Flight critical systems research (fcsr) statement of work (SOW)task order title: regulatory considerations for adaptive SYSTEMS1.0INTRODUCTION / background:adaptive systems, such as adaptive controls,…

Flight critical systems research (fcsr) statement of work (SOW)task order title: regulatory considerations for adaptive SYSTEMS1.0INTRODUCTION / background:adaptive systems, such as adaptive controls, have proven beneficial in some application domains in maintaining needed performance in the presence of uncertainty and in accommodating unanticipated environmental conditions and sensor input. The benefits afforded by adaptive technologies make them attractive for use in systems planned for the next generation air transportation system (nextgen) and the national airspace system (NAS) air traffic control (ATC) system. NASA langley research center (larc) is working with the federal aviation administration (FAA) to ensure that potentially-planned adaptive approaches for nextgen, NAS ATC, and other safety-critical aircraft systems, comply with applicable safety regulations. The FAA and larc are jointly investigating the verification aspects of adaptive systems and identifying the differences between the current approach used for airborne software approval and the planned use of adaptive systems. 2.0 scope and objective(s): the scope of this task order is to examine the verification of adaptive systems and the means by which to ensure compliance of the software to rtca/do-178BF f, identify areas where compliance cannot be achieved, and propose alternative methods of compliance as necessary. This task will also review past results and current plans, as available, by the FAA software and digital systems (SDS) technical community representatives group (tcrg) as related to deterministic and non-deterministic research. In addition, this task will also examine the use of tools to validate and verify adaptive systems. For this study, adaptive systems are considered to be systems that have the ability to change behavior at run-time in response to changes in their operational environment, system configuration, resource availability or other factors, and would include learning systems. Since the primary focus is on systems subject to do-178B, piloted systems (including UAS) are the primary interest at this time. However, there are nextgen plans that propose adaptive systems for air traffic management applications that should also be considered in the study. This study is not expected to focus on one type of adaptive system (or application) before another, but consider the broad range of possible applications. Examples of envisioned adaptive nextgen systems include algorithms that provide for dynamically reconfiguring airspace boundaries to meet prevailing demand and optimize throughput; adaptive weather prediction; adaptive speed control to improve traffic flow management; and autonomous landing of a damaged UAS. Consideration of system development, safety assessment, and safety assurance processes that pertain to do-178B are also appropriate for this study, as are system-level considerations that might be necessary to provide assurance of an adaptive system, including identification of issues and obstacles that adaptive systems pose to these processes. The objective of this task order is to provide NASA with a recommended approach for assuring flight software that uses techniques such as adaptive control and machine learning to modify its behavior during operation. Do-178B and related guidelines have an underlying assumption that all aircraft systems with their software and its assurance artifacts are available for scrutiny at one time, and that the behavior of the software is fully determined at that time. Adaptive software that is, software that adjusts its behavior during operation violates this assumption and therefore likely requires some revision in the current approach to assurance for flight software. 3.0 description of the work/tasks to be performed: NASA is currently documenting preliminary background information relative to terminology definitions (e.g., adaptive system, neural net, adaptive software, a