Certification and Safety of In-Flight Multi-Objective Decision Making Algorithm Techniques

The Advanced Air Mobility (AAM) marketplace has seen rapid growth in recent years with hundreds of possible designs at varied stages of development. Many manufacturers are approaching the design problem under the presumption that each of their vehicles will include a pilot onboard. According to an AP News article from February 2023, the government estimates a shortage of ~30k airline pilots by 2032. Autonomous operation of these vehicles could help mitigate/eliminate the piloted need but only a small subset of designers is considering this as their entry point. The necessary autonomous technologies are currently not mature enough to meet this challenge and will be met with intense regulatory scrutiny. Without the ability to take in information from multiple sources and distill that to an actionable response, these vehicles will be limited in their operation scope. Alternatively, without a clear path to certify these new capabilities, the realization of an autonomous UAS/AAM system and its potential benefits will languish.nbsp;As such, a team led by Systems Technology Inc. (STI), propose to address this problem through the development of Certification and Safety of In-Flight Multi-Objective Decision Making Algorithm Techniques (CERTIFICATE). CERTIFICATE will establish a robustnbsp;certification approach and companion software-based toolset for multi-objective decision making (MODM) autonomous technologies and demonstrate this process with novel envelope protection (EP) schemes based upon neural control certificates and correct-by-construction software paradigms. Phase I will establish the system requirements and demonstrate the CERTIFICATE process using these novel techniques. At the conclusion of the Phase II program, a prototype of the CERTIFICATE software will have been developed and validated for supporting the certification of MODM systems. In addition, two MODM EP methods will have been advanced towards addressing increasingly complex and robust mission scenarios.