Modeling steering using the Queueing Network - Model Human Processor (QN-MHP)

Abstract

The Queueing Network -Model Human Processor (QN-MHP) is a computational architecture that combines the mathematical theories and simulation methods of queueing networks (QN) with the symbolic and procedure methods of a GOMS-style task description and the Model Human Processor (MHP). Using QN-MHP, a steering model was created to represent the concurrent perceptual, cognitive, and motor activities involved in vehicle steering as truly concurrent processes. The model was compared with driving performance of human subjects and demonstrated realistic steering behavior. It steered a simulated vehicle at a fixed speed within the lane boundaries of straight sections and curves of different radii. In a quantitative validation of several basic measures of driving performance, the steering model yielded steering angle and lateral position similar to the human subject data. This work showed the strength of QN-MHP as a model of driving behavior. Ongoing work further develops the model by expanding the scope of the driving task and by adding a concurrent secondary in-vehicle task.

A cognitive constraint model of dual-task trade-offs in a highly dynamic driving task

“…There is, however, a significant limitation in the application of CCM to driving. Previous models of driving [13,14,15,16,18] have tended to interact with a simulation environment that, in essence, provides the mapping from behavior (e.g., steering and acceleration) to changes in the environment (e.g., the vehicle's lane position). Cognitive constraint models in contrast are not intended to interact with a simulated environment; rather, they demand a mathematical understanding of the consequences of action for objectiverelevant features of the world.…”

Computational Modeling of Driver Lateral Control on Curved Roads with Integration of Vehicle Dynamics and Reference Trajectory Tracking

“…In Tsimhoni & Liu (2003), a driver steering model was also successfully developed using processing logics including detecting orientation, selecting a steering strategy, and steering action: the model yielded the steering angle and lateral position in two fixed radii of curvature, similar to the experimental data using human subjects. However, one major contribution of this current study is that the method of reference trajectory tracking can help prediction of the driving performance on different curved conditions (rather than just fixed radius of curvature).…”

Modeling driver car-following based on the queuing network cognitive architecture

“…Computational cognitive models of driving behavior can considerably contribute to driving-related human factors research [1]. They can reveal cognitive bottlenecks, simulate driving behavior, quantitatively predict possible interference of secondary in-vehicle tasks, and thus help develop evaluation methods and human factors guidelines for in-vehicle systems.…”

“…The SOAR models do not meet human psychological findings, whereas the main weakness of the ACT-R models is the constraint of a serial process of cognitive processing. Parallel processes need to be interleaved into one serial process, yielding its dependence on explicit and often task-specific transfer of control between the concurrent activities [1].…”

“…To model concurrent activities of driving in a truly concurrent manner based on an underlying context-free cognitive architecture that is based on human psychological findings, Tsimhoni and Liu investigated modeling driving with a novel cognitive architecture [1], the QN-MHP (Queuing Network-Model Human Processor) that integrates the queuing network approach and the symbolic approaches of the Model Human Processor (MHP) [6]. QN-MHP has a parallel cognitive network.…”