In this paper, we introduce a family of spatio-temporal Gaussian processes specified by a class of covariance functions. Nonparametric prediction based on truncated observations is proposed for mobile sensor networks with limited memory and computational power. We show that there is a trade-off between precision and efficiency when prediction based on truncated observations is used. Next, we propose both centralized and distributed navigation strategies for mobile sensor networks to move in order to reduce prediction error variances at positions of interest. Simulation results demonstrate the effectiveness of the proposed schemes.

This content is only available via PDF.
You do not currently have access to this content.