Is this project an undergraduate, graduate, or faculty project?
Undergraduate
Project Type
group
Campus
Daytona Beach
Authors' Class Standing
Evan Bear, Senior
Lead Presenter's Name
Evan Bear
Lead Presenter's College
DB College of Aviation
Faculty Mentor Name
Dr. Vidhyashree Nagaraju
Abstract
Artificial intelligence and machine learning techniques are increasingly proposed for use in safety-critical civil aviation functions including perception decision support and pilot assistance. Existing aviation safety and certification standards such as ARP4754A and DO-178C were developed under assumptions of determinism explicit requirements and complete behavioral specification which do not directly apply to learning-enabled systems. This mismatch has created uncertainty regarding how artificial intelligence enabled avionics can be safely assured and certified. This paper presents a system safety approach for assuring artificial intelligence enabled functions within existing aviation certification frameworks. In this approach safety assurance is based on explicitly identifying the operational environmental and system assumptions under which safety claims remain valid. Rather than relying solely on verification of learned behavior artificial intelligence components are treated as bounded contributors within a system level safety architecture. Safety is ensured through the allocation of safety responsibility to deterministic and certifiable mechanisms including monitoring safety envelopes and fallback functions that maintain acceptable system behavior when assumptions are violated. The proposed approach is mapped to established aviation system safety processes including hazard analysis safety requirement allocation and architectural design and is consistent with principles reflected in SOTIF and assurance case based standards such as UL 4600. A representative aviation use case illustrates how safety responsibility can be partitioned between artificial intelligence enabled perception functions and independent safety mechanisms. The paper concludes with a discussion of certification and lifecycle considerations for artificial intelligence enabled avionics in safety critical systems.
Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?
No
Included in
Artificial Intelligence and Robotics Commons, Aviation Safety and Security Commons, Risk Analysis Commons
A System Safety Approach to Assuring Artificial Intelligence Enabled Functions in Civil Aviation
Artificial intelligence and machine learning techniques are increasingly proposed for use in safety-critical civil aviation functions including perception decision support and pilot assistance. Existing aviation safety and certification standards such as ARP4754A and DO-178C were developed under assumptions of determinism explicit requirements and complete behavioral specification which do not directly apply to learning-enabled systems. This mismatch has created uncertainty regarding how artificial intelligence enabled avionics can be safely assured and certified. This paper presents a system safety approach for assuring artificial intelligence enabled functions within existing aviation certification frameworks. In this approach safety assurance is based on explicitly identifying the operational environmental and system assumptions under which safety claims remain valid. Rather than relying solely on verification of learned behavior artificial intelligence components are treated as bounded contributors within a system level safety architecture. Safety is ensured through the allocation of safety responsibility to deterministic and certifiable mechanisms including monitoring safety envelopes and fallback functions that maintain acceptable system behavior when assumptions are violated. The proposed approach is mapped to established aviation system safety processes including hazard analysis safety requirement allocation and architectural design and is consistent with principles reflected in SOTIF and assurance case based standards such as UL 4600. A representative aviation use case illustrates how safety responsibility can be partitioned between artificial intelligence enabled perception functions and independent safety mechanisms. The paper concludes with a discussion of certification and lifecycle considerations for artificial intelligence enabled avionics in safety critical systems.