Intellectus: Multi-hop fault detection methodology

Various sensor network studies have reported instances of faults in WSN. This paper focuses specifically on two types of faults, those caused by sensor node crashes resulting from power depletion or unreliable hardware and those caused by unpredictable link connectivity due to unstable wireless links. Intellectus [31] addresses these faults through the use of a group-detection mechanism; whereby neighbouring nodes monitor each other facilitating the detection of crash failures. Link instability leads to constant changing in the routing path. Intellectus nodes maintain state information about self and their neighbourhood and provide a monitoring capability for connectivity and topology change. This paper evaluates the Intellectus methodology through the use of a testbed; during the experiments two parameters, namely transmit power and the radio channel are changed at run time in order to create controlled faults and managed topology change. Intellectus is able to detect injected faults and recognize topology change. Intel-lectus information facilitates the construction of an abstract overview of the network topologies created during the experiments and helps us better understand the network as a whole. Spatial and temporal correlations from the local decision of every sensor nodes can provide a global overview of the topology network.