Monitoring driver status through in-vehicle parameters.

The primary objective of DREAM (Driver Related Evaluation And Monitoring), DRIVE project V 1004, was to demonstrate that it would be possible to develop a device capable of monitoring the state of a vehicle's driver unobtrusively. In the first phase of the project, a review of the state-of-the-art showed that most of the essential conditions for building such devices are already met. Only information to determine the position and/or situation of the vehicle relative to it's environment is not yet available. New developments (for instance in DRIVE and PROMETHEUS) will have to supply this information within due course. In the second phase, a series of demonstration experiments was carried out to demonstrate the existence of interrelationships between the state of the driver, his or her driving performance and the driving environment. For this, a variety of techniques was used, involving laboratory, simulator and field trials examining various aspects of driver status and performance using primarily psychological, physiological, environmental and vehicle handling measures. The results of the experiments show that driving performance is not to be conceived as a simple and unitary phenomenon which can be established reliably by a single measure. A number of unobtrusive in-vehicle measures were identified which provide reliable indices of changes in the state of the driver. However, a considerable amount of further work is necessary to select precisely which vehicle parameters should be incorporated in DREAM, to determine the nature of the interrelationships involved and assess the thresholds or limits allowed with respect to preventing traffic hazards. On the basis of the work done and future research in this field, it will be possible to specify DREAM. It is proposed that the device itself should be operating on the basis of a neuronal network, enabling the system to cope with a multitude of signals adaptively in real time. This would simplify resolving problems such as selection and weight of criteria for limiting hazards or setting the system to individual differences.