A cluster‐based framework for interface analysis in large‐scale aerospace systems

Abstract

This paper proposes a framework using a community maturity level metric to determine the integration readiness of interface elements in a particular network cluster. As a proof of concept this methodology is applied to an aircraft seating system to assess the readiness of complex interfaces before proceeding to full-scale production and systems integration. A multi-objective genetic algorithm, MOGA-Net, is coupled with the Newman-Girvan modularity metric as a clustering algorithm. This algorithm identifies system elements grouped by common interfaces, referred to as community clusters. The TRL and IRL values for these elements is then used to calculate an overall community maturity level. The achieved performance in these clusters is then compared to the target performance to determine overall maturity of the interfaces. This is compared to other system readiness metrics and interface readiness metrics as applied to the aircraft seating system and was found to be more consistent with subject matter expert evaluations during the critical design review. This gives a better representation of the true readiness of system interfaces before entering design reviews to reduce overall integration risks.

Emerging Technology Certification Risk Assessment with ETHICIST

“…Given its weaknesses, several replacements for and extensions to TRL have been proposed for assessing technology in context. For example, "Integration Readiness Levels" for considering different architectural views of a system (Jesus and Chagas Junior, 2022), "Human Readiness Levels" considering the readiness of technology to human operators (Salazar and Russi-Vigoya, 2021), and "Community Maturity Levels" that result from a cluster analysis of communities of system components (Goldschmid and Corns, 2021). However, integration and community maturity assume at least some basic level of system maturity, and human-readiness considers only operators directly using some technology not the broader community of stakeholders, e.g.…”

Machine learning approaches for improving integration of advanced sensors on legacy aircraft

“…Several researchers have addressed integration interface issues. To address aircraft interface complexity, Goldschmid and Corns used a genetic algorithm to group interfaces into clusters 13 for integration readiness level determination based on metrics defined by Sauser et al 14 Morgan, Holzer, and Eveleigh addressed rapid component replacement by requiring interface definition as part of a container modeling method (CMM). 15 Abdolshah et al proposed model-based reinforcement learning, that could be used to plug and play object-oriented software objects within a software application.…”