This report argues that in order to develop reliable intelligent interfaces in motor cars, a driver model should be developed which reflects human information processing mechanisms and, more specifically, mechanisms of skill acquisition. Two mechanisms are proposed to underlie skill acquisition, namely involuntary priming and voluntary preparation. On this basis three alternative models of skill acquisition are proposed that differ with respect to the effects of voluntary control at the perceptual and response level of information processing. Subjects carried out two-choice reactions in rapid succession. The most important experimental manipulations were (1) whether the first choice reaction predicted the second and (2) the degree of transfer of training to conditions where predictivity changed. In addition, stimulus presentation was for some subjects always visual whereas for other subjects only the first stimulus was visual and the second was auditory. The results support a model asserting that involuntary effects of priming evolve only at the perceptual level but not at the response level. In addition, they support earlier findings that preparation for the second reaction occurs, in part, before execution of the first one. Correlational analyses of individual differences indicate that an overlapping strategy during training yields involuntary priming whereas a sequential strategy without overlapping preparation does not. Together, the results are in close agreement with a connectionist-control model of human information processing which consists of separate processing modules each of which can be described as a neutral network (Schneider & Detweiler, 1987, 1988). Finally, implications for the Generic Intelligent Driver Support (GIDS) system are presented. (Author/publisher)
Abstract