Low Rate Error Modeling of Articulated Heavy Vehicle Dynamics and Experimental Validation

In the automotive domain, adequate control and diagnosis rely on the use of state observers and parametric identification systems to estimate the dynamics performances of the vehicle. Unfortunately, the simultaneous use of different methods of observation, estimation and identification is not risk-free. The risks can be expressed mathematically through a problem of error accumulation, posing major risks for the vehicle and its driver (errors of detection, errors in the prediction of dangerous driving situations, vehicle instability, etc.). This paper presents a method of observation and estimation of the dynamic state and parameter identification of an articulated vehicle simultaneously at very low error rates. This method is based on the HOSM (High Order Sliding Modes) approach, with the application of the STA (Super-Twisting Algorithm). Towards to this aim, a 5-DOF (Degree Of Freedom) nonlinear dynamic model for an articulated vehicle is proposed. The model is derived by applying Lagrange's equations. Simulation and experimental results showed that the algorithms generate accurate estimation of articulated vehicle parameters and states dynamics in real driving situations.