Abstract: This paper introduces a burglar alarm monitoring system using multisensor data fusion technology. The paper presents the sensor selection principles and the system's hardware structure, and discusses several models of multisensor data fusion. Keywords: multisensor technology, burglar alarm system, data fusion 1. Introduction Over the past two decades, multisensor data fusion technology has gained widespread attention and application. The term " fusion" has been used almost indiscriminately in numerous application fields. Information fusion is a new research direction in information processing that addresses the specific problem of multiple sensors (multiple or different types) in a system; therefore, information fusion is also known as multisensor fusion. According to foreign research results, the definition of information fusion can be summarized as: the information processing process that uses computer technology to automatically analyze and synthesize the observation information of several sensors obtained in time sequence under certain criteria to complete the required decision-making and estimation tasks. It can be seen that multi-sensor systems are the hardware foundation of information fusion, multi-source information is the processing object of information fusion, and coordinated optimization and integrated processing are the core of information fusion. This paper discusses two aspects: the application of multi-sensor technology and the design of data fusion models. 2. Sensor selection Multi-sensor integration and information fusion technology mainly uses multiple sensors or multiple similar sensors to collect data, applies data fusion technology to process data, extracts useful information, and realizes measurement and control. The principle to be followed when selecting sensors is: select sensors of different types and functions, give full play to the advantages of each sensor, complement each other's strengths and weaknesses, realize information sharing, and reduce the possibility of false alarms and missed alarms. Therefore, based on the above analysis and principles, in order to achieve the purpose of anti-theft monitoring, the following sensors can be selected to form a multi-sensor anti-theft alarm system: (1) Pyroelectric infrared sensor. All objects in nature, such as the human body and flames, emit infrared radiation, only the wavelengths of the emitted infrared radiation are different. The human body temperature is 36-37℃, and it can emit infrared radiation with a center wavelength of 9-10μm. By attaching a filter with a cutoff wavelength of 7-10μm to the surface of a silicon wafer, only infrared radiation with wavelengths greater than 7-10μm is allowed to pass through, while infrared radiation with wavelengths less than 7μm is filtered out, thus obtaining a pyroelectric infrared sensor that is only sensitive to the human body. (2) Ultrasonic sensor. Ultrasonic waves have strong penetrating power, have a certain directionality, have small attenuation during transmission, and have strong reflection ability. When someone stands in front of the sensor, due to the reflection of the human body, a signal will be detected in the receiving circuit. This signal is amplified and input to the microcontroller for judgment and alarm. (3) Doppler sensor. When there is relative motion between the wave source and the observer, the frequency of the wave changes. This phenomenon is called the Doppler effect. The Doppler effect sensor can be made using the Doppler effect. When a human body or object moves relative to the sensor, the reflected signal and the original signal are frequency-shifted. The integrated circuit then amplifies the weak frequency-shifted signal, and after Doppler detection, amplification, and limiting, a DC output level related to the object's movement signal response is finally obtained. (4) CCD sensor. A CCD sensor is a photoelectric sensor array composed of photodiodes. Its structure is an interline transfer type. This device has a large photosensitive area and high target surface utilization. When the optical image of the scene is projected onto this array through the camera lens, different amounts of photocharge are induced due to the different light intensity received by each photodiode. After a certain period of time (field), these induced charges are horizontally transferred to the vertical shift register located next to the photosensitive element under the control of the transfer grid. Then, under the control of the horizontal transfer pulse, the charges are transferred to the horizontal shift register and sequentially transferred to the output end under the control of the horizontal shift clock. Finally, the video signal is output by the output circuit. The video signal output by the CCD sensor already has a relatively large amplitude (above 0.5V). After processing by the processing circuit (including automatic gain control, correction, synchronization signal mixing, power amplification, etc.), a full television signal is output at the terminal. These four sensors monitor human intrusion from different aspects such as infrared radiation emitted by the human body, reflection of ultrasound waves by the human body, changes in wave frequency caused by human movement, and human images. It can basically realize the collection of all source information of human intrusion, thereby eliminating the uncertainty of alarm information to the greatest extent and improving the accuracy of alarm. 3. Composition of the alarm system 3.1 Hardware structure of the system The hardware block diagram of the multi-sensor integrated anti-theft monitoring alarm system is shown in Figure 1. [align=center] Figure 1 Hardware block diagram of multi-sensor integrated anti-theft alarm system[/align] The four sensors, namely pyroelectric infrared sensor, ultrasonic sensor, Doppler sensor, and CCD sensor, respectively collect different characteristic parameters in human intrusion information. The information is transmitted to their respective signal conditioning circuits for amplification, filtering to remove interference, and A/D conversion. The conditioned information is then transmitted to the microcontroller, which performs data preprocessing to realize the judgment of human intrusion and trigger an alarm. In network applications, information is transmitted to a host computer—a PC or industrial control computer—for information processing, enabling overall situation monitoring. 3.2 Data Fusion Technology in Multi-Sensor Technology Multi-sensor integrated data fusion models can be mainly divided into functional, structural, and mathematical models. The functional model starts from the fusion process, the structural model from the composition of information fusion, and the mathematical model is the information fusion algorithm and comprehensive logic. According to the five levels of information abstraction, information fusion is correspondingly divided into five levels: detection-level fusion, location-level fusion, target recognition-level fusion, and situation assessment and threat estimation. Since fusion itself mainly occurs at the detection, location, and target recognition levels, only the fusion structures of the first three levels are considered when discussing structural models. This paper only discusses several data fusion structural models at the detection level. At the data detection level, the main structural models of data fusion include serial, parallel, parallel with feedback, tree-like, and distributed types. For different targets and application areas, one or more structures can be used in combination. In anti-theft monitoring, the serial data detection fusion structural model shown in Figure 2 can be used to accurately obtain information about human intrusion at the scene. First, a pyroelectric infrared sensor (P) detects the presence of infrared radiation with a center wavelength of 9-10 μm emitted by a human body, makes a decision, and transmits the result to the ultrasonic sensor (U) node. Here, the detection results of the two sensors are fused to form a new decision. This process is repeated, and the data is then transmitted to the node consisting of a Doppler sensor (T) and a CCD sensor (C). Finally, the decision result is fused and output. Alternatively, a parallel data fusion structure as shown in Figure 3 can be used. In the parallel structure, the detection data from the four sensors are processed simultaneously at the data fusion center. Regardless of the situation, the data is detected and transmitted to the microcontroller, and then the result is output. The characteristics of the parallel method are that the alarm settings are simple and clear, but the possibility of false alarms is also higher. A hard decision method can be used to simplify the data detection processing and simplify the hardware. [align=center] Figure 2 Schematic diagram of serial structure model[/align] [align=center] Figure 3 Schematic diagram of parallel structure model[/align] A hybrid serial-parallel fusion structure can also be used, which can more comprehensively reflect the human intrusion situation in the detected environment. Its structure is shown in Figure 4. Each linear component can be output as a separate result, and the fused decision can also be output. [align=center] Figure 4 Schematic diagram of the serial-parallel hybrid structure model[/align] 4. Conclusion Multi-sensor integration and information fusion technology has achieved great success in the fields of automatic control, military, and aviation navigation. With the rapid development of sensor technology, computer technology, and information technology, multi-sensor technology will continue to develop and gradually be applied to other fields such as burglar alarm systems. References: [1] Li Hua, ed. Practical Interface Technology of MCS-51 Series Microcontrollers. Beijing: Beijing University of Aeronautics and Astronautics Press, 1999. [2] He Xicai, ed. Sensors and Their Applications. Beijing: National Defense Industry Press, 2002. [3] Horst Ahlers, ed. Multi-Sensor Technology and Its Applications. Beijing: National Defense Industry Press, 2001. [4] He You, ed. Multi-Sensor Information Fusion and Its Applications. Beijing: Electronic Industry Press, 2000. [5] Ding Zhensheng, ed. Sensors and Sensor Technology Applications. Beijing: Electronic Industry Press, 1998. [6] Li Yufang, ed. Sensing Technology. Sichuan: University of Electronic Science and Technology of China Press, 1999. [7] Liu Tongming, ed. Data Fusion Technology and Its Applications. Beijing: National Defense Industry Press, 2000. [8] Liu Yingchun, ed. Principles and Applications of Modern New Sensors. Beijing: National Defense Industry Press, 1998.