Probabilistic Approach to Provisioning Guaranteed QoS for Distributed Event Detection

Abstract

Abstract-It has been of significant importance to provision network-wide guaranteed QoS for a wide range of event detection applications in wireless sensor networks (WSNs). This paper investigates solutions to this QoS provision problem. For event detection applications, there are two key performance metrics, i.e., detection probability and detection latency. This paper focuses on dual-objective QoS provision, taking both metrics into account. This is very challenging due to the stringent resource constraint of sensor nodes and unpredictable randomness of physical events. We propose a novel probabilistic approach to provisioning due-objective QoS. Following a unified framework, we design a distributed algorithm that determines the active probability of every sensor node. The probability is minimized while being sufficient for QoS provision. Our approach is flexible and supports different requirements that may be posed by different applications. Theoretical analysis and comprehensive simulation experiments have been conducted, which jointly demonstrate that our approach is able to deliver guaranteed QoS for distributed event detection while prolonging the system lifetime significantly compared with other alternative schemes.

QoS Support in Event Detection in WSN through Optimal k-Coverage

“…Zhu and Ni [13] first formalized the QoS provisioning problem in event detection applications in WSNs. They presented detection latency and detection probability as the two key performance metrics for event detection systems and proposed a probabilistic approach to provisioning QoS.…”

Event detection and channel allocation in cognitive radio sensor networks

“…[1] focuses on accurate event detection while also catering to QoS constraints. [2] analyzes event detection delay in WSNs.…”

Risk-based adaptive scheduling in randomly deployed video sensor networks for critical surveillance applications

“…The next thing to consider is that surveillance applications have very specific needs due to their inherently critical nature associated to security (He et al, 2004;Oh et al, 2006;Dousse et al, 2006;Zhu and Ni, 2008;Freitas et al, 2009). Early surveillance applications involving WSN have been applied to critical infrastructures such as production systems or oil/water pipeline systems (Stoianov et al, 2007;Albano et al, 2009).…”

“…Many contributions have been made in the last few years and the authors in Gupta et al (2003), Wang et al (2003) and Shakkottai et al (2003), to name a few, have proposed interesting energyefficient approaches that aim at providing the highest detection quality. However, it is also desirable to have differentiated surveillance services for various target areas with different degrees of security requirements (Yan et al, 2003) or to be able to probabilistically support flexible QoS (Zhu and Ni, 2007) without over-provisioning resources. In surveillance applications such as an intrusion detection system having multiple levels of activity is essential because such systems have to be able to operate on a long term basis as no one knows when such intrusions could occur.…”