Estimation of Nominal Direction of Arrival and Angular Spread Using an Array of Sensors
1996 (English)In: Signal Processing, ISSN 0165-1684, E-ISSN 1872-7557, Vol. 50, no 1-2, 57-69 p.Article in journal (Refereed) Published
The problem of estimating the nominal direction of arrival and angular spread of a source surrounded by a large number of local scatterers using an array of sensors is addressed. This type of propagation occurs frequently in, for example, mobile communications. The maximum likelihood estimator is considered and two computationally less complex estimators are also proposed. They are based on least-squares fits of the sample covariance to the theoretical covariance matrix derived from the measurement model. The performance of the least-squares-based algorithm is analyzed and based on this, an optimally weighted least-squares criterion is proposed. The weighted least-squares algorithm is shown to be asymptotically efficient. Results of numerical experiments are presented to indicate small sample behavior of the estimators. The nominal direction-of-arrival (DOA) estimates are compared with those provided by a standard subspace algorithm. Finally, the methods are applied to experimental array data to determine spread angles for non line of sight scenarios.
Place, publisher, year, edition, pages
1996. Vol. 50, no 1-2, 57-69 p.
array of sensors, nominal direction, arrival and angular spread, covariance matrix, least-squares criterion
IdentifiersURN: urn:nbn:se:kth:diva-55821DOI: 10.1016/0165-1684(96)00003-5ISI: A1996UR76300006OAI: oai:DiVA.org:kth-55821DiVA: diva2:471787
QC 201201042012-01-032012-01-032012-01-04Bibliographically approved