I-Corps: Commercialization of Feedback Particle Filter for Target State Estimation

University Of Illinois At Urbana-Champaign, Urbana IL

Investigators

Abstract

The goal of the project is to further develop a commercialization a nonlinear estimation platform technology for applications involving target state estimation. Nonlinear estimation is a widely pervasive enabling technology. Some current algorithms either require specific mathematical characteristics of the underlying model or suffer from robustness issues. In applications involving target state estimation (target tracking), these approaches can lead to poor terminal accuracy and sub-optimal tracking performance. The innovation of the proposed project lies in an algorithm for nonlinear estimation that may provide more accurate results with less computational overhead. In target state estimation applications, existing solutions have difficulty tracking targets that exhibit nonlinear and uncertain motion. These algorithms are successfully able to address these problems. Increasing accuracy and speed for nonlinear estimation at lower computational cost could benefit many applications and industries. The National Defense would benefit from increased radar, laser, and satellite tracking accuracy. Financial firms could make better derivatives models and trading decisions. Applied to robotic vision, autonomous vehicles could navigate more accurately and assembly robots could spot more defects. Software tools based on these algorithms have a potential to impact a number of application areas both in military and civilian domain: Application areas of interest in the military domain include Air Moving Target Indicator (AMTI) and Ground Moving Target Indicator (GMTI) systems deployed in Intelligence, Surveillance, and Reconnaissance (ISR) operations. Civilian applications include air traffic surveillance, weather surveillance, ground mapping, geophysical surveys, remote sensing, autonomous navigation and robotics.