GE Aviation Systems (formerly Smiths Aerospace) is an American aerospace engineering, aircraft engine and aircraft parts manufacturer.Smiths Aerospace was formerly one of four business units of Smiths Group plc., an engineering company and constituent of the FTSE 100 share index. However, it was announced on January 15, 2007 that Smiths Group was divesting Smiths Aerospace to General Electric for US $4.8 billion. Smiths Aerospace, which was an important supplier, became an operating subsidiary of GE Aviation. This acquisition reportedly gives the combined unit the clout to resist pricing pressures from its two largest customers, Boeing Commercial Airplanes and Airbus. Analysts further assert that it will enable General Electric to acquire assets similar to those it desired in its failed bid for Honeywell in 2000. GE Aviation completed the transaction on May 4, 2007.
Artificial Intelligence/Machine Learning (AI/ML) techniques can be useful for autonomous or semi-autonomous operation of aircraft under diverse circumstances. Such safety-critical applications of AI/ML warrant adequate safety assurance before they are certified for operation. Moreover, because of the varying degrees of opacity and complexity in the design of non-trivial AI/ML-based aerospace applications, safety assurance must holistically consider the low-level behavioral and design elements of AI/ML-based components with respect to system-level hazards. However, existing certification standards and guidelines do not support systems utilizing AI/ML. We present an Overarching Properties (OP)-based approach for developing safety arguments that can be used to claim that an AI/ML-based component will be safe when used in the context of a more complex aircraft system. As a motivating use case for the study, we introduce an AI-Assisted Autonomous Runway Alignment (AARA) System that uses Artificial Neural Networks (ANN) to help align the aircraft nose to the runway center line after touchdown. We use the AARA to identify potential safety concerns and mitigate them through appropriate premises and assumptions in the safety arguments. The results show promise in the applicability of our OP-based approach for the holistic safety assurance of AI/ML-based complex aerospace systems.
This paper presents a formal language and framework, OYSTER, to develop correct-by-construction design of Integrated Modular Avionics (IMA). The OYSTER language is created as an annex to the Architecture Analysis and Design Language (AADL) for encoding constraints for aspects of IMA. The OYSTER constraints involve determining the correct locations of hosted applications within an IMA system, validity of the port connections involved in the design, and the conformance of virtual links allocated with bandwidth and jitter requirements. OYSTER also allows synthesis of communication paths for the allocated virtual links. The OYSTER prototype tool is developed as a plugin to the Open Source AADL Tool Environment (OSATE), and invokes Satisfiability Modulo Theories (SMT) solvers to synthesize correct-by-construction architecture designs for IMA. In addition, behaviors of applications running on IMA components and their safety properties can be modeled in the Assume Guarantee REasoning Environment (AGREE) annex and checked by the Kind2 model checker. The verification results are guaranteed to be correct by the independently verifiable proof certificates produced by Kind2. Finally, the paper evaluates OYSTER on a GE Aviation use case – a fuel control system IMA, and discusses the lessons learned.
The aerospace ecosystem is a complex system of systems comprising of many stakeholders in exchanging technical, design, development, certification, operational, and maintenance data across the different lifecycle stages of an aircraft from concept, engineering, manufacturing, operations, and maintenance to its disposal. Many standards have been developed to standardize and improve the effectiveness, efficiency, and security of the data transfer processes in the aerospace ecosystem. There are still challenges in data transfer due to the lack of standards in certain areas and lack of awareness and implementation of some standards. G-31 standards committee of SAE International has conducted a study on the available digital data standards in aircraft asset life cycle to understand the current and future landscapes of the needed digital data standards and identify gaps. This technical paper presents the study conducted by the G-31 technical committee. This paper reviews the data being exchanged between various stakeholders in the aerospace asset lifecycle and the availability of standards for the data transfer within the aerospace ecosystem. It identifies gaps based on the list of currently available data standards, and then creates a future landscape to address the needed digital data standards. This paper focuses on aircraft operations, maintenance, transfer, disposal processes, and post-build stage, and does not address the detailed interactions during the aircraft design, development and manufacturing phases. Its scope is also limited to key stakeholder interactions throughout the different stages of the aircraft operations, maintenance, and retirement.
This paper studies the use of touchscreen displays on the flight deck, focusing on the usability of touch interfaces to complete panning and numeric entry tasks. Results from this study show that the usability of a drag gesture to meet a pan function, surpasses the performance of a simple tapping interface under all turbulent conditions tested. Also, that this drag interface is more useable in an inter‐seat flight deck position, than a central one. It was also found that touchscreen interfaces can surpass the performance of the traditional mechanical dial interface for numeric data entry tasks, under all turbulent conditions tested.
Writing clear and unambiguous requirements that are conflict-free and complete is no easy task. Incorrect requirements lead to errors being introduced early in the design process. The longer the gap between error introduction and error discovery, the higher the cost associated with the error. To address the growing cost of system development, we introduce a tool called Analysis of Semantic Specifications and Efficient generation of Requirements-based Tests (ASSERT™) for capturing requirements, backed by a formal requirements analysis engine. ASSERT also automatically generates a complete set of requirements-based test cases. The requirements are captured in a structured natural language that is both human- and machine-readable. Formal analysis of these requirements with an automated theorem prover identifies errors as soon as requirements are written. It also addresses the historical problem that analysis engines are hard to use and understand for someone without formal methods expertise and analysis results are often difficult for the end-user to understand and make actionable. ASSERT’s major contribution is to bring powerful requirements capture and analysis capability to the domain of the end-user. We provide explainable and automated formal analysis, something we found important for a tool’s adoptability in industry. Automating test case generation in ASSERT also provides clear and measurable productivity gains in system development.
