Date of Graduation
Statler College of Engineering and Mineral Resources
Lane Department of Computer Science and Electrical Engineering
This problem report presents the results of an anonymous online survey that was used to collect information about software systems that use model-based software engineering, contain autonomy and where software reuse plays an important role. Recent advancements in computation ability and the emergence of decision-making algorithms have increased interest in and use of autonomy in applications areas such as aeronautics, automotive, military, and space industries. In these application domains, autonomy is used to reduce costs, reduce reaction times, and improve performance. Due to the emerging nature of autonomy, very little research has been done regarding the level of autonomy of autonomous systems. In addition, developing software systems involving autonomy is particularly challenging if mission’s success or human life depend on it. Also, there is limited information about autonomous systems due to the proprietary nature of the code implementing autonomy.
The main goals of our survey were to: (1) assess the current state-of-the-practice using autonomous systems, (2) identify and quantify the benefits and challenges of autonomy and reuse, (3) explore the processes and standards used to develop autonomous systems, and (4) investigate verification and validation of the models, autonomy, and reuse.
The results of the survey are based on the answers provided by 110 respondents who used model-based software engineering and/or autonomy. The main findings, as they pertain to our goals, are as follows:
Significant percentage of autonomous components used full autonomy. Most common types of reuse were code reuse, model reuse, and data reuse.
2. Benefits and challenges.
The prevalent challenges associated with autonomy were due to high level of environment uncertainty, system complexity, and achieving desired level of autonomy. In projects with reuse, 63% of respondents reported increase productivity, 40% reported increased quality, and 37% observed decreased cost due to reuse.
3. Processes and standards.
48% of respondents did not use modeling standards for their project. Only 24% of the projects went through a certification process.
4. Verification and validation.
Only 11% of the respondents verified and validated security. During the development phase the autonomous components were verified and validated using simulation based testing (20%), unit and integration testing (19%), and manual model/code reviews or inspections (18%). Only 19% of respondents verified and validated reused artifacts. 28% of respondents found bugs due the autonomous functionality, 25% found bugs due to reused artifacts, and 9% due to the use model-based approach.
Hood, Denny LaVerne III, "Autonomous Systems with Reuse: A Survey on the State-of-the-Practice" (2019). Graduate Theses, Dissertations, and Problem Reports. 7391.
Available for download on Saturday, December 05, 2020