1
Value of Information 1 st year review. UCLA 2012 Kickoff VOI Year 2 review. ARL, Sept 9 20 Kickoff 1 st year review. UCLA 2012 Value of Information Kickoff VOI Year 2 review. ARL, Sept 9 20 Information-geometric VoI in Sensor Planning Co-PIs: A. Hero and D. Cochran Motivating question: Can VoI with respect to a sensing objective be described geometrically via Rao’s differential geometric view of statistical manifolds? Relevance: The VoI of a sequence of sensor control actions with respect to an objective may be characterized by geometric properties of curves on a statistical manifold Collaborator: B. Sadler (ARL) Summary 1. Smooth adjustment of sensor parameters defines a curve in a statistical manifold Q 2. Information change can be quantified by measures on Q 3. Sensor planning can be carried out wrt geometry defined by informational objectives ran, A. Hero, ``Information-driven sensor planning,” GlobalSIP, Dec

Information-geometric VoI in Sensor Planning Co-PIs: A. Hero and D. Cochran

Embed Size (px)

DESCRIPTION

Information-geometric VoI in Sensor Planning Co-PIs: A. Hero and D. Cochran. Motivating question : Can VoI with respect to a sensing objective be described geometrically via Rao’s differential geometric view of statistical manifolds? - PowerPoint PPT Presentation

Citation preview

Page 1: Information-geometric  VoI  in Sensor Planning Co-PIs:  A. Hero  and  D. Cochran

Valueof

Information

1st year review. UCLA 2012

Kickoff

VOI

Year 2 review. ARL, Sept 9 2013

Kickoff

1st year review. UCLA 2012

Valueof

Information

Kickoff

VOI

Year 2 review. ARL, Sept 9 2013

Information-geometric VoI in Sensor PlanningCo-PIs: A. Hero and D. Cochran

• Motivating question: Can VoI with respect to a sensing objective be described geometrically via Rao’s differential geometric view of statistical manifolds?

• Relevance: The VoI of a sequence of sensor control actions with respect to an objective may be characterized by geometric properties of curves on a statistical manifold

• Collaborator: B. Sadler (ARL)

Summary1. Smooth adjustment of sensor

parameters defines a curve in a statistical manifold Q

2. Information change can be quantified by measures on Q

3. Sensor planning can be carried out wrt geometry defined by informational objectives

[1] D. Cochran, A. Hero, ``Information-driven sensor planning,” GlobalSIP, December 2013