Multi- sensor information fusion technology was developed as early as the 20th century. It is an information processing process that uses computer technology to automatically analyze and synthesize information and data from multiple sensors or sources under certain criteria to complete necessary decisions and estimations. The use of multi-sensor information fusion technology has brought great convenience to people's work and life, and promoted the development of China's modern society.
I. Definition of Multi-Sensor Information Fusion
Multi-sensor information fusion utilizes well-known computer technology and specific criteria to automatically analyze and process information and data from multiple sensors, thus completing a complete information processing process. Multi-sensor information fusion technology was first used in the military field. In the military, multi-sensor information fusion can be defined as processing detection. It has been extensively studied in estimation and interconnection, yielding more accurate identity, threat, and situation estimates, thus bringing greater convenience to the military industry. As society has evolved, multi-sensors have become experts in information and data fusion. Research shows that the information processing and analysis capabilities of multi-sensor systems far exceed the human brain's ability to understand information. The human brain is a natural system for understanding, analyzing, and processing information and data. Everything the human eye sees and the ear hears converges in the brain, integrating and processing each visual, auditory, or olfactory sensation, and then estimating it using knowledge. In this way, humans understand everything around them. Multi-sensor information fusion technology follows the same concept, but its system's information and data processing capabilities are dozens of times higher than those of ordinary systems. Such systems, applied across various industries, undoubtedly possess extremely high information processing capabilities. By fully utilizing the resources of multiple sensors and using them rationally, and by complementing the information and data from each multi-sensor system in time or space, the measured object can be obtained within the multi-sensor system, thereby further optimizing the information processing capabilities inherent in the multi-sensor systems. However, it is important to note that when applying multi-sensor information fusion, the data from the multi-sensor system may exhibit different characteristics. Sometimes it may be real-time data, and sometimes it may be non-real-time data. Often, the data are mutually supportive or complementary, while at other times they are contradictory. Therefore, these phenomena must be carefully considered when applying multi-sensor information fusion.
II. Functional Description of Multi-Sensor Information Fusion
(a) Calibration
In multi-sensor information fusion, calibration functions are used to unify the spatial and temporal data from each sensor. However, in practical applications, the data acquisition frequencies of each sensor differ, leading to spatial discrepancies. Since each sensor operates independently, improvements are necessary to unify their operating times and acquisition frequencies. Various temporal and spatial calibration methods exist for multi-sensor systems, with interpolation/extrapolation and virtual fusion being commonly used.
(II) Methods for Multi-Sensor Information Fusion
Multi-sensor information fusion technology has been applied in numerous ways, with mature methods including cluster analysis, evidence theory, DS algorithm, and optimality theory. In recent years, a new information fusion method has been developed: the computer intelligent method. This method incorporates neural networks, wavelet analysis, and fuzzy set theory. The computer intelligent method has been widely adopted in a short period, demonstrating its impressive fusion performance.
III. Applications of Multi-Sensor Information Fusion
(I) Application of Multi-Sensor Information Fusion in the Military Field
For national security, every country invests significant financial, human, and material resources in the military field. my country has also invested heavily in multi-sensor information fusion technology and has achieved considerable success. Many countries, including the UK and the US, have developed numerous military systems with strong information fusion capabilities. The UK's multi-sensor information fusion systems include the "Artillery Intelligent Information Fusion System" and the "Mobility and Control System." The most ingenious aspect of the UK's "distributed data fusion" lies in its use of distributed technology within data fusion, while my country's data fusion technology employs centralized data, where all information or data is processed at a single level. A disadvantage of centralized information fusion systems is that if the central node is damaged or attacked, the entire multi-sensor information fusion system will be compromised. Distributed information fusion avoids this problem. All operations occur at individual nodes, meaning that even if one node is attacked or damaged, the entire system will not collapse. The British company developing this system has successfully tested it. The experiment involved eight different interconnected nodes, allowing the system to autonomously deploy. If the positioning radar detects an enemy, the autonomous drone receives the information and flies to the relevant area to assess the situation. The surveillance radar can track designated targets from the ground, maintaining virtual tracking or surveillance even when the target is not within line of sight. If one of the autonomous drones flies past a sensor, it can also integrate that sensor into the existing network—an advantage many systems cannot match. This will make the information collected by the multi-sensor system more accurate and secure.
(II) Application of Multi-Sensor Information Fusion in the Civil Field
With the continuous development of the economy and society, multi-sensor information fusion technology has been applied to the civilian sector and has achieved excellent results in agriculture. Its main applications include intelligent transportation, fault diagnosis of industrial robots, and autonomous driving. In autonomous driving applications, multi-sensor information fusion technology uses a control system to achieve driverless operation. This system is equipped with positioning systems, inertial sensors, digital maps, and other functions, enabling real-time control of the vehicle's direction and automatic detection of the ground shape. Simultaneously, the system can detect road conditions and obstacles ahead, making autonomous vehicles safer and more stable. After the development of autonomous vehicles, many people questioned their safety, but practice has proven that autonomous vehicles using multi-sensor information fusion technology possess sufficient safety.
Despite the continued development of multi-sensor information fusion technology and the existence of certain disadvantages in the system, researchers are constantly working to improve it. However, the application of multi-sensor information fusion has promoted my country's modernization and provided favorable conditions for the development of military and civilian industries. With the continuous maturation and development of related technologies, multi-sensor information fusion technology will surely drive more industries and fields in my country to achieve better development.
