We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.AP

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Applications

Title: Detection and Estimation of Multiple DoA Targets with Single Snapshot Measurements

Abstract: In this paper, we explore the problems of detecting the number of narrow-band, far-field targets and estimating their corresponding directions of arrivals (DoAs) from single snapshot measurements. We use the principles of sparse signal recovery (SSR) for detection and estimation of multiple targets. In the SSR framework, the DoA estimation problem is grid based and can be posed as the lasso optimization problem. The corresponding DoA detection problem reduces to estimating the optimal regularization parameter ($\tau$) of the lasso problem for achieving the required probability of correct detection ($P_c$). We propose finite sample and asymptotic test statistics for detecting the number of sources with the required $P_c$ at moderate to high signal to noise ratios. Once the number of sources are detected, or equivalently the optimal $\hat{\tau}$ is estimated, the corresponding DoAs can be estimated by solving the lasso with regularization parameter set to $\hat{\tau}$.
Comments: This version is contained in 1705.07561 and hence in redundant
Subjects: Applications (stat.AP); Information Theory (cs.IT); Probability (math.PR)
Cite as: arXiv:1609.00677 [stat.AP]
  (or arXiv:1609.00677v4 [stat.AP] for this version)

Submission history

From: Rakshith Jagannath [view email]
[v1] Fri, 2 Sep 2016 17:38:14 GMT (13kb)
[v2] Tue, 25 Apr 2017 14:47:58 GMT (15kb)
[v3] Fri, 8 Sep 2017 17:29:03 GMT (0kb,I)
[v4] Mon, 11 Sep 2017 08:07:17 GMT (0kb,I)

Link back to: arXiv, form interface, contact.