Experimental Performance Evaluation of Smoothing Predictive Redundancy Using Embedded Microcontroller Unit

The dependence of numerous systems on electronic devices is causing rapidly increasing concern over fault tolerance. For example, a car with electronically controlled steering and no mechanical linkage from the steering wheel to the front tires (x-by-wire) must be fault tolerant because a critical failure could arise without warning. Fault-tolerant systems have been studied in detail, mainly in the field of aeronautics. This paper presents a smoothing predictive redundancy system that can remove the effect of faults. In addition, this paper introduces an implementation of such a system using an embedded microcontroller unit to show that the smoothing predictive redundancy system outperforms the well-known average and median voters. The experimental results show that the smoothing predictive redundancy system can be a viable choice for the fault-tolerant aspects of safety-critical systems.