An application-level methodology to guide the design of intelligent-processing, power-aware passive RFIDs

We provide a methodology for reducing the power consumption in wireless sensor networks based on wireless active nodes (also implementing RFID reading ability) and sensor/processing-integrated passive RFIDs. Information acquired by the low cost, redundant, passive units is suitably compressed with an integrated sensor fusion algorithm based on neural networks, hence reducing the data to be transmitted to (and stored in) the reading unit. The methodology, by exploiting the application needs, also identifies the minimum number of bits to be transmitted to the reader (communication bit rate) as well as those required at each neuron output (information which immediately impacts on its complexity and electronics design) composing the sensor fusion algorithm. Simulation results applied to a comfort monitoring application in buildings show the effectiveness of the method with an energy saving factor of about 55% in the considered application.