PRODUCING A NEURAL-NETWORK FOR MONITORING DRIVER ALERTNESS FROM STEERING ACTIONS

There is a limit to the accuracy with which we can predict a person's state of alertness from their behaviour. Driver behaviour and alertnes s are, however, clearly related, and this should allow us to build a p redictive model. For such a model to be of use it must be very general in its ability. Such generality is available at the expense of accura cy and a trade-off between overall error rate and quantity of usable p redictions must consequently be made. This paper discusses a set of me thods which were applied to the task of building a neural network base d system for predicting driver alertness from steering behaviour. We s how how an acceptable level of generality was achieved and how the tra de-off between error rare and quantity of usable predictions was manag ed.