Sensor Accuracy

Sensors have varying levels of accuracy, which quantifies how close a measured value is to the actual value. This is measurement error:

$$\text{Measurement error}=\text{Actual Value}-\text{Measured Value}$$

Measurement accuracy describes the measurement error guaranteed by the manufacturer:

$$\text{Actual Value is Within Measure Value }\pm \text{Measurement Accuracy}$$

So, the actual measurement error of a particular device should not exceed the manufacturer-specified measurement accuracy. This is usually expressed as:

Here are the different definitions:

• “Full-scale” is the max value at $90°$
• “Instrument span” is the max range at $180°$

If the manufacturer said the measurement accuracy was $\pm 0.2\text{ mm}$, anything between 9.8 mm to 10.2 mm is okay.

The system accuracy considers all the individual sensor and system errors to determine overall accuracy.

• Care must be taken in assessing how individual errors contribute to the overall accuracy. In some cases the errors add, whereas in others the contribution is root-sum-squared.

Worse case analysis: Choose extremes to make the system as bad as possible. Many unknowns: Monte Carlo analysis. Each random variable is sampled, run simulation, plot all results. Gives family of curves that show possible results.