Motorcycle safety is essential for the future of riding.Honda has been pioneering motorcycle safety systems from the first disc brake in 1969 up to the latest airbag and C-ABS systems. In recent years, the focus is more and more shifting towards the rider and rider skills.
Online analysis and quantification of riding skills allow to improve rider education by precise individual feedback and engage potentially safety relevant systems such as stability control or future emergency braking systems.
HRI-EU developed a machine-learning based algorithm that is trained with everyday riding data. It automatically understands riding maneuvers such as riding straights or preparing, going through and exiting curves.
The algorithm analyzes how precise each of the rider’s turns and corners are and how stable the motorcycle is during these maneuvers. Notably, all this is under typical riding conditions on popular weekend routes and riding appropriately in a safe manner. By testing with a number of riders from beginners, intermediates, experts as well as professional test riders and even racing champions, we found that we can reliably classify rider skill from a ride of about 50km. The algorithm is also able to find riding inconsistencies and can measure how smoothly each curve was handled.
We are currently investigating the usefulness of this technology for rider education, first for the Honda internal education of professional test riders and possibly for general driving schools in the future.
We are also researching ways to extend these results by measuring the rider’s body posture and, e.g., handle grip strength and body stiffness as well as eye movements with respect to the road and other vehicles using gaze tracking glasses in combination with cameras and convolutional network based object recognition. By this, we want to advance the understanding of the human rider and make Honda motorcycles safer and more adaptable to individual skill and behavior.