Extended Sensor Reliability Evaluation Method in multi-sensor control systems

• It is possible to estimate a level of sensor credibility in order to improve a reliability of multi-sensor control systems.
• The applied mechanism discards readings of the broken sensors.
• The effectiveness of the method was proved by simulation results.

This article discusses a reliability evaluation method of sensors in multi-sensor control systems. Based on performed analysis and research work a universal method for the credibility evaluation of information from sensors was developed. The purpose of this method is to estimate a level of effectiveness of sensors, assign a proper weight to them during the voting process and utilization in multi-sensor control systems as a decision unit. The Sensor Reliability Evaluation Method for estimation of credibility coefficients processes results of a decision module, which, on the other hand, uses credibility coefficients evaluation of participating voters to reach a right verdict. In this case, the voting process takes a form of a weighted voting system. The method can be also used to identify sensors with some malfunctions. When a sensor reliability falls to 0 or other declared value an alarm signal can be generated. In order to forward the tests of the Sensor Reliability Evaluation Method a simulation was proposed. Presented results verify assumptions of described method and show high effectiveness of this solution.