The buildings designed by architects are settings for human activities. A good design can foster interactions among users, boost productivity, and maximize satisfaction, while a poor design can hinder users’ performance, reduce satisfaction, even undermine people’s safety. Assessing whether a proposed building will be considered good or poor before it has been built and occupied is one of the main challenges for architectural design.
Different from most other design practices, where products can be prototyped and tested before mass production, architectural artifacts are too big, complex and costly to be prototyped and tested before their construction. Mistakes can lead to user-dissatisfaction, lower productivity, and in the case of hospitals – to medical mistakes.
It is therefore imperative for architects and clients to be able to predict and evaluate a building’s performance during the design process itself. While current prediction methods involve calculations related to structural aspects, light, energy and sound. However, predicting, analyzing and evaluating human behavior in not-yet built environments is a more complex task, which involves psychology, sociology, and culture, in addition to physical and ergonomic factors. At present, architects refer to norms, regulations, and case studies to predict and assess human behavior in design projects. These methods, however, cannot assess the dynamic and stochastic components of human behavior in built environments. For this reason, a gap exists between the behavior that architects and planners expect to be performed, and the actual behavior of end-users’ in the designed environments.
To narrow this gap, the research investigated the potential of assessing human behavior in simulated built environments by means of 3D game engine technology. In addition to geometrical built environment representations generated with Computer Aided Design (CAD) and Building Information Modeling (BIM), the simulation system uses autonomous, goal-oriented, anthropomorphic agents (called Virtual Users) that perform a predetermined set of activities in 3D virtual environments resembling specific use processes of the intended environment. Rules of behavior, which are based on observed activities in similar facilities, are computed for each Virtual User, which dynamically responds both to the physical properties of the built environment, and to the presence of other Virtual Users, adapting its behavior accordingly. The simulation allows for coordinating multiple agents’ behavior to achieve larger goals as determined by the organization that occupies the built environment. Agents’ behavior is driven both by a scheduled set of activities that need to be performed, as well as by an unscheduled set of activities that originate from agents’ situated interactions with their physical and social environments.
Hospitals have been chosen as test bed for the system development and use. Designing hospitals is a complicated task because of the wide range of users and the variety of functions that are carried out in the same location. Yet, given the general aim of hospitals to heal patients, hospitals can be considered as highly specialized “machines” that implement state-of the art medical technologies and procedures to “fix” ailing patients. This formalization was found to be advantageous for the purposes of the research, because it provides a comprehensive and agreed-upon set of behavior patterns on which to build the model.
The simulation has been tested in four major hospitals in Israel. We found that the simulation enabled the architects and hospital managers to analyze and evaluate specific aspects of their design solutions, such as the scale and position of physical elements within the designed setting, in support of the activities of the indented users. In particular, the simulation process enabled the analysis of circulation patterns, times, distances, as well as the identification of bottlenecks and density areas. By running “what-if” scenarios, the research was able to test at the same time multiple activities scenarios and alternative design solutions for the built environments, in order to generate a better match between spatial features and human activities.
Simulating human behavior has a strong potential for analyzing and evaluating environmental design solutions from a human behavior point of view, both in professional design practice and education.
