Indoor WSN Node Fault And Health Diagnosis

As a branch of wireless sensor networks (WSN), indoor wireless sensor network has been widely used in industrial and agricultural production. Because of the constraints of the sensor node and environmental impact, network performance inevitably decreases during the operation of the network. Hence, considering the characteristics of the indoor wireless sensor network to study the diagnosis of sensor node fault and the health of the network has great significance.First, take a typical indoor wireless sensor network application, the green house wireless sensor network as example, an indoor wireless sensor network health diagnosis system is established, including short-range wireless communication option, ZigBee star network networking, hardware and software design of the center node and sensor nodes, and the debugging of the whole system. Then, based on the system established, sensor nodes fault detecting method is studied by adopting principal component analysis (PCA) that divides the data space into the principal component subspace (PCS) and the residual subspace (RS), and sensor nodes fault identification location method is discussed by using high-frequency coefficients of wavelet decomposition. Finally, the health evaluation model of indoor wireless sensor network to evaluate the operation of the network is built by using fuzzy analytic hierarchy process (FAHP) analyzing the factors that influence the healthy running of the indoor wireless sensor network to determine hierarchical single order vectors and hierarchical general order vectors.Indoor wireless sensor network health diagnosis system commissioning shows that system running stable and reliable. Sensor nodes transmit the temperature, humidity and light intensity information to the upper computer with low power sleep and wake up mode in real time. Only two-byte diagnosis information-Square Prediction Error (SPE) values-is added, and the time cluster node detecting the fault is far less than the system collecting the data, therefore, this method is energy efficiency and real time. In all the factors that influence the indoor wireless sensor networks, failure nodes have the largest weight. Combined the weights of the factors and the single index evaluations, the health degree index given is effective that provides the basis for the network maintenance personnel to understand the overall operation of the network.