Definitions of calibration and testing
Let's first clarify the definitions of calibration and testing.

Calibration means obtaining the various parameters of the internal algorithm of the six-dimensional force sensor by loading the load of the theoretical value and recording the corresponding original signal output by the sensor at the same time. That is, establishing a mapping relationship between the original signal of the sensor and the force.

Testing means that by loading the load of the known theoretical true value and recording the sensor measurement results at the same time, the difference between the measurement results and the theoretical true value is counted and compared to obtain the precision and accuracy of the sensor, that is, to test whether the sensor is accurate.

Simply put, calibration is to obtain the sensor firmware parameters, and testing is to obtain the accuracy of the sensor.

Calibration
Let's talk about calibration first. For a six-dimensional force sensor, calibration needs to consider six dimensions at the same time.

If we only consider the precise loading calibration of the sensor in the X-axis direction according to the four steps of ±25%FS, ±50%FS, ±75%FS and ±100%FS (Figure 1). Then the red dots in this figure represent the calibration sample points. Four steps plus no load, and considering the positive and negative directions, there are 9 sample points on one coordinate axis.
If the joint loading calibration of the X-axis and Y-axis is considered at the same time, then not only the sample points on the X-axis and Y-axis must be considered (Figure 2), but also other sample points on the XY plane, which are the yellow points in the figure (Figure 3, the yellow points gradually appear), which are also called cross sample points. Obviously, when considering the joint loading of the sensor on the X and Y axes, the calibrated sample space changes from a one-dimensional straight line to a two-dimensional plane.
Advantages of six-dimensional joint loading calibration
Using so many sample points for six-dimensional joint loading calibration will bring three advantages:

First, the cross sample points can simulate the force conditions of the sensor very close to the actual use conditions.
Secondly, such calibration is convenient for examining the nonlinear mechanical characteristics of the sensor under the simultaneous action of multi-dimensional loads, which can effectively improve the design of the sensor structure.
Finally, this is a calibration based on the nonlinear mechanical characteristics of the sensor, which can greatly optimize the mathematical model of the decoupling algorithm.
In short, the use of six-dimensional force joint loading calibration can make the sensor more accurate and have lower crosstalk. When the six-dimensional force sensor is subjected to forces in multiple dimensions at the same time, the nonlinear characteristics are very significant, and the superposition of six-dimensional linear models cannot accurately describe this nonlinear effect.