Direct target localization and deghosting in active radar network

Recently proposed, the one step localization approach for active sensor network estimates the target location directly from the sensor raw signals, as opposed to the classical decentralized two steps approach where intermediate signal processing (e.g., measurement estimation, detection) is performed on each sensor node. A composite multiple hypothesis test based on the generalized maximum likelihood (GML) rule is proposed to cope with the sidelobes of the one step cost function that may cause detection and localization of ghost targets. The proposed method has been verified both on synthetic and real data.