Sensing represents a critical issue in the design of steer-by-wire systems. Specifically, the road feedback information resulting from the interaction between the tire, vehicle dynamics and road is important for the quality of force feedback. Its sensing is however related to disadvantages with respect to cost, robustness and failure issues. In this paper instead, the estimation strategies are proposed as favorable approach. The friction forces are modelled as stochastic Gauss-Markov process and an extended Kalman filter is designed for its estimation. This signal is further coupled to a steer-by-wire control setup. Simulation data with a non-linear vehicle dynamics model validate the applicability and efficiency of the approach.