Table 1.

Principle, main advantage and disadvantage of existing methods.

Existing Methods	Principle	Advantages	Disadvantages/Limitations
Learning a Curve Guardian for Motorcycles [14].	Using IA to improve the curve undertaking. Analyzing the instruments’ measured data.	Lane localization, adding a learned roll prediction approach, using standard maps for real-world evaluation.	Does not anticipate the danger before undertaking the curve. Does not consider road grip in the estimation of roll angle.
Powered Two-Wheeler Riding Pattern Recognition Using a Machine-Learning Framework [15].	Using IA to predict rider behavior. Analyzing the instruments’ measured data.	Recognition of the driver action by applying ML on a dataset of measurement collected by sensors.	Does not reply completely to the challenge of anticipating dangers on the curve. Uses a set of sensors that may raise costs.
Powered Two-Wheelers Critical Events Detection and Recognition Using Data-Driven Approaches [16].	Development of critical event detection methodology by using AI classification algorithms.	The classification of the events with ML techniques, which could constitute a good database to characterize critical events.	The experiences collected based on a single model HONDA CBF 1000 which is not beneficial in case of generalization.
Estimation of Mental Workload during Motorcycle Operation [17].	Development of a method for mental workload when riding a PTW.	Contributes to characterize the driver behaviour by estimating the level of fatigue.	Deal partially with the challenges of PTW concerning curve undertaking.
Lateral & Steering Dynamics Estimation for Single Track Vehicle: Experimental Tests [18].	Development of on-board instrumentations for lateral and steering dynamic estimation.	Deals with lateral and steering dynamics estimation the reconstruction of unknown inputs.	Does not anticipate the danger before undertaking the curve. The loss of information due to linearization is considerable.