Reliability Analysis for Small Unmanned Air Vehicle with Algorithmic Redundancy

Abstract

Integrating Unmanned Aerial Systems (UAS) into the US National Airspace (NAS) is necessary for advancement of UAS applications in different fields like transport, surveillance, disaster management, and geospatial applications. Airworthiness certification by the Federal Aviation Administration (FAA) is an important step towards NAS integration. To match the expected FAA safety and reliability standards for UAS certification, novel technologies are needed. Conventional manned air vehicles meet FAA standards via hardware redundancy, which is not feasible for small UAS due to size, weight, and power constraints. This challenge is addressed in this paper via the so-called algorithmically redundant approach, where multiple fault detection and isolation (FDI) algorithms work in parallel to detect faults more reliably. Reliability of algorithmically redundant systems is computed using well-understood procedures like fault tree analysis or failure modes and effects analysis. This approach is also extended to evaluate a hybrid redundant design featuring flight control system hardware, control laws in the loop and multiple FDI algorithms. Extensive simulations are carried out to validate the reliability estimates and simulation model itself is validated via flight testing. This step-by-step approach for overall reliability estimation based on theoretical analysis, simulation runs, and flight tests, can provide a path forward toward certification.