On environment-aware channel estimation for wireless sensor networks using Magnetic Induction

The Magnetic Induction (MI)-based communication techniques enable the applications of wireless sensor networks (WSNs) in complex environments, such as underground and underwater environments. However, the complex environments usually consist of many conductive objects that can significantly enhance or block the MI signals. Efficient communications can be achieved if we can smartly utilize the positive factors and prevent the channel from the negative factors by selecting the optimal operating frequency and routing relays in the networks. However, this strategy cannot be realized if the environment is unknown to us. To address such a problem, we propose an environment-aware method for WSNs applications in the complex environments. In this paper, a channel model is first developed by considering the complex environments and the positive/negative factors are analyzed based on the model. Then, based on the training data obtained by the handshake between transceivers, the Kernel method is used to learn the positive/negative factors in the environment. Finally, an example of environment-aware routing protocol is presented to show that the proposed environment-aware method can be used to significantly improve the efficiency of the networks by optimally selecting the channel and the routing relays.