A driver-automation system for brake assistance in intelligent vehicles

In this paper, we present a driver-automation system for providing brake assistance to the driver using the concepts of various levels of intervention. The system offers potential for further development and integration into main stream cars using by-wire technology, both from passenger safety and system usability perspectives. The criticality of the situation of a potential crash is estimated using a Bayesian framework from literature which considers not only the vehicle´s situational need for braking but also the driver´s intent to apply brakes. The novel idea of our work is a well-defined dictionary for critical zone segmentation is incorporated into the system to decide the appropriate intervention level by the automation system for each situation, and appropriate assistance is provided to the driver, in terms of not only communicating auditory messages to the driver but also applying automatic brakes if need be to shift the criticality from a high/very high to a low zone. The second idea of our work is the feedback control of the amount of automatic braking necessary to be applied by the system. Results from sample runs of simulated testing of the system have been reported and future scope of work towards improvement is discussed.