Rule-Based Approach for Simulating Age-Related Usability Problems

Ambient Assisted Living requires easy to use interfaces, making usability a critical feature. Because usability evaluations are resource and time consuming, several automation efforts have been made, one of which is the simulation of users interacting with UIs. In this article, we present ongoing work of a tool for automated usability simulations that allows simulating agerelated deficits. The tool is specifically intended to be used by IT practitioners, i.e. in difference to cognitive architectures that allow similar simulations, this tool does not require extensive knowledge in cognitive science. A core component of the simulation tool is its rule-based User Model (UM). During a simulated interaction, the UM selects actions causing a model of the UI to change states until a specified task goal is satisfied or the UM “gives up”. Interactions of the UM are calculated from probabilities which are informed by rules drawing on user and UI attributes. Using a Monte Carlo approach, the simulation is iterated, resulting in a set of task solutions where non-optimal solutions may indicate usability problems. By analyzing which rules led the UM to interact non-optimally, our approach can offer hints on how to improve the UI. While our approach cannot render user-based evaluations unnecessary, our aim is to substantially reduce the effort involved in usability testing of UIs as well as to provide an automated tool that can be used early on in the development process.