Degradable Hardware MM Estimators for Resilient USV Search and Surveillance

Abstract

An Unmanned Surface Vehicle (USV) operates in an environment of constrained resources carrying enough power and computing to complete its mission. A cyber attack may primarily aim to degrade the USV’s computational devises and, ultimately, incapacitate its sensing systems. To build resilience against such casualties, this research investigates a tracking system and proposes hardware implementation of a Multiple Model estimator on several FPGA units with the ability of dynamic partial reconfiguration so that the most likely models of the estimator can be activated in blocks of the surviving FPGAs.