This work presents a concept for a collision avoidance system simulated using 3D-models of three different vehicles implemented in MATLAB. Two of the vehicle data sets were built with generic information, used to characterize mid-size and full-size vehicles. These standard vehicles were used in initial simulations and for demonstration of some concepts. The third data set was built with information from the Apollo N, a super sportive car. These different data sets were used to evaluate the controller s capacity to work with a range of vehicles, with different sizes and driving characteristics. The approach for triggering the system is based on the time-to-colision (TTC) estimation. This concept was adopted to recognize when the driver is not able to avoid an accident. After being triggered, the system must decide which maneuver is the most appropriate for the given friction and risk conditions. The first goal of this work is to develop an autonomous braking system which evaluates the risk of a possible collision and decides if the driver is able to avoid it. Once the driver has not enough time to react, the system must trigger the brakes automatically in order to avoid the accident. The vehicle is equipped with an embedded Anti-lock Brake System (ABS) developed using Fuzzy control. The ABS controller s performance was evaluated in simulations using the data sets and tested in a scaled vehicle. In more critical cases, when there is low friction, the vehicle is not able to brake in a reasonable distance. Considering this situation, an autonomous steering control was implemented in order to make an alternative avoidance maneuver. This second system was evaluated in simulations using vehicles with understeering and oversteering characteristics. The results pointed out that the autonomous steering control was able to perform avoidance maneuvers in a reasonable range of lateral accelerations, in vehicles with different driving tendencies.