Steering maneuvers determine the lateral stability of a running vehicle and are considered as an important indication of driving performance. To effectively measure the steering wheel angle, we propose an estimation approach by using a built-in smartphone tri-axis gyroscope. First, coordinate transition was applied to the angle velocity about the Z-axis of the phone to calculate the vehicle yaw rate. A linear regression method was then used for the modeling of correlations between the resulting yaw rate and the steering angle. Comprehensive experiments show that the steering angle can be well approximated through the angle velocity when the angle remains within a specified range. In addition, sudden changes of angle can be viewed as occurrences of abnormal driving behaviors. This study show that off-the-shelf sensor technologies provide a cheap and portable solution for driving behavior monitoring.
展开▼
