April 11, 2022
Article written by Dr. Phil Yang
Figure 1. IOT flow chart
Artificial intelligence is familiar in language but still unfamiliar in its use. One of the specific directions is AIOT. AIOT is the combination of artificial intelligence and the Internet of things. Compared with AIOT, I personally prefer "AI driven IOT". Similarly, we call it "Ai enabled IOT".
IOT, translated as the Internet of Things, is a technology that connects sensors to various things and connects them with each other. In other words, the devices connected to the sensors are interconnected systems. IOT does not have to be connected via the Internet. However, in most cases, the Internet of things operates through the Internet.
AIOT, the AI driven IOT, combines the power of artificial intelligence. It is used for both ‘Edge AI’ and ‘Cloud AI’. It's just that the design and architecture of IOT are different. This means that AI technology is combined with IOT infrastructure, which improves the efficiency of IOT, enhances human-computer interaction, and improves data management and analysis. In particular, real-time data processing and required functions run faster without delay.
With AIOT, we can achieve economies of scale. For example, there is AIOT-enabled smart homes, AIOT-enabled smart buildings, AIOT-enabled smart cities, and AIOT-embedded factories. AIOT solutions and applications are countless. Autonomous car is also the product of AIOT, and wearable car equipped with artificial intelligence function is also the product of AIOT. Let's look at a specific case. Google uses AIOT to improve efficiency and reduce costs in the data center.
MAILab paid attention to AIOT from the start. First, we developed an edge device with built-in AI chip. One of the most representative examples is the AI camera. Although it can be learned on the server, the learning of data and algorithms is scattered to the edge, which is of course to effectively realize the connection between devices through AIOT. Moreover, to combine multiple functions, AIOT was essential. For example, for our QSS solution that integrates quarantine, security, and safety, AIOT technology had to be used because not only software but also hardware had to be connected through various devices.
Also, our urban micro-fulfillment service is the flower of AIOT technology. Due to the need to integrate more than 50 technologies, and each hardware device needs to use different algorithms, a high degree of AIOT design and architecture are adopted.
Finally, AIOT is an essential technology to improve the practicability and availability of AI. Of course, because there are many connection points, network security cannot be free from problems. Thus, MAILab has already been preparing for a year to establish a separate cybersecurity company, and the source of the core technology has been developed by a prominent American cybersecurity professor and has entered the stage of commercialization. For more effective development, we have set up a team overseas.
Now is not the time to get bogged down in the big word of AI. It is urgent to develop a specific application plan and business plan and application technology suitable for it.