A Coherent Multi-User Frequency Shift Keying Receiver for Wireless Sensor Networks Based on Clustering Algorithms

Abstract

We explore the use of ${M}$ -ary frequency shift keying (FSK) as an integrated modulation/random access scheme for wireless sensor networks whereby multiple sensor nodes may transmit data simultaneously to a master node. Specifically, we propose exploiting the presence of multiple signal dimensions and using a clustering algorithm to estimate blindly the channel gains required for coherent multi-user detection. To demonstrate the feasibility of this concept, we simulate the bit error rates (BER) of 4FSK and 8FSK systems with three and four simultaneously transmitting sensor nodes, respectively. It was found that the BER of the proposed blind multi-user detector is very close to that achieved under perfect channel state information. This suggests that the proposed integrated modulation/random-access approach can significantly improve spectral efficiency and mitigate multi-user interference. Compared to no multi-user detection, less energy is wasted on the retranmissions caused by packet collisions and this helps to extend the lifetime of the sensor nodes.