An Approach to System Architecture Design Through Usability Heuristics

Abstract

This research explores the application of usability heuristics in the design of system architectures, addressing the current lack of clear guidelines in systems engineering. Systems engineering has evolved into a distinct scientific field, emphasizing the importance of robust and user-friendly system designs. The integration of human factors and usability principles is crucial for developing effective system architectures that can meet both technical and user needs. This study investigates how usability heuristics can be applied to enhance system architecture design.The methodology involved two primary usability testing approaches: timed performance assessments and A/B testing. The timed performance assessments evaluated two differently designed architectures of an autonomous vehicle's active safety features. Sixteen participants, selected based on their STEM background and basic design knowledge, were timed on their ability to locate and interpret information from the system architectures. The architectures were analyzed for node identification, data flow paths, and nested node structures. Additionally, qualitative feedback was gathered to understand user satisfaction and perceived usability.A/B testing was conducted on an aircraft avionics system architecture, comparing a patented design with a newly designed configuration. Participants provided qualitative feedback on their understanding, satisfaction, and perceived complexity of each architecture. The aim was to identify specific design elements that contribute to better usability and to develop a set of heuristics tailored for system architecture design.The analysis revealed significant differences in usability between the two architectures tested in the timed performance assessments. Participants identified elements faster in the redesigned architecture with fewer line intersections and clearer node organization. The results highlighted the importance of minimizing line intersections, using straight lines instead of diagonal ones, and providing clear labels for all elements. In A/B testing, the newly designed architecture received higher satisfaction ratings and was easier for participants to understand, emphasizing the value of organized node grouping and clear flow of information.Based on the findings, a refined set of usability heuristics for system architecture design was developed. These heuristics include minimizing the intersection of arrows, creating straight lines, labeling and identifying all elements, ensuring a clear information flow, designing with enough space to reduce noise and promote flexibility, and attaching a key for all abbreviations and symbols. These guidelines aim to help system architects create more intuitive and efficient designs.The study concludes that applying usability heuristics can significantly improve the design of system architectures, leading to better performance, user satisfaction, and overall business success. The research underscores the need for further studies involving a larger and more diverse participant pool to validate and expand the proposed heuristics. By integrating human factors research into systems engineering, businesses can develop more robust and user-friendly products that meet the demands of both technical and non-technical stakeholders.