Abstract
For many object tracking systems, how to quickly and efficiently estimate the direction of arrival (DOA) of radio waves impinging on the antenna array is an urgent task. In this paper, a new efficient DOA estimation approach based on the deep neural networks (DNN) is proposed, in which a nonlinear mapping that relates the outputs of the receiving antennas with its associated DOAs is learned by using the DNN-based network. The novel network architecture is divided into two phases, the detection phase and the DOA estimation phase. Additional detection network dramatically reduces the size of the training set and the process of the training data preparation is discussed in detail. After finishing the training phase, the corresponding DOAs can be identified based on current input data during testing phase. It has been shown that the proposed method can not only achieve reasonably high DOA estimation accuracy, but also reduce the computational complexity required by traditional superresolution DOA estimation methods such as multiple signal classification (MUSIC) and estimation of signal parameters via rotation invariance (ESPRIT). The computer simulation results are performed to investigate the generalization and effectiveness of the proposed approach in different scenarios.
Highlights
In modern electromagnetic research, there has been a growing research interest in the development of mobile communication devices [1]–[5]
SUPERVISED LEARNING POLICY OF THE PROPOSED ALGORITHM Unlike the existing methods which mainly rely on the array geometry, the proposed deep neural networks (DNN)-based framework uses a set of given samples of the input/output values to learn the relationship between the received signals and the direction of arrival (DOA) of radio waves as well as possible
In this paper, we have presented a new approach based on DNNs for detecting and estimating the DOAs of radio waves
Summary
There has been a growing research interest in the development of mobile communication devices [1]–[5]. Methods based on the use of support vector regression (SVR) [21], [22] and radial basis function (RBF) [26] have been efficiently applied for the problem of DOA estimation by establishing training data sets of the possible configurations of the impinging sources first, and deriving a mapping from array outputs to signal directions. A novel framework that performs both detection and DOA estimation, which is based on DNN, is presented and the design aspects are studied under a scenario where multiple uncorrelated narrowband signals arriving from different directions are incident on a uniform linear array. From the perspective of machine learning, the DOA estimation problem is dealt with as a nonlinear mapping that relates the outputs of the receiving antennas with its associated DOAs by means of the DNN-based network. Since the neural network does not deal with complex numbers directly, the each complex-valued entity of Rx is considered to be two dimensional real values except for the diagonal elements
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
