Experimental Comparison of Dynamic Spectrum Access Techniques for Wireless Sensor Networks

Communication schemes exploiting dynamic spectrum access have been recently envisaged in the context of wireless sensor networks: these access techniques allow cognitive sensor devices to avoid transmissions of interfering networks by opportunistically exploiting spectrum holes in the time or frequency domain. Such spectrum opportunities need to be identified through spectrum sensing, however this might introduce significant energy overhead. Selecting the most appropriate spectrum sensing and access technique is therefore a key issue for the efficient design of these communication schemes. In this paper we present the experimental evaluation of two different cognitive approaches, respectively exploiting spectrum holes in the time and frequency domain. We quantify the average energy required by a two-node system for the successful delivery of a packet and investigate the effect of the energy overhead introduced by spectrum sensing. Our results show that the time domain approach results in poor performance even for relatively low values of channel occupancy; on the other hand channel surfing schemes introduce significant energy overhead due to spectrum sensing but allow to effectively avoid interference and should therefore be preferred to time-domain techniques when the risk of operating in heavily interfered channels is high or when a large amount of data has to be transmitted.