Neighbor discovery in wireless networks: A multiuser-detection approach

We consider the problem of determining, in a distributed and scalable manner, which nodes are neighbors in a wireless sensor network. We consider an un-supervised network operating on a frequency-flat Gaussian multipleaccess channel, shared by K + 1 nodes, transmitting synchronously and independently a set of known signal waveforms. A number of ND algorithms, relying on different optimization criteria, are introduced. A thorough statistical analysis for the proposed ND algorithms is presented, and the results demonstrate that a number of reduced complexity receivers exist achieving satisfactory performances.