The design of a robotic vehicle of Autonomous Mobility Laboratory (LMA) at UNICAMP is developed in-vehicle platform FIAT-PUNTO. In addition to a set of sensors, actuators, mechanism and components (hardware and software), new technologies should be developed to support the Automation, Control, Perception, Localization and Navigation. The purposes of this work is to present a non-linear vehicle model with a tire dynamic model, and the design of virtual sensors applying a stochastic estimation technique for acquiring indicators of the intrinsic safety of the vehicle. The virtual sensors allow to estimate several critical parameters of the vehicle dynamics. The tests simulations of the vehicle model are executed applying the extended Kalman filter (EKF). The estimated responses data are evaluated with the simulated vehicle and tires model, allowing the validation of the vehicle model and the virtual sensors proposed.