An Alternate Data Gathering Method Based on Sparse Random Projections

Due to the fact that each sensor node of wireless sensor networks (WSNs) is usually powered by batteries which cannot be recharged in most cases, there exist a contradiction between the limited energy supply and the demand of a long network lifetime. The emerging compressed sensing (CS) theory provides a promising scheme to alleviate this contradiction. The existing methods based on both CS and sparse random projections may save energy to some extent, they still result in some unnecessary energy consumption. In order to further save the energy, this paper proposes an alternate data gathering method based on sparse random projections. In the proposed method, we adopt an alternate strategy, that is, the sensor nodes are randomly divided into two groups, one group monitors the data and the other keeps silence in a period, and they switch the duty in the next period. In addition, the ill-posed problem of signal recovery may be solved because monitoring data may be viewed as compressible signal in general. So the proposed method can save the energy and guarantee the balanced load. Experimental results show that the proposed method consumes less energy to obtain almost same reconstruction quality of the signal, compared with the existing methods based on random projections.