Psychologically Plausible Cognitive Models for Simulating Interactive Human Behaviors

Abstract

The intent of Sandia National Laboratories' Human Interactions (HI) project is to demonstrate initial virtual human interaction modeling capability. To accomplish this, we have begun the process of simulating human behavior in a manner that produces life-like characteristics and movement, as well as creating the framework for models that are based on the most current experimental research in cognition, perception, physiology, and cognitive modeling. Currently the simulated human models can sense each other, react to each other, and move about in a simulated 3D environment. A preliminary action generation or motor-level cognition model, which transforms abstract actions generated by high-level cognition to actions that can be carried out by a simulated physical human model, has also been developed. Our work has yielded models of perceptual, spatial, and motor functioning and memory that will be embedded in upgrades to the cognitive framework.

Applying cognitive science to digital human modelling for user centred design

“…During the last decade, there have been several attempts at incorporating cognitive modelling in DHM, most of which have focused on using cognitive architectures to predict human performance. A research group at Sandia National Laboratories in New Mexico have created a framework based on a modular and symbol processing view of human cognition and others have focused on a rule based system built on architectures such as ACT-R and Soar (Bernard et al, 2005;Carruth et al, 2007). Though not built on exactly the same architecture, several others have gone about the problem in similar ways, ultimately trying to reach a state where the system can, at the press of a button, perform a cognitive evaluation (Gore, 2006).…”

“…Previous suggestions on integrating cognition in DHM tools have largely taken their basis in symbol processing architectures such as ACT-R, Soar etc. (Bernard et al, 2005;Gore, 2006;Carruth et al, 2007); architectures that disregard embodiment and situatedness of cognition. This paper places the computer manikins used in DHM tools within a context, a context where cognitive offloading and scaffolding onto the environment is supported.…”

Preparing for the aftermath: Using emotional agents in game-based training for disaster response

“…Using subsequently-developed simulation software that extended practicality, [21] developed a prototype human augmentation system based on discrepancy detection with respect to a task-based runtime cognitive model. A high-level, psychological framework was fleshed out (Figure 4) to enable application within an embodied agent, as distinguished from a disembodied "decider" [22], [23]. At the same time, the Sandia Cognitive Runtime Engine with Active Memory (SCREAM) was developed to provide a practical cognitive simulation capability that supported that psychological framework.…”

“…Applying SHERCA for the conceptual and psychological model structure and SCREAM for the computational cognitive engine, [24] used runtime cognitive models to control the behavior of cognitive characters in a virtual 3D environment in a prototype training application emphasizing cultural awareness [22]. Standard embodied-agent simula-tion techniques implemented in Sandia's Umbra simulation framework [25] model the characters' ability to "perceive" their environment, other entities, and those entities' various attributes and actions.…”

Bridging the Gap: Exploring Interactions Between Digital Human Models and Cognitive Models

“…For this same reason, physical and cognitive models of human performance also need to be integrated. Approaches such as HOS [12] and MIDAS [13] as well as work by researchers at Sandia National Laboratories [14] represent initial efforts to integrate physical and cognitive models into a single approach by including independent "modules" (micromodels containing algorithms based on human experimental literature) that can be used as necessary to predict an aspect of performance. One criticism is that performance predicted by these "plug and play" modules actually may not be independent; that is, interactions between different aspects of performance may produce behaviors and performance that the modular approach would have difficulty predicting accurately.…”