European statistics show that motorbikes road accidents are extremely high and the reduction of such accidents is one of the main concern for the European community. Advanced Driver Assistance Systems are safety electronic systems used to assist the driver in avoiding risks and road accidents, by means of warnings sent before the situation becomes critical. The use of such systems in motorcycle context is currently lacking due to numerous variables that it is necessary to consider for making sure the riding. This paper presents an innovative research for the safety improvement of Powered-Two-Wheelers (PTW) by means of the development of effective and rider-friendly interfaces and interaction elements for the on-bike assistance systems. In particular, the paper presents the experimental results on comfort and safety aspects of two advanced rider assistance systems: the Frontal Collision Warning (FCW) and the Lane Change Support (LCS). The study starts from analyzing results of motorcycle simulator tests performed in 3D Virtual Reality environments which aim is to find recursive rider’s behavior patterns in FCW and LCS situations according to different multimodal type of warnings (visual, audio and haptic). Afterward, the paper presents three different machine learning models, Hidden Markov Models, Support Vector Machines and Artificial Neural Networks, that have been considered for simulating the riders’ behavior patterns according to the reaction time needful for avoiding a front collision. These simulation behavior models enabled to design a warning delivery strategy for apprising the rider of possible dangerous situations due to front collisions. Finally, the paper describes how this warning delivery strategy has been implemented in a HMI (Human Machine Interface) installed on motorbikes. This HMI is thought to offer an effective FCW system based on an understandable but, at the same time, discreet and unobtrusive rider-friendly solution.

This content is only available via PDF.
You do not currently have access to this content.