Source reconstruction via mobile agents in sensor networks: throughput-distortion characteristics

We consider the problem of reconstructing a signal field measured by a large scale sensor network with mobile agents. Sensors transmit packets containing measurement data to mobile agents using either random or deterministic medium access control (MAC) schemes, and the signal field is reconstructed by mobile agents that minimize the mean square error of the reconstruction. For the one-dimensional Gauss-Markov field, we investigate the relation between the system throughput and reconstruction distortion, for different types of MAC schemes. We show that at low throughput level, increasing system throughput decreases the reconstruction distortion considerably. But the improvement is much less when the throughput is relatively high. We also show that the choice of MAC schemes can affect the reconstruction performance significantly, especially when the measurement noise is low.