the equity. Orient Online will also help JIBO improve its video calling and localization. In addition, GQY Video also invested RMB 9 million in 1.1% equity in November 2015. However, at present, JIBO's product release has been repeatedly postponed, and its market potential remains to be verified. Microsoft Xiao Hai simulates human singing and intends to enter the virtual singer market. In August 2016, Microsoft held a conference for the fourth generation of Xiao Bing. The Microsoft team took advantage of big data training. The new Xiao Bing has been able to imitate human singing. Her "Invisible Wings" and "Summer Sweetheart" theme songs have received extensive attention. At present, the most famous virtual singer is Japan's Hatsune Miku. His singing is mainly generated using electronic sound synthesis technology, and he conducts concerts through 3D modeling and holographic projection technology. Microsoft Xiao Bing learns and sings through artificial intelligence technology and is expected to step out a new growth path for virtual singers.
In 2015, the global market for professional service robots was US$4.6 billion, and the market for personal/home service robots was US$2.2 billion. It is expected to see sustained and rapid growth in 2016-2019. IFR divides service robots into professional service robots and personal/home service robots. According to IFR statistics, the CAGR of the global professional service robot market was 5.1% in 2010-2015, and the CAGR of personal/home service robot market was 33.6% in the same period. IFR predicts that in 2016-2019, the global professional service robot market will accumulatively sell 23.1 billion U.S. dollars, and the personal/home service robot market will be 24.2 billion U.S. dollars. Based on this, CAGR for professional service robot market was 9.4% in 2016-2019, and CAGR for personal/home service robot market was 45.0%. Technology side: Robot intelligence reaches commercial level, API and modular acceleration accelerate product development 1 Deep learning greatly advances the intelligence of service robot Deep learning is an important milestone in the development of artificial intelligence, and it promotes the rapid development of voice interaction and image recognition. From the Dartmouth Conference in 1956, the development of artificial intelligence has gone through three stages. (1) The first stage: the reasoning period (1956-1960s). The subject of this stage is to assign logical reasoning capabilities to computer systems, the main achievement being the development of an automatic theorem proving system, etc.; (2) the second stage: the knowledge period (1970s-1980s). At this stage, the researchers are committed to summarizing human knowledge and handing it over to the computer system. The main achievement is the development of an expert system; (3) The third stage: the learning period (1990s to date). Researchers are trying to make computers learn their own knowledge. An important breakthrough at this stage is the introduction and development of deep learning. Hinton proposed a deep learning algorithm in 2006. In 2012, Hinton's team took advantage of the deep learning algorithm to shine on ImageNet. Subsequent deep learning gained widespread attention in the industry, while attracting huge amounts of investment from IT giants, which in turn drove the rapid development of speech recognition, image recognition, and natural language processing. Voice interaction is becoming one of the mainstream technologies for man-machine interaction of service robots. The way of human-machine voice interaction is more similar to the mode of communication between people. At present, the service robot products mostly adopt the voice interaction mode. Domestic service robots mainly use voice interaction modules provided by enterprises such as HKUST, Baidu, Sibi Chi, and Cloud Sound. Apple, Google, Microsoft, Amazon, Facebook, IBM and other IT giants have already deployed their own speech recognition technology. Thanks to deep learning technology, speech recognition has made great strides and is now approaching or even exceeding humans in some areas. Since 1970, speech recognition technology has developed rapidly, and the number of machine recognition words has grown exponentially. At present, the word recognition accuracy of all mainstream speech recognition platforms has exceeded 90%. In October 2016, Microsoft speech recognition technology achieved an error rate of 5.9% in the industry-standard Switchboard speech recognition task test, which has reached the level of a professional stenographer. The image recognition ability has reached the human level. Image recognition capability is also an important part of the field of human-machine interaction for service robots. In 2012, Hinton's team used GPUs to train neural network models, dramatically reducing the error rate from 26% to 15%. Since then, the error rate has been declining year by year. By 2015, the error rate of the Microsoft Asia Research Institute computer group has been as low as 3.6%, and the human eye recognition error rate was 5.1%. 2 Development and iteration speeds of APIs and modular upgrade service robotics APIs and modular upgrade service robots development and iteration speeds. The intelligent modules of service robots include sensory interaction, cognitive decision-making, positioning and navigation, and motion control. The technical span is extremely wide. Service robot start-ups can generally study in depth one module at a time, and even many startups only do platform integration. In this context, the opening of the API interface of artificial intelligence technology and the modularization of core components will undoubtedly greatly reduce the difficulty of product development and iteration for service robot manufacturers. At present, a number of outstanding companies have emerged in the core technology of service robots, and the service robot integration industry is being promoted to mature, enabling service robot integrators to devote their energy to the exploration of industrial applications and business models. The artificial intelligence API enables service robot manufacturers to use external forces to achieve product intelligence. There are a large number of open artificial intelligence APIs (Application Programming Interfaces) that cover (1) machine learning and prediction, (2) image recognition, (3) text analysis, NLP, and sentiment analysis, (4) Machine translation, (5) speech recognition. The Internet giant also began to join the camp of open artificial intelligence API. In 2015-2016, Google launched Google Cloud Vision API and Google Cloud Speech API, which can be applied to software or devices. Devices connected to these two APIs can upload pictures and voices to the Google Cloud. After processing in the cloud, they will return images and voice content that need to be recognized. Microsoft's open API includes expression recognition, speech recognition, video recognition, and more. The opening of artificial intelligence API interfaces has greatly reduced the difficulty for service robot manufacturers to realize product intelligence. The modularization of the core components accelerates the production R&D process of service robot manufacturers. The modularization of the core components facilitates the service robot manufacturers to acquire the functions required by the robot products through outsourcing. Taking the navigation mobile module as an example, Sci-Tech launched the SLAMWARE, a modular robot autonomous positioning and navigation system, which is only coin-sized and integrates laser radar-based synchronous positioning (SLAM) and matching path planning capabilities. The host system can use the communication interface provided by SLAMWARE, combined with the high-precision robot position information acquired by SDK in real time and the environmental map data constructed by SLAMWARE. Application side: Two types of target users, three types of products, building a "six-dimensional model" 1 "six-dimensional model of service robots" Building a "six-dimensional model" based on the role of target users and products. Various types of service robots are extremely different from each other. For example, Da Vinci surgical robots, Cobos sweeping robots, and UAV drones are service robots, but the logic behind them is obviously different. We categorized the application of the service robot from two perspectives: the target user type and the product role type, and proposed the "six-dimensional model" of the service robot for the first time. The advantages of this classification are: (1) Avoiding general discussion of all service robots under one framework. For example, we will point out that rising labor costs can only promote the development of “replacement human†service robot products; and (2) avoid being trapped in the list of poor companies listed by the company. IFR divides service robots into 2 categories, 13 sub-categories, and hundreds of specific products. The analysis of the main lines cannot be done in an exhaustive manner. Classification point one: According to the target customer type, it is divided into to B and to C two types of products. The same type of service robot, the main to B market or to C market, determines its different development model. For example, to C's aerial drones and to B's vicious drones have significant differences in price positioning, sales channels, and business models. (1) To B's service robot: Mainly refers to professional-grade service robots. The customers faced are corporations, governments, and institutions. Types include military, agriculture, animal husbandry, logistics, transportation, medical, construction, cleaning, testing, underwater, and rescue. (2) To C's service robots: Mainly referring to home and personal service robots. The customers faced are individuals or families. The types include household chores, entertainment, education, nursing care, and security. At present, we also observe the trend of mutual integration between products of To B and companies of C. For example, after Dadong occupied the aerial drone market of To C, it also launched the agricultural plant protection machine to B. The second perspective of classification: According to the type of product role, it is divided into alternative humans, auxiliary humans, new fields of creation, and three types of products. (1) Replacement of humans: In order to reduce labor costs or improve work safety, service robots have been used in many fields to replace human work, such as hotel delivery robots and explosion-proof robots; (2) Assisting humans: in some areas, services Robots are not intended to replace humans, but to improve the quality of work. Typical examples are Da Vinci Surgical Robots and IBM Waston. (3) Creation of new areas: The development of service robots has also created many new areas of trials. Most of these areas are currently At the stage of market exploration and verification, such as intelligent virtual assistants, emotional robots, and programmed toy robots. 2 Two types of target customers: to B products and to C products (1) to C products To C's products focus on price, ease of use, and product ecology. To C's products are directly facing the end-consumer. In the early stage, a large amount of market cultivation and promotion expenses were required. Therefore, compared to the to B products, the cycle of income and profit is generally longer. The advantage of to C's products is its ability to gradually develop its own brand and product ecosystem. We believe that the to C service robotics company that meets the following three criteria deserves our attention: (1) Low-priced but cost-effective products; (2) Products that are easy to use and can be quickly learned and used by consumer-level users (3) Establish product ecosystem around core products. Concern 1: Cheap and cost-effective products. Opening the to C market requires two things: (1) product intelligence can satisfy the customer's rational expectations, which is mainly achieved through the integration of more and higher quality intelligent modules; (2) the product price is low enough to satisfy the C-side Customer's affordability. Based on these two points, manufacturers of to C need to strike a balance between technology pursuit and cost control. Currently, the to C service robot companies that have emerged in China, such as Dajiang, Coworth, and Youbixi, all open up to the C market through low-cost and cost-effective products, and achieve large-scale sales. Pepper robots are used in Nestlé stores. Taking overseas markets with high consumer price tolerance as a breakthrough is a way for C companies to “curve the countryâ€. The domestic market's C-side customer's price bearing capacity is weaker than the foreign market, and the same cost of service robotic products is easier to open the market in foreign countries. In addition to the high price bearing capacity, overseas C-side customers have a good history of DIY and robot use, while domestic robot consumer groups are still under cultivation. Companies such as Dajiang and Youdian must first open overseas markets and establish industry status, and then turn to open up the domestic market. In 2015, Alpha 1S domestic sales accounted for approximately 15%, and DJI domestic sales accounted for approximately 20%. Concern 2: Easily accessible products that can be quickly learned and used by consumer-level users. To C's service robot integrates more complex intelligent modules than ordinary consumer electronics products, and it also increases its difficulty of use. The C-side customer itself does not have professional capabilities. Ease of use is a prerequisite for opening the to C market. Typical examples of successful operations include: (1) You must choose to develop a 3D visual programming system. Even child users can use the graphical interface to arrange the Alpha1S. Its ease of use is significantly ahead of competing products that use programming languages ​​for action planning; (2) DJI allows consumers to take high-quality picture content by increasing the stability of PTZ and other technologies. Focus 3: Build an ecosystem around core products. To C's service robot products have the characteristics of low industry threshold and short product iteration cycle. A star product can easily attract a large number of imitators. It is difficult to maintain a leading edge in the industry only by relying on a star product. Enterprises need to build their own ecosystem to maintain their leading position. (1) Dajiang UAV: ​​In terms of products, Dajiang established and improved its drone product line in response to different consumer needs and industrial-level needs. In terms of users, Dajiang established a friendly ecosystem, through sharing and competitions. , The community accumulates data to accurately characterize user characteristics; (2) Ubiquitous: Ubiquitous must choose multiple product lines including Alpha 1S, Alpha 2, and Disney Cooperative IP toy robots and building block robots to form home applications. The data is closed-loop. (2) to B Products to B products focus on product efficacy, stability/security/durability, and application scenario selection. To B's service robot products, the customer groups faced are enterprises, governments, institutions, etc. Compared with C-side customers, the characteristics of B-side customers are more rational, and they are more from the cost and utility of purchasing service robot products. Make a decision. We believe that the to B service robotics company that meets the following three criteria deserves our attention: (1) products can effectively solve customers' pain points; (2) products have high stability, safety, and durability; (3) products The application scenario has a deep understanding. Concern 1: The product can effectively solve customers' pain points. The B-side customer's purchase decision is more rational, and it pays more attention to whether the service robot product can effectively solve its business pain points, and will not easily pay for unpractical functions. The NAO and Pepper of Aldebaran Robotics, a French subsidiary of Softbank, are excellent humanoid robots abroad. Due to the higher unit price, the two products are currently used in the To B market. (1) NAO robots are currently mainly used in schools, Research institutions, government related project departments, etc. For example, the NAO robot was selected by the European L2TOR project to use the voice interaction function of the NAO robot to help European immigrant children learn the language of the new country; (2) The Pepper robot is used in the commercial field, for example, Nestlé launches the Pepper robot at the appliance store in Japan. To sell its coffee machine. However, at present, there is still a gap between the development of Pepper and the original expectation. Apart from attracting customers through novelty, how to effectively help merchants to serve customers is still a problem that they need to solve. B-end customers are more tolerant of prices, and the unit price is higher than to C products. According to the IFR report, in the service robot market in 2014, sales of to B products accounted for only 0.5%, but their sales accounted for 63.1% of sales. The B-side customers are more rational than the C-side customers, and they are more concerned with the cost-benefit comparison of purchasing service robot products. Especially in the field of medical service robots, Da Vinci surgical robots have a price of tens of millions of dollars. Concern 2: The product has high stability, safety, and durability. (1) Stability: B-side customers have low tolerance to service robot failures. For example, a temporary failure of a hospital delivery robot may cause a medical accident; (2) Security: It is used in crowded areas such as shopping malls, hotels, and conference venues. Service robot products emphasize safety more; (3) Durability: Products entering the B-side market have strong capital attributes, and their service life, depreciation speed, and maintenance cost are all important indicators for B-side customers. At present, most of the to B service robots are in the initial stage of market development. It is necessary to perfect the product's stability, safety, and durability by continuously conducting field trials and product update iterations. Taking service robots used in shopping malls at home as an example, problems that need to be resolved include human-machine dialogue capabilities in high-noise environments, reaction capabilities in human interference, and emergency capabilities such as weak WIFI signals. Focus 3: Have a deep understanding of the product application scenario. Service robot = intelligent module integration + specific functional components. Under the same intelligent module integration, what kind of application scenario is selected and matched with the corresponding functional components is the key to the development of the market for toroidal robot products. Also with indoor navigation as the core intelligent module, Aethon and Savioke chose different application scenarios. (1) Aethon's TUG robots mainly work in hospitals and provide services such as food delivery, medical waste delivery, and communication between medical staff and patients in streaming media; (2) Savioke's Relay robots work mainly in hotels and can be used as replacements. Hotel staff deliver hair dryers, towels, takeaways and other items for the rooms. Domestic cloud-robot robots are also service robots designed for hotel delivery. The selection of application scenarios and the complex approval process of domestic medical equipment are related to the reasons. 3 The role of three types of products: replacing humans, supporting humans, and opening up new areas (1) Replacement of humans Service robots that replace humans mainly focus on whether they can reduce labor costs or improve work safety, as well as the extension of application scenarios. Due to the rising labor costs and the aging of the population, a large number of companies are currently trying to replace human work by serving robots. Typical examples are storage robots. In addition, service robots have also achieved human replacement in parts of high-risk work. Because such robots are replacing human work in the original field, their prices need to be based on human costs. Their pricing cannot generally be higher than the artificial compensation costs that they can replace. At present, the prices of mainstream welcome, shopping guide, and consulting robots in the market are all between 50,000 and 200,000 yuan, which is not as significant as the labor cost. However, if such service robots can achieve mass production and sales, prices are expected to decline, and the competitiveness of their replacement laborers will also increase. There are indeed service robot products that are much higher than labor costs, such as welcome robots worth more than 300,000 yuan. We believe that merchants purchase such products not only to replace manual labor, but also to use robots to attract customer attention. These service robot companies that meet the following two criteria deserve our attention: (1) Effectively reducing labor costs or effectively improving work safety; (2) Products have the ability to extend application scenarios. Concern 1: Effectively reduce labor costs. Service robots are widely used in areas where there is a need to reduce labor costs. (1) A typical example is Amazon's KIVA warehouse robot. In 2012, Amazon acquired KIVA for US$750 million and obtained its logistics automation technology, which can help Amazon save US$900 million in labor costs annually. According to MWPVL, a logistics consulting company, after using KIVA, Amazon will save 21.3 cents per item, which is 48% of the cost; and (2) A typical example of replacing human work in the to C area is Coworth. Irobot's sweeping robot. Effectively improve work safety. Examples include (1) large-scale deployment of service robots in the military field to replace humans for high-risk work. For example, a military mine-clearing robot equipped with a mine detector and a mine failure device can speed up the speed of mine clearance and greatly reduce the number of casualties; (2) Pac Technology's glass exterior cleaning robot ispider can be independent of the building's external walls. Cleansing is replacing the high-risk external wall cleaning of Spiderman. Concern 2: The product has the ability to extend the application scenario. After successful sweeping robots, a repetitive housework replacement market was verified. On this basis, the service robot manufacturers further tap into the relevant market potential, mainly from two directions, (1) near the market expansion: the same housework labor substitution market, manufacturers developed automatic cleaning glass robots and automatic weeding robots, examples include Cobos WINBOT Glass Cleaning Robot and JOHN DEERE's Automated Weeding Robot, (2) Expansion of target customers: Cobos sweeping robot was originally aimed at the to C market, and in 2016 the company launched to B for solar photovoltaic power station. Solar panel cleaning robot Rui Bao RAYBOT. (2) Assisting Humans Service robots assisting humans pay attention to technological advantages and the ability of technology to land in application scenarios. Service robots and humans can be cooperative relationships, and can improve human existing work efficiency by effectively utilizing technologies such as artificial intelligence, motion control, and human-computer interaction. The main driving force for the development of "assisting human" service robots is related technological progress. In this field, enterprises with technological advantages are generally occupying the leading position in the industry. Typical examples are Da Vinci surgical robots and IBM Waston. The prices of such service robots are mainly based on their value for improving human work. Take the Da Vinci surgical robot as an example. Using its assisted surgery, it is generally necessary to add 30,000 yuan to the cost of traditional surgery. The incremental value of the 30,000 yuan reflects the improvement of Da Vinci surgical robots' surgery. At present, Da Vinci surgical robots are sold at a price of 100-1.5 million US dollars in foreign countries, and are sold in the country for 3 million US dollars. We believe that such service robot companies that meet the following two criteria deserve our attention: (1) they have a leading edge in technology; and (2) they have the ability to use technology to land in application scenarios. Concern 1: The product has technical advantages. A typical example is a surgical robot whose requirements for mechanical motion control are extremely high. The Da Vinci surgical robot consists of a surgeon's console, a bedside robotic arm system, and an imaging system. With the Da Vinci robot, doctors can obtain 10 times more magnification of the three-dimensional vision, a more comfortable surgical environment, and higher-precision surgical procedures to improve the quality of surgery. Other star medical robots include Auris Surgical, an ophthalmic surgical robot, ARTAS, a hair-transplant robot, and Medrobotics, a flexible medical robot. Concern 2: Have the ability to use technology in the application scenario. A typical example is IBM Waston. The purpose of IBM's development of artificial intelligence technology is not to use machines instead of humans, but to extract core technologies from artificial intelligence, such as deep learning and pattern recognition, and then use these core technologies to help humans overcome their own deficiencies. Currently, IBM Waston is widely used in the medical field. In addition, IBM Waston also tried in many areas, including: (1) Analysis of team data: The NBA Raptors teamed up with IBM to use data analysis to select players, formulate game countermeasures, and temporarily adjust match plans. Changed the situation where previous coaches took brains to make decisions; (2) Smart wear to monitor health status: Waston provided users with exercise suggestions by analyzing user data; (3) Cooperated with robot body manufacturers: IBM implanted Watson in robot products. , To enable human-computer interaction; (4) To help the government manage haze: Waston assists the Beijing Municipal Government in combating haze by analyzing data such as air flow, sewage, social network commentary, and photographs; (5) assisting the government in implementing grading clinics : Waston recognizes public information, hospital information, patient information, etc., provides decision-makers with timely reports, and evaluates the effectiveness and problems of implementing policies. (3) Creating new fields Service robots used in the field of innovation need to pay attention to whether there are real needs. With the development of service robotics technology, manufacturers are not satisfied with replacing or assisting human work only in the original field, and they are also starting to tap new demand areas. Currently, they are mostly in the stage of market exploration and verification, such as intelligent virtual assistants, emotional robots, and programming. Toy robots and so on. Such products generally enter a new market area and need to follow and verify whether the new demand areas created by them are real demand. The success of Amazon Echo smart speakers has attracted a large number of imitators. Echo is a smart speaker introduced by Amazon in 2014. Amazon pre-installed the personal virtual assistant Alexa (Siri-like) in Echo. Users can interact with Echo to implement news broadcasts, weather inquiries, task reminders, music playback, and smart home functions. Echo has the potential to become a smart home of the future. The control platform. Echo is priced at $180 and cumulative sales have already exceeded 3 million units. Echo's success factors include: (1) The market demand is real: Echo knocks on the market through the form of home speakers, which is a huge and just-needed market; (2) The technical solution effectively solves the user's pain points: The original speakers need to be booted and selected. Music, tuning, and other steps, and Echo can immediately start working according to the user's voice instructions after hearing the wakeup word “Alexaâ€; (3) Amazon uses Echo as a strategic product to invest in R&D. Echo is more than just a speaker. It is also the home entrance of the Internet of Things. Amazon has established a R&D team with more than 1,000 people and has established its advantages in technology to its competitors. For example, Echo's average response speed has been within 1.5 seconds. . Echo's success has enabled competitors to see the potential of the smart speaker market. In 2015, HKUST and JD.com established a joint venture to establish Linglong Technology and launched the audio system. In 2016, Google launched Google Home. Compared to Echo, Google Home features include (1) deep integration with various services provided by Google; (2) support for simultaneous use of multiple rooms and multiple devices; (3) use only 2 microphone combinations, and Echo is 6+1 Microphone array; (4) More customized appearance. MIT launched emotional robot JIBO. JIBO is a family social robot developed by a team at the MIT Media Lab. According to the development team, JIBO has emotion-aware functions that can recognize and interact with user emotions. JIBO will focus on the mid- to high-end family market. At the same time, some B-side real estate, hotels, etc. are also potential users. In August 2015, JIBO introduced investors from China, Japan, and South Korea to pave the way for entering the Asian market. Among them, Eastern Power Company invested 2 million US dollars accounted for 2.34% of the equity. Orient Online will also help JIBO improve its video calling and localization. In addition, GQY Video also invested RMB 9 million in 1.1% equity in November 2015. However, at present, JIBO's product release has been repeatedly postponed, and its market potential remains to be verified. Microsoft Xiao Hai simulates human singing and intends to enter the virtual singer market. In August 2016, Microsoft held a conference for the fourth generation of Xiao Bing. The Microsoft team took advantage of big data training. The new Xiao Bing has been able to imitate human singing. Her "Invisible Wings" and "Summer Sweetheart" theme songs have received extensive attention. At present, the most famous virtual singer is Japan's Hatsune Miku. His singing is mainly generated using electronic sound synthesis technology, and he conducts concerts through 3D modeling and holographic projection technology. Microsoft Xiao Bing learns and sings through artificial intelligence technology and is expected to step out a new growth path for virtual singers.
GALOCE Canister Load cells are an integral part of a weighbridge, and must act as durable mounting hardware to facilitate the harshest of outdoor environments. Weightron offer two models of stainless steel weighbridge Load Cell,which are based on the same mechanical envelope with key design features ensure optimum performance, precision and long-term durability.
Canister Load Cell,Tank Weighing Load Cells,Tank Load Cells,Tank Scales Load Cells
Xi'an Gavin Electronic Technology Co., Ltd , https://www.galoce-meas.com