• articulated vehicle
• data analysis
• gray-box model
• identification algorithm
• inertial parameters estimation
• least squares method
• mathematical modelling
• multibody simulation
• system identification
• truck trailer

In this thesis, which is the result of six months of work at Fraunhofer LBF research centre, a truck trailer inertial parameters estimation technique is presented. The mass, the position of the centre of gravity, the mass moments of inertia of the trailer were estimated. After bibliographic research of the state of the art of parameters estimation and state identification, both applied to the automotive field, the problem has been divided in two sections: an offline technique was carried out for the estimation of the parameters while an online technique was used for the state identification. Two models, of both the truck and the trailer, were developed: a longitudinal model for the evaluation of the mass and the position of the centre of gravity, a ride model for the evaluation of the roll and pitch mass moments of inertia. To correctly identify the different driving conditions, and thus enable the correct application of the different models, an identification algorithm has been developed. The input data to the algorithm are obtained by means of a multibody model (developed in Adams View), which was validated with the actual test vehicle. The multibody model was also used for the validation of the results obtained with the identification algorithm. The identification was carried out on a longitudinal trajectory, performed on two surfaces with different degrees of road roughness (A and B ISO road class). In order to take into account a stochastic variability of the results, different (five) road profiles for each ISO class were simulated (by varying a so-called seed number in the Adams road model). In addition, the results are shown when varying the available signals used for the estimation. The results obtained for the estimation of the mass and position of the centre of gravity show an increment of the error as the degree of roughness increases, while the error remains almost constant for the estimation of the roll and pitch mass moments of inertia. The developed vehicle model also allows the estimation of vertical tyre forces.