In this paper real-time estimation methods for identifying the tire-road friction coefficient are presented. Taking advantage of the Magic Formula Tire Model, the similarity technique and the specific model for the vehicle dynamics, a reduced order observer/filtered-regressor-based method is proposed. The Proposed method is evaluated on simulations of a full-vehicle model with an eight state nonlinear vehicle/transmission model and nonlinear suspension model. It has been shown through simulations that it is possible to estimate the tire-road friction from measurements of engine rpm, transmission output speed and wheel speeds using the proposed identification method. The proposed method can be used as a useful option as a part of vehicle collision warning/avoidance systems and will be useful in the implementation of a warning algorithm since the tire-road friction can be estimated only using RPM sensors.