Adaptive Analog Function Computation via Fading Multiple-Access Channels

Abstract

In this letter, we propose an adaptive analog function computation (AFC) via fading multiple-access channels in which multiple sensors simultaneously send their observations and then the fusion center computes the desired function via the superposition property of wireless channels. In particular, each sensor adaptively sends its observation to the fusion center based on its causal channel state information (CSI). Numerical results show that the adaptive AFC significantly outperforms the conventional non-adaptive AFC in terms of the outage probability of function estimation error. The adaptive AFC operates in a fully distributed manner with local and causal CSI, applicable to various practical sensor network applications.