With the development of information technology, instruments based on intelligent-interface-design principles emerged in cars, such as the e-vehicle or telematics system with the function of mobile communication are called. Speech communication is a most important part of information communication during driving. But, it is arguable to question its effect on driving. Would it be safe, although more convenient? Would it be more dangerous as more actions requiring shared attention and more mental workload (MWL)? The three major aims of the study were (1) to develop and standardize reliable diagnostic methods for MWL measurements [i.e., subjective ratings, objective driving difficulty (ODD) measures and achievement in mental performance] in different traffic situations; (2) to investigate the interaction between speech communication and driving behavior, and (3) to detect the variations
in the amount of spare mental resources as the function of driving difficulties and speech processing. The methods of laboratory experiments, field study and survey were used in the study.
Survey 103 drivers were interviewed with the questions concerning the state-of-the-art of telematics equipment in their cars, and the frequency, purposes and contents of their uses communication means while driving.
Field study. Based on their reports, experimental tasks were designed to test the effect of different speech processing processes (i.e., verbal comprehension and speech production) and the effect of different levels of language processing (i.e., sentence processing and paragraph processing), namely, (a) sentence comprehension, (b) paragraph comprehension, (c) sentence-making and (d) answering questions by talking within a fixed extension of time. A 2 X 2 X 3 design was used in the field study. That is, the four experimental tasks were conducted under different traffic situations, respectively, easy driving in the highway, medium-difficult driving in auxiliary busy roads, and difficult driving in crowded downtown streets.
Laboratorv experiments. A driving The subjects simulator with the 3D visual scenes was
used for the laboratory tasks. were asked to drive while conducting the same verbal tasks under situations varied in driving difficulty (i.e., along a straight road or in a road with many turns) and in driving environment (i.e., with or without road-blocks).
The main results were as follows:
(1) NASA TLX rating scale was sensitive to driving task. A significant increase was found with the increasing driving difficulty. Time pressure perceived by the subjects under dual-task situation was the main factor that increased MWL. The ODD was also a reliable criterion for detecting MWL in driving, as it could reflect the dynamic variation of MWL in real situation.
(2) There was a significant effect of verbal communication on driving. Comparing the two kinds of speech processes, driving had more influence on verbal comprehension than it did on speech production. In verbal comprehension, the response time was significantly delayed. In speech production process, significant differences were found in the amount of words produced, time and speed of speech, and lexical fluency between the single-task and the dual-task conditions. With the increasing of driving difficulty, the speed of speech slowed down, while repetitions of words increased.
(3) . The spare mental resources were changed significantly as a function of difficulty both in driving and in verbal communication. Under the highest dual-task dii}culty condition, "red line" effect was detected in visual selection task, and subjects slowed down their driving speed to decrease perceived MWL, which showed the limit of human's mental resources.