Design and analysis of a new brake-by-wire system using machine learning

Abstract

One of the main aims of the recent research on brake-by-wire systems is to decrease mechanical components. In this paper, we propose replacing the brake pedal with a driving wheel that is fully covered by pressure braking batch sensors. The new mechanism for braking translates pressure exerted through the driver’s hands on the driving wheel to a corresponding electrical signal. A proposed design for the pressure braking batch (PBB) is made out of a mesh of conducting threads separated by a resistive sheet. To the best of our knowledge, this idea has not been raised before in other research papers. Different people have different muscle strengths and so the problem of identifying the intention of the user when pressing the PBB is tackled. For this aim, a new dataset of its kind is created by several volunteers. From each volunteer, age, gender, body mass index (BMI), and maximum pressure exerted on the driving wheel are collected. Using Weka software, the detection accuracy is calculated for a new volunteer to know the intention of his/her pressure on PBB. Among the three algorithms tried, the regression tree gives the best results in predicting the class of the pressure exerted by the volunteers.