Highly Adaptable Uncertainity Estimation Methodology for Sensor Fusion Systems

Period of Performance: 12/15/1998 - 06/15/1999

$120K

Phase 1 SBIR

Recipient Firm

Physical Optics Corp.
1845 West 205th Street Array
Torrance, CA 90501
Principal Investigator

Research Topics

Abstract

Physical Optics Corporation (POC) proposes to develop uncertainty estimation technology based on geometric data management (6DM). The proposed 6DM is a uniquely integrated data analysis model for state estimation and system self-calibration for adaptive uncertainty management. 6DM will be built with interconnecting logical sensors with three types of modules according to the given system structure: Feature Transformation Modules transform raw data into logical sensors; Data Fusion Modules fuse multiple sources of logical sensors and generate optimal features; and the Constraint Satisfaction Module represents system knowledge that imposes a constraint upon a set of feature values. This unique structure will have several advantages over error source investigation: continuous consistency and error monitoring; increased stability and robustness; maximum utilization of information; simultaneous performance of uncertainty update, error source identification, and error recovery; and improved computational efficiency for real-time estimates of uncertainties. In Phase I, POC will focus on identifying and analyzing the source of errors by building an appropriate model based on a current ALTAIR model with the proposed 6DM. Designing the 6DM structure will establish the basic parameters, and will isolate and identify the sources of errors. Computer simulation will be performed to evaluate 6DM real-time operating capability. BENEFITS: The proposed error analysis model will have numerous applications such as air traffic control, subway and railroad control, autonomous robotic vehicle navigation, intelligent security systems, and many other control applications.