A stochastic approach to resolution based on information distances computed from the geometry of data models which is characterized by the Fisher information is explored. Stochastic resolution includes probability of resolution and signal-to-noise ratio (SNR). The probability of resolution is assessed from a hypothesis test by exploiting information distances in a likelihood ratio. Taking SNR into account is especially relevant in compressive sensing (CS) due to its fewer measurements. Based on this information-geometry approach, we demonstrate the stochastic resolution analysis in test cases from array processing. In addition, we also compare our stochastic resolution bounds with the actual resolution obtained numerically from sparse signal processing which nowadays is a major component of the back end of any CS sensor. Results demonstrate the suitability of the proposed stochastic resolution analysis due to its ability to include crucial features in the resolution performance guarantees: array configuration or sensor design, SNR, separation and probability of resolution.
|Title of host publication||2017 IEEE 7th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP)|
|Place of Publication||Piscataway, NJ|
|Number of pages||5|
|Publication status||Published - 2018|
|Event||2017 IEEE 7th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing - Willemstad, Curaçao|
Duration: 10 Dec 2017 → 13 Dec 2017
Conference number: 7
|Workshop||2017 IEEE 7th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing|
|Period||10/12/17 → 13/12/17|
Bibliographical noteGreen Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.
- array processing
- compressive sensing
- information geometry
- likelihood ratio