Nowadays, vehicles have been widely equipped with power assistance steering systems and the way of providing assist torque is various. Electronic Power Assistance Steering (EPAS) systems have the tendency to replace the Hydraulic Power Assistance Steering (HPAS) systems, due to better fuel consumption, better road feel feedback to the driver and better return-to-center performance of the steering system. Actual EPAS systems are designed basing on a general driver’s profile and they try to mimic the behavior of the HPAS system. Main goal of this thesis is to propose a general methodology to adapt standard EPAS for drivers with reduced mobility. Toward this general goal, following contributions are integrant part of this thesis.First, a mathematical model of the steering system that keeps into account the torsion of the steering column is introduced. This model is influenced by two exogenous inputs: the tire-road friction torque and the driver's torque. Mathematical models for these inputs are also proposed. Due its importance on the steering system, great attention is dedicated to the description of the tire-road friction torque. This torque includes the contribution of two main phenomena: the self-alignment and the sticking friction torques. They appear, respectively, at high and low speeds of the vehicle. Dynamical friction models, based on the LuGre theory, are used to describe these phenomena.Then, a study of the steering model subject to the influence of exogenous inputs is done in open loop. This study shows that the steering column starts to oscillate when an exogenous input is applied. To delete low-frequency oscillations of the column, an optimal state-space control is introduced. Moreover, an observer is designed to estimate state variables of the system and exogenous inputs, acting on it.Third contribution of the thesis deals with the aim to provide some rationality to actual methods that are used to implement electronic power assistance on commercial vehicles. Indeed, investigations about actual assistance systems do not provide ground foundations or evidence that the current assistance is optimal for the driver in any sense. At this aim, an optimization procedure is proposed to retrieve the shape of the steering assistance, starting from the analysis of the steering movement, in case of a general driver's profile, and the analysis of the human perception face to an external stimulus.After that, a procedure to adapt the standard steering assistance to the driving of disabled people is proposed. As there is not any classification of drivers according to their pathology, attention focuses on drivers affected by diseases to the upper limbs that cause an asymmetry into the torque production on the steering wheels. A study about main problems that these patients are dealing during some critical driving manoeuvres is done. Starting from this, a compensation procedure is proposed to modify the standard power steering assistance. This procedure is validated in simulation. Moreover, objective evaluation criteria are derived to test improvements due to the proposed assistance.Finally, the experimental validation of all elements proposed in the general methodology is done on the tele-operation platform, present at the laboratory Gipsa-LAB. This system is made up of a control station with the operator, a pc-unit and a tele-operated vehicle, that allows retrieving a real profile of tire-road friction torque. Results about these experimental tests validate benefits of the adaptation methodology. They represent also the starting point for the installation of the proposed assistance on a real vehicle and set the basis for the application of the methodology for a wider class of patients, affected by different pathologies.
