Lean angle estimation for single-track vehicles using inertial drift compensation

Accurate real-time estimation of a motorcycle's roll angle is essential for deploying advanced safety systems, such as stability control. This work proposes a novel and computationally efficient roll angle estimation method based on the integration of longitudinal angular velocity, augmented by an innovative drift compensation strategy. Unlike traditional approaches, such as Kalman filters, the proposed technique does not rely on prior knowledge of the vehicle's dynamic model or the tuning of multiple filter parameters. Instead, it leverages gravitational vector measurements and estimates of centrifugal acceleration to correct angular velocity bias in real time. The proposed method was experimentally validated on an instrumented Yamaha MT-03 motorcycle under naturalistic riding conditions. The results show that the proposed approach outperforms existing methods, including Kalman filter-based estimators by achieving higher accuracy with significantly lower computational cost, without requiring the tuning of multiple parameters. The method achieved a Root Mean Squared Error (RMSE) of less than 2∘ and a Mean Absolute Error (MAE) below 1.5∘, which represents an average reduction of 9.39% in RMSE and 5.60% in MAE in comparison to the existing literature.