CAPRIO with Inclusive Pedestrian Path Recommendations

Accessibility and usability have been key concerns in the design of computer interfaces through which users interact with applications and systems. Recently, chatbots have gained popularity with service providers for improvements in this area. In this paper, we present our experience in designing and implementing CAPRIO’s inclusive chatbot-based interface for pedestrian path recommendations. Our CAPRIO system provides inclusive usability by extracting user preferences in a non-intrusive dialog and using them to build a more accurate model for the user’s intent. It uses the Microsoft Bot Framework (MBF) and NLP modeling to support text and voice dialog.

I. Introduction

Context-awareness in human-computer interactions is essential for the everyday activities of an individual to maximize the utility of the computer systems and applications which the individual interacts with. For example, a context-aware path recommendation system can provide recommendations that are robust to a wide range of constraints of great importance to many user groups. For instance, during inclement weather, such as the 2021 heatwave in the USA and the 2021 polar vertex in Canada, pedestrians are more vulnerable to heat and cold-related injuries and illnesses. In addition, moving crews need to avoid paths with many steps and narrow staircases. As the examples suggest, context is subjective and while these systems can provide utility for many users under several scenarios, benefiting users from all backgrounds requires holistic thinking about their design.