Unmanned factories are usually considered the ultimate achievement in smart manufacturing but they are actually rare because key technologies needed to construct them far complicated than imagined.Progress in realizing the vision of completely automated production is not as smooth as originally expected, and in fact, unmanned factories are likely to be in decreasing demand with human-machine collaboration favored more for its flexibility.Unmanned factories are intended to completely replace humans with automated machines to reduce labor cost and hike production efficiency and product quality. But such an approach is built on the assumption machines can do better than humans.Obviously, robots cannot match humans in intelligence for the time being and machines lack the level of human sensing capability. Many vendors have integrated robotic arms with machine vision and force sensors to give them capability of precise collaboration between eyes and hands, but developers robots need tactile sensing to further hike efficiency.Robots also lack human adaptability. They are unlike humans who can adapt themselves to unpredictable changes in the surrounding and predict what is going to happen based on observation. Currently, most robots in use can work nimbly but there is not yet competent AI (artificial intelligence) technology to support them.The lack of flexibility is a key issue for robots, said Alan Chen technical support manager for Siemens PLM Software. When unexpected changes happen to manufacturing, or when new products come into production, manufacturers have to design production lines and place machines anew or adjust robot calibration. While completely automated production lines can effectively reduce labor input, for more complicated production lines, unpredictable mistakes are more likely to happen. Since robots are not yet capable of flexibly coping with such conditions, it may incur cost continually to solve problems.According to Masaru Takeuchi, general manager for Intelligent System Research Center under Omron, manpower, money and time to be spent on adopting robots are far more than imagined, with engineering cost likely to be 3-8 or even 20 times hardware cost. In addition, robots need to be maintained. As robots entail high total cost and are not flexible enough to meet production purposes, the benefits of using robots will diminish along with decreasing product lifecycles, which contradicts the mainstream trend in development of manufacturing.If manufacturers need mixed production lines or flexible production for diverse product models each in small volumes, basically they will not choose unmanned factories with 100% automated production lines because such a choice would incur higher production cost, Chen said.In comparison, 100% automated production is likely to be suitable for highly standardized products with production lines infrequently adjusted. Production of automobiles was previously thought to be suitable for total automation, but Daimler and BMW, for example, have instead incorporated collaborative robots at production lines to maintain flexible operation based on human labor, and have robots do dangerous, dirty and heavy-duty work in place of human labor.Collaborative robots change the interaction between humans and machines, allowing both to create as much respective values as possible.But collaborative robots are by no means limited to collaboration with human labor; they can be used in dark factories of a new type because such use is flexible, easy and inexpensive, Denmark-based Universal Robots (UR) has pointed out.UR recommends that small- to medium-size or startup manufacturers use automated production like a dark factory for manufacturing processes that are unsafe to workers, such as those involving very high temperatures or toxic gases, by adopting collaborative robots. In addition to the type of dark factories, collaborative robots can collaborate with workers in daytime and work independently during workers' absence at night, becoming another type of dark factories.

Taiwan makers are highlighting their robust prowess in robot production and smart manufacturing systems at the 2018 Taiwan Automation Intelligence and Robot Show running August 29-31 in Taipei, by unveiling their latest smart production solutions, and there are many iconic exhibitors such as Hiwin Technologies, Mirle Automation, Gallant Precision Machining, Compal Electronics and Nexcom International.As the most representative Taiwan maker of robots and automation equipment, Hiwin showcases a variety of Industry 4.0-based automation systems, including automated equipment for semiconductor uses; mechanical robotic arms incorporating vision systems and sensors for handling grinding, packaging and graving; smart ball screws, high-performance direct drive rotary table, and smart automation components.Besides industrial-use robots and smart manufacturing solutions, many firms also display service-oriented robotic applications. Taiwan Shin Kong Security, for instance, has displayed robots to support smart living security, and healthcare-use robotic arms designed to help nurses transfer patients and intelligent walking aids can also be seen at the show.Taiwan Intelligent Robotics exhibits customized robots for use at recreation, dinning, and biotech fields, while Arobot Innovation also showcases small-size robots that can accompany little children at home.

Taiwan-based Anchor Tech, a provider of smart buildings control systems, is actively striving to apply its control, monitoring and maintenance solutions to commercial properties including Carrefour, RT-Mart, and convenience stores, so as to keep the premises in an optimal operating state through predictive maintenance, according to the firm's CEO Duncan Huang. Huang said in the IoT era, the niche market for startups like his company, founded in 2010, rests with the space for integration, flow optimization and maintenance services, which accounts for 5% of business opportunities which big enterprises can hardly tap as they are busy developing, producing and selling products.In terms of smart illumination, he said, radar sensors and LED driver can be combined to allow indoor lights to automatically switch on and off, or to detect sunlight brightness at signage boards to determine where to switch on or off lights and luminance. As to predictive maintenance, sensors can be used to collect data on the frequencies and waveforms of lights to accurately predict when every light will wear out, so as to reduce the cost for patrol personnel and provide customers with proper maintenance solutions.Huang stressed that his company will move to increase long-term added values for products already sold by offering maintenance solutions to customers, aiming to become a service provider and ecosystem builder rather than a product supplier.

While mature infrastructures, stable system technologies, sound supervision regulations and public acceptability are required to support the development of driverless vehicles, such vehicles for public transportation will work in three main scenarios: grid network in smart city, smart tourism in theme parks, and shuttle services in remote areas, according to Ting Yen-yun, president of 7Starlake, a Taiwan startup dedicated to developing smart ride-sharing shuttle buses.Ting said that public transportation highly relies on the support of good road networks, and only after the network coverage reaches a certain scale can efficient transportation be achieved. He continued that autonomous-driving minibuses can be applied to build virtual road networks and sharply reduce the lead time and cost for public transportation.In terms of smart tourism application, Ting noted, the mini buses used in amusement parks must boast smart mobility, digital experience and ride-sharing functions, with AR (augmented reality) technologies applicable to mark the locations of scenic spots and introduce the natural landscapes and attractions.Driverless vehicles can also be applied for flexible shuttle services in remote areas, where maintaining a fixed transportation route usually involves unaffordable costs. Automated-driving minibuses can also help bring medical services to remote countryside, Ting continued.

In 2017, 380,600 industrial robots were sold globally, increasing 29.3% on year, according to World Industrial Robot 2018 published by International Federation of Robotics (IFR).Of the total, 125,400 were purchased by users in the automaking industry, growing 21% on year. Shipments to the manufacturing industries of metals, electronics/electric appliances and food rose 54%, 27% and 19% respectively. Although growth in global car sales is slowing down, automakers keep hiking automation of production, especially China-based ones and makers of electric vehicles.Among regional markets, 255,000 industrial robots were sold in Asia and Australia, up 34% on year; 67,000 units in Europe, up 20%; and 50,000 units in North and Latin Americas, up 22%.China, South Korea and Japan were the three largest country markets with sales of 138,000, 40,000 and 39,000 units respectively. The US ranked fourth with 33,000 units, followed by Germany (22,000) and Taiwan (11,000).Manufacturing industries around the world recorded an average use density of 74 industrial robots per 10,000 workers, with South Korea posting the highest density of 631 units per 10,000 workers. Taiwan's use density stood at 177 industrial robots per 10,000 workers, ranking 10th globally. While China ranked 23rd, its government aims to advance the country to the top-10 in terms of industrial automation with use density to rise to over 150 industrial robots per 10,000 workers in 2020.

Developing AI (artificial intelligence) solutions for industries face many challenges mainly because of lack of AI talent and availability of necessary data, according to experts.And AI solutions are practically case-specific rather than for common use within a specific industry, the experts added.Some market research organizations predict that AI will be mainly applied to four areas: financial services, medical/health care, retail operation and manufacturing. But manufacturing has the biggest potential for AI applications in Taiwan because of the country's industrial structure, followed by financial services and medical/health care, said Yu Shaw-shian, senior VP for government-sponsored Industrial Technology Research Institute (ITRI). Big data is the core to drive AI, and Taiwan's manufacturing industry consists of comprehensive supply chains that offer sufficient data for development of AI solutions.Taiwna-based manufacturers fall into three categories in terms of readiness in collecting and accumulating operating data from production equipment, Yu said. The first is high-tech firms such as semiconductor and flat panel makers and/or large-size enterprises which have collected such data, Yu noted.The second is makers which have collect insufficient data, but much of the data can be specially processed for use in developing AI solutions, Yu indicated.The third refers to makers whose machines are of old models that cannot be web-connected to collect data, Yu said. Most of them are small- to medium-size firms and the third category accounts for over 95% of the total number of manufacturers in Taiwan, Yu noted.External assistance, such as the use of smart machine boxes, is needed to enable such old machines to collect operating data, resulting in much time taken and less efficiency in developing AI solutions, Yu explained, adding thethird group pose the main difficulty in boosting industrial AI applications in Taiwan, Yu pointed out.The time taken to accumulate enough data varies: for example, at least one year is needed for retail operation to reflect seasonal effects, and a few months to half a year for PCB makers, Yu said.However, makers of different product lines may differ much in time taken to accumulate sufficient data and this is why it is difficult to develop AI solutions for common use in an industry.While training in developing AI solutions is commonly based on deep learning, the purposes of using AI solutions vary from enterprise to enterprise, involving much higher multiplicity and customization than adoption of ERP (enterprise resource planning) or CRM (customer relationship management) systems, said CEO Chen Sheng-wei for Taiwan AI Academy.

Manufacturing of electronic components including ICs and PCBs will account for 36% of collaborative robots to be deployed in Taiwan in the future, followed by metal working with 32%, retail logistics with 15%, food processing with 11% and shoe making with 6%, according to Industry, Science and Technology International Strategy Center under government-sponsored Industrial Technology Research Institute (ITRI).While demand for all types of industrial robot in China, Germany, Japan, South Korea and the US kept increasing at a CAGR of about 13% during 2011-2017, global demand for collaborative robots alone hiked 89% on year in 2016, 75% in 2017 and is forecast to hike 65% in 2018, the center said. Human-machine collaboration functionally fills the gap between 100% manually operated production and completely automated production and is becoming the mainstream mode in the global industrial automation market, the center explained.For manufacturing of electronic components, collaborative robots will paired with AGVs (automated guided vehicles) for feeding, cutting and moving of materials mostly, the center noted. For metal working processes, collaborative robots will be used in feeding, cutting, moving and carrying of materials as well as grinding, paint spraying, gluing and polishing, the center indicated.

Artificial intelligence (AI) can be applied to the travel service industry to better gauge the real needs of consumers through deep machine learning, according to CH Kuo, chief strategy officer of Lotus Group, a travel service and technology group based in Canada.Kuo, who is leading the Taiwan branch of Lotus toward a smart operation, said that Taiwan travel service firms must strive to work out travel itineraries that can increase consumer experiences through the support of AI.Kuo said that if consumers have a language barrier in travel, AI can help incorporate online translation service into their travel itineraries to let users of the service more easily know travel situations in advance and remind them of what to do before arriving at destinations. This is part of the service products developed by Lotus based on "thick data," Kuo added.Unlike big data, thick data highlights the depth of data, utilizing qualitative research to analyze a few data samples to dig out insightful points which are likely to be neglected by big data, Kuo said, adding that the deep machine learning technology develop by his company is based on thick data, able to conduct detailed analysis of traditional travel packages and relevant situations consumers may encounter.He stressed that AI should be based on the accumulation of intelligence rather than focused only on the development of technology, and therefore his company insists on "design for customer's need" by utilizing IT technologies to help upgrade traditional travel services.

Taiwan government-sponsored Industrial Technology Research Institute (ITRI) has kicked off operation of AIdea, a platform for developing AI (artifical intelligence) solutions to enterprises' problems, according to Yu Shiaw-shian, ITRI senior vice president and general director for Computational Intelligence Technology Center.AIdea first allows enterprises to present operational problems and related data. Sensitive information such as product names will be deleted from the data to protect business secrets. The data will be made available on the platform for open use by academic and research organizations to build models or develop algorithms that form AI solutions to the specific problems.AIdea plays the role of collecting industries' operational problems and related data and matching them with academic and research organizations' capability of developing AI solutions, Yu said. Enterprises are required to present a certain level of accumulated data on their operation because AI is based on big data analysis, Yu noted, adding the richer the data, the more likely AI can be applied.AIdea's main target users are small- to medium-size enterprises, Yu noted.As Taiwan has well developed medical care systems and a comprehensive national health insurance database, it offers good opportunities for developing AI technologies for medical care, such as precision medicine, Yu indicated.Several local enterprises or organizations have participated in AIdea, including Taipei Medical University, Taiwan Taxi, China Petrochemical Development and Asustek Computer.

A Taipei park has kicked the official use of Taiwan's first smart public lavatory allowing citizens to experience IoT-backed smart living, and the lavatory will double as an emergency call center in the next stage to enhance night safety for citizens, according to Spencer Liang, chairman of Flowring Technology playing a role in the project.Liang said the smart lavatory mainly relies on the support of IoT and cloud management technologies, with sensors utilized to detect the usage volumes of tissues and hand cleansers as well odors such as heavy ammonia before reminding service staff to do the replenishing and cleaning jobs. Liang said the sensing technology is just a basic application of IoT, but the smart lavatory project has attracted great attention from various sectors, with the Environmental Protection Administration even asking his company to help promote similar projects to other places in Taiwan.The pilot smart lavatory in the Daan Forest Park will be connected to the police systems later to become also a night emergency call center to better safeguard park visitors at night, according to Liang.The pilot project will also gradually incorporate AI, big data analysis, as well as facial recognition, user flow recognition and dangerous behavior recognition applications in the future, Liang stressed, adding that the most valuable part of the project is the backend use of the accumulated data for more efficient management.