Model-driven design of self-observing products

Current complex electronic products are designed and developed by multi-disciplinary teams of stakeholders, and they are sold to customers world-wide who are often overwhelmed with the complexity of products. Product creation processes lack information about product usage and user experience in the field. This information deficit clearly has a negative effect on current and future product creation, but also on the entire business development. Prior approaches to acquire usage data via logging, extensive prototype evaluation, and market research studies, miss benefits of connected and rich information. In this thesis, new technology is researched and developed to bridge this gap and to reveal actual customer needs and expectations. This thesis first presents a theoretical approach to remote data collection that begins with "observables" within complex electronic products. It is shown how high-quality information about usage and context of products in the field can be obtained from these observables; precisely the information that is meaningful to stakeholders. It is described how additional local processing and semantics have to be applied in a stakeholder-friendly way such that desired information can be extracted. Finally, data quality can be further improved by jointly collecting objective and subjective data, that is, what users do and why they do it, in a structured way. Building on the previously described theoretical approach to remote data collection we pursued a generic framework that illustrates how operational observation systems can be created, including a dedicated development process and a run-time system architecture. This enables low-impact, real-time data collection on the one hand, and, on the other hand, allows stakeholders to visually specify observation and data analysis with a graphical language. The extensive body of knowledge about the domain of observation has also been captured in models, design patterns, and ontologies that allows for reuse and extension of observation concepts and tools for other application domains. The research project is embedded in a larger multi-disciplinary research project, the Soft Reliability IOP-IPCR project, sponsored by the Dutch Ministry of Economic Affairs, which enables tackling the lack of information from different perspectives. This collaboration and industrial case-studies showed the feasibility and relevance of the research. The technology presented in this thesis is currently being commercialized for real-life application in the business contexts of industrial stakeholders.