In an usual embedded product we see following set up . Processing logic runs inside MCU and both sensors and actuators are connected to it . MCU takes input from sensor and based on that input processing logic decide how to control actuators . Same like the example given in here .
During the life time of the product , MCU could fail due to several reasons Ex : High voltage or current , static , reset due to internal logic state etc. The example here is a safety critical scenario. As in case of a fire if water is not sprinkled in time it may lead to loss of human life. So in such a safety critical scenario , how can we ensure higher reliability. As it is possible MCU failed to deliver when needed.
This is where we use SBC . It continuously monitor MCU health and based on that decide to take over the control of actuator . It does not have the processing logic to decide actuator state based on sensor input. So SBC push actuator to a default safe state. This safe state is decided as part of system design. For example here start high pressure pump to sprinkle water could be a safe state. So regardless of whether there is fire or not , if SBC detect MCU failure , then it start pressure pump. So in any case risk is averted. Now one can argue that , it is possible both MCU and SBC failed. Yes it is possible, after all they are electronics chips. So we can not ensure 100% reliability , but we increased the reliability and decreased the probability of failure by using SBC.