I. Introduction
Distributed data processing has become a hot topic in recent years [1]–[3]. In this paper, we consider the estimation issue. Distributed estimation over networks is to estimate some parameters of interest from noisy measurements [4]–[7]. It performs the estimation task at each node, and exchanges information between immediate neighbors [8]–[11]. That is, it does not need a fusion center, which is indispensable for the centralized approach. Thus, distributed estimation is reliable and robust to node and/or link failure. Besides, the communication cost of distributed implementation is low. Due to these merits, distributed estimation has been applied to a wide range of fields, such as cognitive radios, environmental monitoring, and military surveillance. Many distributed algorithms have been developed for the distributed estimation problem [12]–[19].