Accuracy describe the closeness of the measured value to the true value. In a theoretical perfect world the output signal for a proportional position sensor shall be the exact representation of the position i.e. the output shall be an ideal straight-line. But in reality different types of accuracy errors occur.
Linearity or non-linearity is the deviation between the ideal straight-line and the real output signal. Linearity is expressed in absolute numbers; millimeters [mm], milliampere [mA] etc., or relative to the full scale [FS] ex 0,05% FS.
Hysteresis is the deviation on signal characteristics between increasing output and decreasing output. Hysteresis can be seen as an error in proportional position sensors but is used as a feature in many switching applications. For example, proximity sensors and thermostats uses hysteresis to avoid noise (rapid, repeated switching on/off) at the switching point.
A position sensor shall present the same output characteristics between repeated strokes. If the characteristics change between strokes there is a repeatability error. Linearity and hysteresis error are systematic and don’t change from one stroke to another and can therefore be compensated for with calibration. Repeatability is random always changing and is very difficult to calibrate.
Resolution is the scale of the smallest change a sensor can detect. If resolution shall be categorized as an accuracy error or sensor feature is an endless discussion. As the measurement is not necessarily wrong, however low resolution can give a too rough representation of the true value. Like a low pixilated photo, it’s not an error just not good enough.