An efficient closed-form solution for wideband source direction-of-arrival estimation
Metadata[+] Show full item record
[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] We propose and develop in this research an efficient closed-form solution to estimate the direction-of-arrivals (DOAs) for a wideband source signal, based on the data collected by a number of sensors. It has great potential for use in many applications including sonar, navigation and emergency rescue, which might require fast response and accuracy. A closed-form solution avoids the time-consuming grid search to obtain the DOA estimate, thus improving the efficiency and accuracy of the localization results. The proposed algorithm uses the two-stage least-squares minimization techniques and estimates the DOA using the time-difference-of-arrivals (TDOAs) of the source signal at different sensors. The algorithm is applicable for two-dimensional and three-dimensional DOA estimation, and it can be generalized to include the situation where the receiver positions have errors. In addition, this can work with a sensor array of arbitrary configurations. This is in contrast to most of the DOA techniques found in literature that require a linear sensor array configuration. We show theoretically that the performance of the proposed algorithm attains the Cramer-Rao lower bound (CRLB) accuracy. Simulations in MATLAB program are generated to compare the performance of the proposed algorithm with that of the maximum likelihood estimator (MLE) implemented by the iterative Taylor series linearization method.
Access is limited to the campuses of the University of Missouri.