A Driver Lateral Control Model By Integrating Neuromuscular Dynamics Into The QN-based Driver Model

Driver models can not only reveal the driving mechanism, but also compute,simulate, and predict driver behavior, and thus can help improve the development processof vehicles and driver assistance systems. Furthermore, they can provide new insightsinto the research and development of the control technology of unmanned vehicles andintelligent driver assistance systems. Therefore, the study of driver models has importantscientific significance and practical values.The existing driver models focus on either the control (including the aspect of driverneuromuscular system) or cognitive aspects of driving behavior. Therefore, these modelscannot account for driving behavior well, which limits the value of these models insupporting the development of driver assistance systems. To deal with this issue, wepropose a novel driver lateral control model by integrating the driver’s neuromusculardynamics into the queuing network (QN)-based driver model previously developed by us.Based on the proposed model, we propose a dual-task modeling of driving with a motorsecondary task. The proposed driver lateral control model is based on the QN cognitivearchitecture, and thus they can capture psychological and physiological capacities andlimitations of real drivers. The combination of control methods into the cognitivearchitecture can mathematically formulate the driver vehicle control. Further, theintegration of the driver’s neuromuscular dynamics into the QN-based driver model cancapture the dynamic interaction between the vehicular steering system and the driver’sneuromuscular system.The main accomplishments that this thesis achieves are listed as follows:1) We improved the previously proposed QN-based driver model and developed theneuromuscular dynamics model for the driver steering control. On the base of the twomodels, we developed a novel driver lateral control model by integrating the driver’sneuromuscular dynamics into the QN-based driver model.2) We explored the impact of the parameters (the muscle co-contraction stiffness andthe gain of reflex control module) of the driver’s neuromuscular dynamics model on theperformance of the driver lateral control, by using the proposed driver lateral controlmodel. 3) Based on the proposed driver lateral control model in the driving task condition,we proposed a dual-task model of driving with a motor secondary task by using thescheduling method for dual tasks.These above achievements can not only help reveal the driving mechanism betterand account for the influence of the secondary task on the driving task, but also developcorresponding tools to assist the research and development of the driver assistancesystems.