LONDON — Transformative technologies, such as artificial intelligence, the cloud, connectivity, robots, drones, and virtualization are shaping the semiconductor, display, and electronics manufacturing sectors, according to IHS Markit, which claims to be a world leader in critical information, analytics and solutions.
IHS Markit (Nasdaq: INFO) recently published the “Manufacturing Technology Vertical Intelligence Service,” including a manufacturer survey for adoption of transformative technologies in the semiconductor, display and electronics manufacturing sectors.
The survey found that 24% of respondents think transformative technologies have already been used widely in their factories, and 24% of the respondents say that they already have limited adoption. Only 15% of the respondents’ companies were found to be slower to adopt transformative technologies, with no current adoption or planned strategies in place.
These results look very positive for transformative technology adoption in the technology manufacturing space. IHS Markit said it thinks this trend is probably related more to transformative technologies such as the cloud, connectivity, and physical automation, while AI in manufacturing is still very much in its infancy.
Investments in transformative technologies include robotics
The following question is about which transformative technologies companies invested in most heavily last year.
As seen, opinions of how much AI influences the manufacturing vary widely. Responses regarding the importance of AI were split, with around 25% choosing “very important” and 25% choose “Not at all important.”
IHS Markit said that AI still needs more successful user cases to prove its value among transformative technologies in manufacturing.
The “Manufacturing Technology Vertical Intelligence Service” published by IHS Markit is not only a manufacturer survey, but it is also a combination of manufacturer user surveys, analyst insights, and vertical-/industry-specific studies. The analyst firm examines the six categories of transformative technologies with the industry domain knowledge by sector shown below:
AI moves into manufacturing
Focusing on AI, this transformative technology is being implemented in three ways in manufacturing: cloud-based AI, in-device, and hybrid mode.
Cloud-based AI has more computing power to analyze data, but there are potential issues around privacy, latency, and stability, which are very critical for manufacturing. Cloud-based AI could pave the way for technologically immature companies to utilize AI, radically transforming their usage and understanding of data.
Some industrial platform companies cooperate extensively with AI companies, such as Siemens (MindSphere), IBM Watson, and FANUC with Preferred Networks Inc.
Hybrid mode, combined with a stronger in-device AI and cloud-based AI, can help offset the aforementioned risks, albeit to only some degree. For instance, machine builders that deploy built-in AI machines are able to store data locally and thus safeguard their privacy.
The in-device AI or hybrid mode technology also has the benefit of real-time operation, which is very critical for manufacturing. Even without a network connection, the machines can still run AI functions properly. This will require the machines to have more computing power for AI capability.
Drones and machine vision
In the manufacturing and logistics sectors, industrial robots and drones have been deployed mostly in automated warehouses to help employees work more efficiently. Some larger industrial robot companies are trying to make robots smarter.
For example, FANUC is working with Preferred Networks to develop an industrial robot that uses deep learning to learning new tasks by itself. Google is also working on industrial robots that can learn from one another through large-scale interaction with the cloud.
Automated optical inspection (AOI) and machine vision with AI also have promising use cases in semiconductor, display, and electronics manufacturing. Machine vision is an embedded technology that extracts information from digital images through image sensors. Specific applications include image acquisition, processing, analyzing, and understanding.
There are many applications in the manufacturing sector, such as robot guidance, positioning for intelligent transportation systems, high-precision quality control and sorting, collision avoidance, and obstacle detection for aerial drones.
In the beginning, AOI machines used a set of simple rules to do the inspection, though they had a high false-alarm rate. Now, the machine learning and AI methods use self-adaptive, self-learning algorithms to improve the success rate of recognition. The big data analytics also help detect the potential failure of critical testing units.
For instance, IBM works with display panel makers on AOI machines to perform visual inspections. The AI system transforms raw data — such as edges, impurities, high-contrast areas, geometric features, abnormal texture, color and brightness features — into useful, high-level features such as position, similarity, distance, presence, and quality.
Based on large amount of defects image, human knowledges and labelling work, the AI creates models which can be deployed easily to automate quality-control inspections across key manufacturing processes in the semiconductor, display, and electronics industries.
Other transformative tech
In addition, edge computing, the Internet of Things, and augmented and virtual reality are other transformative technologies that could bring new opportunities to industrial automation suppliers, machine builders, and various players in the manufacturing ecosystem.
The IHS Markit Manufacturing Technology Vertical Intelligence Service forecasts the adoption of transformative technology in the display, semiconductor, electronic automotive, oil and gas, and other verticals through 2019, as well as:
- Critical intelligence of transformative technology, such as AI, cloud, connectivity, and robot adoption by sector
- Quantitative data and qualitative insight on manufacturing investment decisions and technology preferences
- Supply chain and ecosystem analysis covering materials, components, modules, and equipment for manufacturers
- End-product trends and influences on the manufacturing process, equipment and supply chain