What Is a Sensor Fusion?

The so-called Multi-sensor Information Fusion (MSIF) is the use of computer technology to automatically analyze and synthesize information and data from multiple sensors or sources under certain criteria to complete the required decisions and estimates. And the process of information processing. ^[1]

Multi-sensor information fusion

The basic principle of multi-sensor information fusion technology is like the process of the human brain's comprehensive processing of information. Various sensors perform multi-level and multi-space information complementation and optimized combination processing, and finally produce a consistent interpretation of the observation environment. In this process, we should make full use of multi-source data for reasonable control and use, and the ultimate goal of information fusion is to derive more useful information through multi-level and multi-faceted combination of information based on the separated observation information obtained by each sensor. This is not only to take advantage of the mutual operation of multiple sensors, but also to comprehensively process data from other information sources to improve the overall sensor system.

According to different data processing methods, there are three types of architectures for information fusion systems: distributed, centralized, and hybrid.

1) Distributed: The raw data obtained by each independent sensor is processed locally, and then the results are sent to the information fusion center for intelligent optimization and combination to obtain the final result. Distributed requires low communication bandwidth, fast calculation speed, good reliability and continuity, but the tracking accuracy is far from centralized. The distributed fusion structure can be divided into distributed fusion structure with feedback and without Distributed Fusion Structure of Feedback.

2) Centralized type: The raw data obtained by each sensor is sent directly to the central processor for fusion processing. Real-time fusion can be achieved. The data processing accuracy is high, and the algorithm is flexible. The disadvantages are high processor requirements and reliability Low, large amount of data, so it is difficult to achieve;

3) Hybrid: In the hybrid multi-sensor information fusion framework, some sensors adopt a centralized fusion method, and the remaining sensors adopt a distributed fusion method. The hybrid fusion framework has strong adaptability, taking into account the advantages of centralized fusion and distributed, and has strong stability. The structure of the hybrid fusion method is more complex than that of the first two fusion methods, which increases the communication and computing costs.

(1)

Multi-sensor information fusion technology has a wide range of applications, not only in the military, but also in civil applications. Military applications are the cornerstone of the birth of multi-sensor information fusion technology. Specific applications include marine surveillance systems and military defense systems. In the field of civil applications, it is mainly used for intelligent processing and industrial control. Intelligent processing includes robotic micro-surgery and disease monitoring in medicine, especially smart home.

Problems in multi-sensor information fusion technology

1) The ambiguity of correlation is the primary deficiency of technology. In a multi-sensor system, the data obtained by each scattered sensor will inevitably be restricted by many factors (such as the state of the environment and the characteristics of the sensor itself). Better development of multi-sensor fusion technology, and the issue of reducing the ambiguity of the association should be paid full attention to;

2) The basic theoretical system and perfect fusion algorithm need to be improved. Because the vast majority of fusion technologies are carried out in specific application areas. We must establish intuitive fusion guidelines for practical problems and form effective data fusion solutions. If there is a perfect theoretical system and fusion model, the blindness of fusion technology can be avoided. Such as asynchronous information fusion algorithm ^[5]

The information fusion system is a complex large system with strong uncertainty. The processing method is limited by existing theories, technologies, and equipment. Although this is a newly developed discipline and many theories are not sound, it will continue to improve with the development of various emerging related disciplines and technologies, and will gain more practical value.