In the field of industrial automation, the technology of sensors has always been a concern. The integration of these years of sensors is a hot topic in industrial automation manufacturing. The concept of sensor fusion began in the 1960s and has now evolved into a new level, that is, data fusion is a new topic closely surrounding sensor grids, and the discussion of sensor grids covers (wireless) sensors, the Internet of Things and so on.

Sensor fusion can also be understood in two ways: sensor prior knowledge (indirect fusion) or from scratch (direct fusion). It is also a simpler idea: use all available information, combine and form better information. Better information means more accurate, complete, and reliable than single or independent data source information. Distinguishing and categorizing is how data is combined. Assume that batch processing is identified, verified, and controlled by a large number of machines or various production steps. A large amount of information can be used later in the production process, but since many of the current information is not verified by the controller of a given application/device, the information is not shared. And if information is available, production can adapt, inspection systems become simpler and cheaper, and waste is reduced, so sensor integration is especially important.

Sensor fusion can significantly promote the production of small and medium-volume manufacturing, because the data it collects and the information of individual objects can achieve more flexible production. In the early stages of production, more and more users moved from final inspection to online inspection. By combining the results of the checkpoint with the sensor data in the manufacturing process, errors can be detected more quickly. Sensor fusion is especially useful for existing plant energy management, condition monitoring, and new systems that require extensive information for predictive maintenance. A lot of information is available, but collecting information specifically is not the end goal. For example, energy data can be collected from devices such as flow meters or AC drives; vibration is measured by sensors within the controller.