SysTems Research on Intelligent Design & Eng.

SysTems Research on Intelligent Design & Eng.



Below are some of the current projects taking place in the STRIDE lab.

Utilizing Gaze Data to Develop Product Preference Design Models

This study aims to provide product designers with feedback on designs earlier in the design process in order to obviate the need for manufacturers to develop expensive and time-consuming prototypes. For example, the prototype BMW X6 cost millions to manufacture with the intent of gathering user feedback on the design. In order to accomplish this, eye-tracking equipment is used to collect gaze data to determine consumer’s expectations or perceptions of an ideal product design. Furthermore, a model will be formalized to understand how gaze fixations (the length of time a user stares at a specific point in space) and scan path (the path of fixations through which a system is visualized) can assist in determining the cognitive intent and interest in the user. This model can then be used to reverse engineer products that are tailor-made to suit the customer’s design demands. The models will be developed targeting different demographics categorized by a variety of social, economic and physical factors such as gender, age, height, education, income, hobbies, and marital status. The research will also be used to validate the effect of these social, economic, and physical factors on consumer’s product preference. 

Exploring Requirement Change Propagation through Physical and Functional Domain

Current research focuses on mechanical design, specifically on change management through engineering requirements. Requirements play very important roles in the engineering design process as their establishment is one of the initial steps that develop solutions to the identified problem. Requirements are translations of customer expectations into objective statements that specify how each expectation will be satisfied and therefore, play an important role in the design process. Requirements are frequently revised, due to iterative nature of the design process, to reflect engineering changes caused by changes in customer expectations, changes in the design process, or both. These changes, if not properly managed, may result in financial and time losses leading to project failure due to possible undesired propagating effect of requirement changes. Prior research developed the requirement change propagation prediction (ARCPP) tool that is able to predict change propagation as a result of an initial requirement change using the natural language data extracted from requirement statements. However, why requirement statement are able to predict change propagation remains to be understood. This work tries to explore the functional and physical relationships of requirements and their role in predicting and managing requirement changes in order to improve ARCPP tool in finer resolution and to automate it.

Application of Complex Network Metrics to Support Requirement Change Propagation Prediction

Requirements management is an integral part of the engineering design process. Requirements are frequently changed to reflect up-to-date stakeholder expectations and improve product design characteristics, and hence cannot be fully eliminated from the engineering design process. Since requirements drive product development from concept ideation to mass production, mismanaged changes to requirements can adversely affect project timelines and budgets. This research seeks to support computational reasoning of requirement change propagation by modeling requirements as networks using linguistic information contained therein, and applying appropriate complex network metrics to deduce pertinent information from them. Prior research in engineering change management has culminated in a tool that uses requirements data to predict change propagation in complex systems. The ARCPP (A Requirement Change Propagation Prediction) tool utilizes natural language information in requirements documents to deduce relationships between product entities and implements a root mean square (RMS) scoring technique to detect change propagation triggered by an initiating change. Complex networks provides a myriad of tools to assess static and dynamic features of physical phenomena represented as networks. This research seeks to enhance the ARCPP tool by providing an alternate means to assess requirement change propagation, by using complex networks.