The dependence of intelligent vehicles on electronic devices is rapidly increasing the concern over fault tolerance due to safety issues. For example, an x-by-wire system, such as electromechanical brake system in which rigid mechanical components are replaced with dynamically configurable electronic elements, should be fault-tolerant because a critical failure could arise without warning. Therefore, in order to guarantee the reliability of safety-critical systems, fault-tolerant functions have been studied in detail. This paper presents a Kalman predictive redundancy system with a fault-detection algorithm using the Kalman filter that can remove the effect of faults. This paper also describes the detailed implementation of such a system using an embedded microcontroller to demonstrate that the Kalman predictive redundancy system outperforms well-known average and median voters. The experimental results show that the Kalman predictive redundancy system can ensure the fault-tolerance of safety-critical systems such as x-by-wire systems.