No abstract

No abstract

This paper deals with two modeling approaches to car driving. The first one is a system theoretic approach to describe adaptive human driving behavior. The second approach utilizes neural networks.As an illustrative example the overtaking task is considered and modeled in system theoretic terms. Model results are used to teach a neural network. The results show that a neural network is able to learn this task even when certain task variables change. The next step is to perform an experiment with real human operators in order to assess the validity of both modeling approaches and their relative merit.

No abstract

The visual scene is an important source of information for the manual approach and landing task. This paper deals vith the effect of this information in combination with baaic display information on the approach perfoxmauce. In this context, a pre-experimental model analysis has been performed in terms of the optimal control -1 . The resulting aircraft approach performance predictions were caupared vith the results of a moving base simulator program.The results illustrate that the model provides a neanbgful description of the visual (scene) perception process involved in the c a q l u (multivariable, tine varying) manual approach task with a useful predictive capability. The theoretical fismnrork has been s h m to allow a straightIomard investigation of the complex interaction of a variety of task variables.The manual approach and landing is a complex manual control task. Tke process is t h e (range) va.rying and involves multivariable task objectives, visual scene and display information and a colnplex pilot's control strategy. Although many studies have dealt with s variety or aspects of this approach and landing task, accident statistics indicate that there are still important unsnavved questions. This p6per s\pmrsrizes the results of a theoretical and experimental program addressing the effect of v i s u a l information on the manual approach and landing. SpecificaUy, this concerned Visual scene information which vas the subject of a previous study (Refs. 1 and 2) and basic (head-up) display infomation. From that study it could be concluded that the Visuil acene perception process can be modelled (described) on the basis of linear perspective geumetry and relative motion cues.In the present study the effect o? visual scene information was investigated by considering three (good, poor and night) visibiiity conditions. These three conditions were combined with three basic head-up https://ntrs.nasa.gov/search.jsp?R=19820026165 2018-05-12T05:06:22+00:00Z