| Embedded Intelligent Real-Time Systems (EIRTS) are proliferating in many safety-critical large-scale environments, including aviation, ship or space shuttle control, air traffic control, nuclear power plant control, intelligent transportation, manufacturing, medical, and military and defense. In all of these systems, safety and reliability are important. During the last two decades, increasing attention has been paid to these safety-critical large scale systems, particularly because of the enormous damage to human life, environment, property, and society associated with such events as the Chernobyl nuclear power plant accident in 1986, the Therac-25 accidents between 1985–1987, and the Exxon Valdez oil spill in 1989.; EIRTS are often introduced in safety-critical large-scale systems to improve the systems' reliability and safety. In many cases, they also increase the system's complexity. EIRTS complexity Is a multi-dimensional subject, and has important impacts on software and operator performance, and on the performance of safety-critical large-scale system. In this thesis, the particular challenges associated with complexity in embedded intelligent real-time systems in safety-critical large-scale settings are studied. A multidisciplinary model of EIRTS complexity and 4 different complexities are proposed. The impacts of the proposed complexities on operator and software performance were investigated. The model and hypotheses were tested utilizing an operational EIRTS, the Navigation and Piloting Expert System, as a research vehicle. Understanding and measuring complexity in such systems can provide important insights about the design of future safety-critical large-scale systems, and can aid in designing and building more reliable and effective EIRTS, particularly in safety-critical settings. |
