Event Detection Algorithm In Wireless Sensor Network Based On Max-Flow/Min-Cut Principle

The wireless sensor network is constituted by a large number of micro sensor nodes with a data processing unit and a communication unit, these nodes are usually cheap and its computing power and storage capacity is limited. The detection of events in the physical environment is one of the important applications of wireless sensor networks. The energy saving is very important, when using the wireless sensor networks for event detection in the wild, because the energy of the sensor nodes is not easy to supplement, energy saving can prolong the life of the entire network. Further, in order to avoid unnecessary manpower and cost, the detection accuracy is very important. The algorithm proposed in this paper, compared with the previous algorithm, it can reduce the traffic, save energy effectively with a higher accuracy. Meanwhile, a practical application of the project provides a detailed theoretical foundation and basic hardware knowledge for the algorithm design. And it Can also be expanded for the algorithm implementation and verification.There are many works focus on saving energy and improving accuracy,but for event detection in the wild via wireless sensor network, energy-saving is an important topic since it is hard for the sensors to refuel energy. On the other hand, the accuracy of detection is also very important taking into account of the resources consumed. Previously, many algorithms have been proposed to save energy and/or to improve the accuracy of detection. However, these algorithms do not perform well on a certain kind of boundary-abrupt event, such as wild fire, in which the difference of the sensor readings between neighboring sensors is maximized on the boundary. In this paper, we propose an algorithm, BMinCut, based on the Max-Flow/Min-Cut algorithm to detect such kind of events. In our algorithm, we first convert all the sensor readings into an image. To reduce traffic and computation complexity, we further cut down the image. After that, we segment the image based on the Max-Flow/Min-Cut algorithm. Finally, we detect the event region by identifying foreground and background. Taking the wildfire events as an example, we have conducted simulation experiments to test the performance of our algorithm. Experiments show that our algorithm can reduce traffic complexity and save energy, while achieving high accuracy of detection. Besides, our algorithm can guarantee the accuracy of detection without increasing the complexity for multi-event region.for design an algorithm to solve the existing problems,must to understand the problem and the inherent limitations. So we want to solve the wireless sensor network event detection accuracy and energy efficiency issues, we must understand the limitations inherent in the working mechanism of the entire sensor network and hardware. This paper designs and implements the college classroom intelligent energy-efficient lighting systems, not only to solve the phenomenon of the college classrooms electricity wasted, and also understand the basic knowledge of the wireless sensor network from the ground up, and lay a solid foundation for future algorithm design.In addition, this system can be changed to implement the Max-Flow/Min-Cut algorithm.There are design of sensor node and sink, as well as corresponding development of control program and upper-computer software. The system sets single-chip Amegl6as control center, realizes communication between nodes via nRF24L01wireless transceiver module, and realizes communication between sink and upper computer via w5100wireless internet module. It perceives illumination intensity via photoconductor, detects the human body position via infrared pyroelectric sensor, and places the sensor node on the lamp, so the light can be controlled according to position of human body and current illumination intensity, which can realize energy saving to a large extent on condition that lighting requirement is satisfied. The system has low cost as well as simple arrangement, and there is no need to change the original lighting circuit. The light can be turned off by hand, and when multi-media are used for the class, light can keep off even it is dim. In addition, this system has the function of automatic fault report, which is convenient for property maintenance.