Accession Number : ADA305289

Title :   Contact Management Model Assessment: An Approach and System Description.

Descriptive Note : Final rept.,

Corporate Author : NAVAL UNDERSEA WARFARE CENTER NEWPORT DIV RI

Personal Author(s) : Nardone, S. C. ; Ferkinhoff, D. J. ; Hammel, S. E. ; Gong, K. F.

PDF Url : ADA305289

Report Date : 10 APR 1995

Pagination or Media Count : 69

Abstract : Contact Management Model Assessment (CM MA) is a problem of state estimation in the presence of modeling ambiguity and uncertainty. In the general context of CM MA, noise corrupted measurements from a contact are processed to obtain estimates of the contact's position and velocity. The process by which the contact state is estimated typically requires a model that relates contact dynamics, the sensor, and the environment to the measurements. Classical hypothesis testing methodologies are applicable to CMMA but have the adverse effect of requiring multiple state estimates; some of which may experience convergence difficulties. The approach taken here uses DempsterShafer (Ds) evidential reasoning to pre. select, for further processing. only those model hypotheses that are consistent with the observed data. In the DS approach. evidential reasoning takes place in the frame of discernment; a set of mutually exclusive and exhaustive hypotheses. Evidence, leading to belief in the subsets of the frame of discernment, is mapped into basic probability assignment over the power set of the frame. The evidence, or belief, from multiple sources that are related to the same frame of discernment can be combined with Dempster's combination rule. Dissimilar but related frames can be mapped to a common frame of discernment with compatibility relations. The collection of frames and compatibility relations is referred to as a gallery. This report details the CMMA gallery and provides a system description. An overview of DS evidential reasoning is presented along with the specific CMMA application. Functions that map evidence from statistical inferences on predicted residual sequences to belief in model hypotheses are developed. Additionally, a method for hypotheses selection from the DS frame is developed.

Descriptors :   *SIGNAL PROCESSING, *ACOUSTIC TRACKING, *UNDERWATER TRACKING, MATHEMATICAL MODELS, ALGORITHMS, COMPUTERIZED SIMULATION, UNCERTAINTY, MANAGEMENT PLANNING AND CONTROL, TOWED ARRAYS, MAXIMUM LIKELIHOOD ESTIMATION, PROBABILITY DISTRIBUTION FUNCTIONS, STATISTICAL INFERENCE, REASONING, SIGNAL TO NOISE RATIO, BEARING(DIRECTION), COMPATIBILITY, FALSE ALARMS, PROBABILITY DENSITY FUNCTIONS, DRIFT, ERROR ANALYSIS, CONVERGENCE, PERTURBATIONS, HYPOTHESES, AMPLITUDE, AZIMUTH, LITTORAL ZONES, OCEAN ENVIRONMENTS, ACOUSTIC DATA, SONAR SIGNALS, ACOUSTIC DETECTORS, TIME LAG THEORY, MOVING TARGET INDICATORS, SONAR ARRAYS.

Subject Categories : Acoustic Detection and Detectors
      Operations Research

Distribution Statement : APPROVED FOR PUBLIC RELEASE