Self-diagnosis for large scale wireless sensor networks

Abstract

Abstract-Existing approaches to diagnosing sensor networks are generally sink-based, which rely on actively pulling state information from all sensor nodes so as to conduct centralized analysis. However, the sink-based diagnosis tools incur huge communication overhead to the traffic sensitive sensor networks. Also, due to the unreliable wireless communications, sink often obtains incomplete and sometimes suspicious information, leading to highly inaccurate judgments. Even worse, we observe that it is always more difficult to obtain state information from the problematic or critical regions. To address the above issues, we present the concept of self-diagnosis, which encourages each single sensor to join the fault decision process. We design a series of novel fault detectors through which multiple nodes can cooperate with each other in a diagnosis task. The fault detectors encode the diagnosis process to state transitions. Each sensor can participate in the fault diagnosis by transiting the detector's current state to a new one based on local evidences and then pass the fault detector to other nodes. Having sufficient evidences, the fault detector achieves the Accept state and outputs the final diagnosis report. We examine the performance of our self-diagnosis tool called TinyD2 on a 100 nodes testbed.

Assessing Diagnosis Approaches for Wireless Sensor Networks: Concepts and Analysis

“…One is out-network information level, we compare two typical diagnostic tools, called PAD [11] and Powertracing [19]. The other is in-network information level, we compare Sympathy [10] and TinyD2 [18] on this level. Since the diagnostic tools of global information level usually contain code level debugging techniques, we cannot employ this change after the network is established.…”

“…PAD [11] reports the concept of passive diagnosis which leverages a packet marking strategy to derive network state and deduces the faults with a probabilistic inference model. TinyD2 [18] is a self-diagnosis tool, which combine the view of the node itself to the diagnosis process.…”

An Estimation Algorithm for Missing Data in Wireless Sensor Networks

“…For simple calculation, it is assumed that ΔT can be divisible by Tth and the error process is a second-order moment, that is, expectation and variance functions exist in the event procedure. Sensor networks are commonly used to detect specific errors, so we can assume that the expectation function Expe(t) and the square root of variance function VarSRe(t) can be stored in the sensor memory before deployment or the Expe(t) and VarSRe(t) can be distributed to various sensors through the meeting points (sink node) by message after deployment [17][18][19][20][21][22][23].…”

Sherlock is around: Detecting network failures with local evidence fusion

“…In [1], the authors use monitoring paths and cycles to localize single link and Shared Risk Link Group failures. Self-diagnosis [11] plants a finite state machine into each sensor node, enabling them to accordingly change the diagnosis state. Nevertheless, it proves difficult to achieve consensus between the nodes as each node can change the state whenever it receives the diagnosis requests.…”

“…These approaches, however, may suffer from the narrow scope of single node. Some of the selfdiagnosis methods like TinyD2 propose to let fault detectors travel among different sensors [11]. Such a method, however, does not well handle the diverse judgements from different nodes so that it leads to inconsistent diagnosis reports at different network regions.…”