Imaging SAS Performance Estimation
Abstract
At present there exists no complete method for quantitatively estimating overall system performance for automated underwater acoustic detection and classification systems being developed for use in mine countermeasure (MCM) operations. The lack of a capability for estimating or predicting performance will limit a systemÕs ability to adapt to changing environmental conditions or to attain higher levels of autonomy. A framework for performance estimation and prediction would allow pre?mission or in?mission decisions to be made that could maximize the probability of detection and classification by adapting operations based on the environmental constraints to performance (e.g., due to multipath interference, occlusion, etc.) or adjusting ATR operating parameters based on the calculated data quality and complexity. The overall goal for the proposed work is to establish the framework for linking the environment, sonar system, and signal processing to ATR detection and classification performance. We will work with two fundamental metrics, quality and complexity, as these seem to be currently supported by the consensus of the MCM research community. These metrics respectively describe the fidelity of sensor data and the environmental effects on ATR performance. To achieve our goal, we will relate data quality and complexity (i.e., the ÔsensedÕ seafloor complexity) to changes in ATR feature vector distributions and ultimately to performance via a loss in target/environment separability. Specifically, this program will develop quality and complexity metrics and then quantify the correspondence between these metrics and system performance through statistical (and modelbased, where appropriate) analysis of experimental data. This work will produce methods for performance estimation and prediction tools based on the quality of processed sensor output and environmental complexity as sensed by a given sonar system. External and prior information will be considered as well, but only to the extent that doing so is operationally feasible and materially enhances the result.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 10, 2016
- Source ID
- N000141612268
Entities
People
- Anthony P. Lyons
Organizations
- Office of Naval Research
- United States Navy
- University System of New Hampshire