Abstract:

According to the B-cell and T-cell model of immune systems, the similarity between artificial immune system and wireless sensor networks is compared. A neighborhood node selection algorithm is proposed to decide if a sensor node is activated or not. Considering the feature of event estimation distortion at the sink node, the relations between distortion and number of activated sensor nodes as well as distortion and reporting frequency are constructed by the spatial-temporal correlation and adaptive least mean square filter. The minimum number of sensor nodes and the best reporting frequency are achieved. The simulation results in different conditions show that the optimization strategy for wireless sensor networks has the capability of reducing the number of sensor nodes, decreasing the reporting frequency and saving energy consumption.