Partially Binarized Deep MUSIC for Multiple Target Angle Estimation Using Wireless Sensor Array Systems

In this paper, a partially binarized deep learning-based MUltiple SIgnal Classification (MUSIC) algorithm for estimating the angle-of-arrivals (AoAs) of multiple targets using wireless sensor array systems is proposed. Since the sensor array system has limited computing power, it is not desirable to put the entire neural network on a single sensor node. Accordingly, the neural network was partitioned into two parts: the sensor node and the ground server. The neural network output (that is, the partially processed data) at the node was forwarded to the server through a noisy backhaul link channel. By modeling the noisy backhaul link as a binarized feedforward layer, we developed a new neural network architecture suitable for AoA estimation using wireless sensor array systems, and we trained it using a straight-through gradient estimator. Furthermore, unlike conventional deep learning-based MUSIC in which the MUSIC pseudospectrum for multiple targets is exploited as a label for training neural networks, a new training dataset generation method is proposed. Specifically, we generated the label by using the weighted sum of the MUSIC pseudospectra of each single target, resulting in more apparent peaks in the target angles and enhancing multiple target angle estimation accuracy.