Road profile estimation using an adaptive Youla-Kučera parametric observer: Comparison to real profilers

Road profile acts as a disturbance input to the vehicle dynamics and results in undesirable vibrations affecting the vehicle stability. An accurate knowledge of this data is a key for a better understanding of the vehicle dynamics behavior and active vehicle control systems design. However, direct measurements of the road profile are not trivial for technical and economical reasons, and thus alternative solutions are needed. This paper develops a novel observer, known as a virtual sensor, suitable for real-time estimation of the road profile. The developed approach is built on a quarter-car model and uses measurements of the vehicle body. The road roughness is modeled as a sinusoidal disturbance signal acting on the vehicle system. Since this signal has unknown and time-varying characteristics, the proposed estimation method implements an adaptive control scheme based on the internal model principle and on the use of Youla-Kučera (YK) parameterization technique (also known as Q-parameterization). For performances assessment, estimations are comparatively evaluated with respect to measurements issued from Longitudinal Profile Analyzer (LPA) and Inertial Profiler (IP) instruments during experimental trials. The proposed method is also compared to the approach provided in Doumiati, Victorino, Charara, and Lechner (2011), where a stochastic Kalman filter is applied assuming a linear road model. Results show the effectiveness and pertinence of the present observation scheme.