Abstract-Sensor failure have become a major bottleneck, which restricted the wide application of wireless sensor networks WSNs (Wireless Sensor Networks). In order to satisfy the needs of industrial production and daily living environment monitoring, it is important to improve the survivability of wireless sensor networks. This paper analyzed the damage types and causes of WSNs and the measurement methods of WSNs survivability. Then, based on the whole system model and sensor node circuit, we studied the fault detection method, which can improve the survivability of the scheme. Additionally, the robust guarantee scheme through hardware design and algorithm model was proposed. The performance results shown that the proposed scheme can realize the remote wireless communication services and prolong the network life cycle, so as to improve the survivability of WSNs. Through the combination of theoretical research and practical analysis system, the network survivability would be improved through soft and hard optimization combination strategy.Keywords-Wireless sensor networks; invulnerability; indoor Complex environment monitoring; hardware design; robust guarantee
IntroductionThe environmental monitoring system [1, 2] based on WSNs (Wireless Sensor Networks) is used to reduce the cost of the entire system, mostly using mass low-cost sensor nodes [3] for networking. Invulnerability of WSNs is the ability measurement metric of WSNs, which could be sustainable and stable to provide reliable services for the entire network system. In the case of random or deterministic failures [4] in the network communication, the system is able to complete the key services.About the robust stability guarantees, the article [5] presented a new soft constrained model predictive control approach for tracking that provides stability guarantees even for unstable systems. The research of article [6] exploited the cubature rule in the unknown input observer structure to overcome nonlinear calculations in the presence of external disturbances for sensor fault detection purposes. In article [7], it is demonstrated that, under mild assumptions, the distributional robust optimization problems over Wasserstein balls can in fact be reformulated as finite convex programs---in many interesting cases even as tractable linear programs. A robust adaptive beamforming scheme based on two-component electromagnetic vector-sensor 4