An engineering approach to determining sampling rates for switches and sensors in real-time systems

An objective of our work is to find more systematic methods of designing and implementing real-time device drivers for embedded systems. As part of this objective, we answer the question “what is the best sampling rate to use for reading data from sensors that provide continuous data?” Our experiences in answering this question for digital switch and analog sensor inputs is described. We first present a model of a real-time device driver that enables the driver to execute as its own thread of control, independent of whatever control task is using the data. We then present an engineering approach towards determining a good sample rate for reading digital switches and analog sensors that provide continuous data. Rather than providing a single value for the sample rate, ranges that are based on application parameters are derived analytically. An application designer can use these equations to quickly determine the minimum and maximum sampling rates for their device driver task