The global smart robot market is expected to grow to $33.6 billion in 2021, and Asia will be the region with the most growth. In recent years, all countries have regarded robots as a strategic industry. To some extent, they can even be a manifestation of comprehensive national strength. Leading manufacturers in various fields are actively investing and developing.
Robots have undoubtedly become the next star of technology tomorrow. All countries in the world are actively promoting the robot industry. Recently, the development of artificial intelligence and deep learning technology has become more and more important, and it has become an important driving force for the development of intelligent robots. According to the IEK research report of the ITRI, the global smart robot market is expected to grow to $33.6 billion in 2021, and Asia will be the region with the most growth.
According to the definition of the Precision Machinery Research and Development Center, intelligent robots can sense the environment through sensors and achieve intelligent understanding through programmatics, and finally reflect the required actions to perform various production activities, provide services or interactive. It is a platform that integrates various technologies, including mechanical, control automation, electronics, motors, imaging, optics, communications, software and security systems and other related technologies and applications, of which software and hardware integration technology is of paramount importance. This course explores the future of the smart robot industry and analyzes key technologies, components, and hardware and software architectures.
Service robots have development potentialAccording to the MIC research data of the CMC (Figure 1), the total size of the robotics market in the four major application areas was approximately 26.9 billion US dollars in 2015, of which the industrial robots accounted for the highest proportion of 11 billion US dollars, but by 2025 the overall market size will expand to 66.9 billion. In the US dollar, although the market size is still the largest in the industrial robots of 24.4 billion US dollars, the composite growth rate (CAGR) of commercial robots and personal robots in 2000-2025 is 11.6% and 17.4%, respectively, and the MIC industry analyst Zhang Jiaxuan ( Figure 2) points out that especially after 2015, these two types of applications have grown more significantly. In the service-oriented application market, there are many emerging fields that have not been imported into robots in the past, driving their growth potential.
Figure 1 The scale of the robot industry in various application fields in the world from 2000 to 2025 Source: BCG, 2014; MIC, MIC, 2016/1
Figure 2 MIC industry analyst Zhang Jiaxuan pointed out that after 2015, the service-oriented application market has many emerging fields that have not been imported into robots in the past, driving its growth potential.
Japan's software bank has made great strides in the field of robots in recent years, and a series of actions have attracted market attention. Zhang Jialu said that including the acquisition of the French humanoid robot company Aldebaran RoboTIcs in 2012, the humanoid robot Pepper launched in 2014 has cooperated with IBM Watson and Microsoft Azure. Softbank proposes a vision of providing home and business applications based on communication. Pepper is set to be a robot that wants to be loved. It understands the members of the home through interactive communication and becomes a part of the family. Based on artificial intelligence, Let Pepper assist in the marketing of corporate products, with both entertainment and learning effects in the home. In addition, there are currently more well-known service robots that have been put on the market, including Leka and Savioke.
In recent years, all countries have regarded robots as a strategic industry. Japan has been developing robots for a long time. In 2015, the Japanese government established a robot revolution initiative agreement to promote the development of the robot industry. South Korea is led by the Ministry of Industry, Trade and Resources, and develops basic plans every five years. The goal is to become a robot in 2022, with a production scale of 25 trillion won; the United States, starting in 2011, is led by the National Science Foundation (NSF) to develop robotics that work safely with people. At present, Korea is the main force for the development of home robots, while the United States is leading the country in disaster relief. Zhang Jiaxuan suggested that the layout of home applications and public applications has been deep, and commercial applications have recently emerged. Taiwan can wait for opportunities.
Deep learning, speech recognition and other technologies have developed significantly in recent years, which has led to the rise of the service-oriented robot industry and applications. The robot has evolved from the past one-way communication and execution commands, and can understand the meaning of the semantic response dialogue, and the application service is the follow-up development of the robot. Focus. The application environment of robots is diversified. In different occasions, it is necessary to combine the expertise of various fields and the understanding of user needs. Therefore, manufacturers should accelerate the application of robots in various fields through an open platform.
Neural network-like technologyThe term deep learning is because the artificial intelligence AlphaGo lost the South Korean chess king in 2016, the machine successfully challenged the human brain for the first time, and in the game game, which is generally considered to be the most difficult, is once again concerned by the world public. AlphaGo's deep learning core is the neural network technology. As early as 1943, Warren McCulloch and Walter Pitts first proposed the mathematical model of neurons. Then in 1958, psychologist Rosenblatt proposed the concept of Perceptron. The mechanism of training correction parameters is added to the structure of the former neuron. At this time, the basic academic structure of the neural network is completed. Neurons of the neural network actually collect various signals (like neural dendrites) from the front end, and then weight each signal according to the weights and then transfer them to a new signal through the activation function (similar to neurons). Axon).
The related technical architecture was actually completed as early as the 1970s. Yin Xiangzhi (Fig. 3), the chief of data decision technology, said that deep learning is actually another kind of neural network. Its success comes from a deeper understanding of the operation of the human brain. The ConvoluTIonal Neural Network assists machines in developing true vision. The two main principles are: local perception and weight sharing. Let the machine understand the overall meaning from the fragment features, and then find out the clustering of the features, and continuously analyze the layered and refined, no matter how detailed the features are: as long as they are not grayed out, the features can be extracted.
Figure 3 Data decision-making technology director Yin Xiangzhi said that the deep learning technology architecture was completed in the 1970s, and its success comes from a deeper understanding of the operation of the human brain.
Among them, graphic recognition is the key point. In the past, the central processing unit (CPU) and the graphics processing unit (GPU) handle different computing functions. For the deep learning function of graphic recognition, the performance of the GPU is One hundred thousand times of the CPU, Yin Xiangzhi further said that through deep learning, the machine can even remove the original mosaic image effect. However, in the recognition of speech and text, Chinese is still a big challenge for machines. Chinese vocabulary is more than one million. No new words can be created and words can be created without convention. There are also many Chinese, English, Chinese, and Chinese Words such as: blue skinny, shiitake mushrooms, 94 madness, etc.
Dachang collects data layout for the futureThe prospects of the robot industry have attracted attention from all sides. In particular, deep learning and artificial intelligence have become the next wave of enterprise development in major manufacturers, including Facebook, Microsoft, Google and Amazon. The commonality of these enterprises is that they interact with consumers through products and services, and have accumulated many years of primary data. In the future, artificial intelligence and deep learning will be filled with a large amount of data collection, sorting and classification, and tagging to make these primary. The data becomes information, and finally the results of the quick search and reaction through a powerful processor.
Observed from such an architecture, Qiu Renjun (Fig. 4), the general manager of Shuowang Information, believes that the data structure of FB among the four major factories is the highest, because each user has already sorted the content when uploading articles or pictures. The image resolution is high and even the characters in the photo are directly labeled. In the future, when the FB is further sorted or utilized by the data, it can take the least time or perform higher quality finishing. At present, 80% of the data in the world is unstructured, and cognitive computing can enhance and simplify the learning process.
Figure 4 Qiu Renzhen, general manager of Shuowang Information, believes that 80% of the data in the world is unstructured, and cognitive computing can enhance and simplify the learning process.
Therefore, the robot should reduce the error rate, focusing on the integrity and structure of the data. Qiu Renjun further explained that the application process of deep learning is from the underlying neural network computing, massive data analysis, discovery rules/automatic classification, and media generation. Combine/recommend strategy, record user behavior, and feed back to the model/improve accuracy. The ultimate goal is to improve data quality and generate self-learning correction mechanisms.
Portable Power Station For Power Tool
COVID-19 Impact:
The COVID-19 pandemic has hit the global portable power station market badly. Closure of public spaces and recreational activities can lead to a slump in sales. Restrictions on travel plans and closure of tourist destinations can negatively impact the market. But areas with frequent blackouts are commanding a huge demand for portable power stations to survive the pandemic.
Industry Trends:
The need for reliable survival gear in treks and areas with low power can drive the demand for portable power stations. Manufacturers in the market are increasing the capacity of these generators by providing extra outlets for charging more devices. Rise of recreational activities and surge in outdoor events has led to sale of power stations. New generation trekkers are opting for camping and caravanning with backup charging equipment being on top of the priority list. Crowdfunding campaigns for driving development of larger capacity power generators will fare well for the market.
Segmentation:
By operation type, the global portable power station market is divided into direct power and solar power. The solar power segment is expected to grow at the fastest rate in the portable power station market. This is attributed to the low dependence on fossil fuels for energy generation and storage. Renewable schemes and subsidies designed to create public awareness and interest can drive the segment demand.
Portable Power Station,Portable Cordless Power Station,Portable Power Station For Power Tools,Portable Power Supply For Power Tools
Shenzhen Zhifu New Energy Co., Ltd. , https://www.sunbeambattery.com