1. Potential of Artificial Intelligence in Aviation/Aerospace Systems
Steve Beland, Panelist
SAE 2018 Aerospace Standards Summit, October 2, 2018
Flight Controls Associate Technical Fellow; Modeling & Simulation IPT; Authorized Representative (FAA) Product Strategy & Future Airplane Development, Boeing Commercial Airplanes

2. What we’re exploring now
Autonomous Taxi
Machine learning
Continuing to achieve safety goals
According to the Aviation Safety Network (ASN), 2016 was the second safest year on record.
2017 was even better! ZERO passenger fatalities!
October 2,

3. Existing & Growing Automation – Layered Loops
Contingency
Examples:
Surface:
Actuation Servos
Maneuver/Stability:
Attitude
Speed
Protection
Load Alleviation
Path Following:
Autoland
Route Planning:
Flight Mgt Function
Contingency:
Detect & Avoid
Situation decisions
Etc.
Route Planning
Path Following
Maneuver/Stability
Surfaces
Automation aids crews, reduces workload, augments stability or responding faster.
Complexity increases going outward (# of inputs, outputs, variations, validation effort)
Reaction time for failures is faster going inward
More loops could be added for fleets and airspace
October 2,

4. SAE S-18 Committee
Aircraft and Systems Development and Safety Assessment Committee
S-18
ADDITIONAL S-18 DOCUMENTS:
ARP5150 – Safety Assessment of Transport Airplanes in Commercial Service
ARP5151 – Safety Assessment of General Aviation Airplanes and Rotorcraft in Commercial Service
AIR6110 – Contiguous Aircraft/System Development Process Example
AIR6218 – Constructing Development Assurance Plan for Integrated Systems
AIR6219 – Incorporation of Atmospheric Neutron Single Event Effects Analysis into Safety Assessment (imminent)
AIR6276 – Use Of Modeling And Tools For Aircraft Systems Development (in work)
AIR6913 – Using STPA During Development and Safety Assessment of Civil Aircraft (in work)
S-18
Introduce S-18 ARPs. For todays systems and can help, but advanced safety methods are needed to accommodate autonomy and AI.
October 2,
SAE ARP4754A Figure 1

5. Challenges and Needs for Standards & Practices
Assuring Autonomous Aircraft Functions
Safety Assessment is itself an enabling technology and can be our guide:
Defines design space: safety requirements, design & operational constraints.
Standards and recognized practices as we validate system behavior:
To establish confidence for stakeholders & achieve today’s level of confidence
To guide completion criteria: see our targets and what “done” looks like
Specific topics:
Validation objectives of AI/ML behavior as they relate to assurance & safety
Degree of instrumentation of AI/ML functions (why did it do that?)
Commonality with automotive sector (similar challenges)
S-18 is sponsoring Aerospace Information Reports (AIRs) for emerging topics :
AIR6276 – Use Of Modeling And Tools For Aircraft Systems Development
AIR6913 – Use Of System-Theoretic Process Analysis (STPA) …..
Acknowledge value of ISO also
Validation stakeholders: developers, users, regulators, public
October 2,

6. Contact Info……
Contact:
Steve Beland, Associate Technical Fellow
Product Strategy & Future Airplane Development
Boeing Commercial Airplanes
Phone:
S18 Committee Website:
October 2,