*5.2. Results Using the Inertial Sensor*

The data obtained from the inertial sensor show more accurately the real behavior of the end of the soft joint.

The inclination results in Test 1 show that the position of the joint does not reach the reference inclination. The 90◦ orientation test (downwards, in the sense of gravity) is the one that presents lower errors when tracking the reference. The tests for 0◦ and 180◦ orientation angles show higher tracking errors, as shown in the attached video. The kinematics designed for these positions assumes that the length of tendon 1 (lower motor) should not change. These theoretical results, when taken to the experimental field, are not fulfilled because the tendons are not perfectly tensioned, and the two upper wires cause the position rise. This rise is reflected in the orientation that has a negative phase shift when the reference is 0◦ and a positive phase shift when the reference is 180◦.

We also observed that the orientation results do not reach the zero position when the reference is zero. This is because it is difficult to move the orientation to zero due to the fact that the inclination is not exactly zero when returning to the resting position, as the inclination graphs show. This causes a slight inclination while maintaining the same orientation. As discussed above for the encoder data, orientation is very sensitive to inclination.

For the sensor results in Test 2, the inclination graphs show how the experimental inclination does not reach the reference value. However, it should be noted that it has a sinusoidal behavior over time. As in the previous test, the reason for both is that the theoretical behavior of the joint is not the same as the real behavior, because the model assumes aspects such as a continuous curvature, and because there are also other influencing mechanical aspects, such as the precision in the tendon length or the tendon winding in the winches.

This undulatory behavior is observed again in the orientation graphs. However, it can be seen that for angles 90◦, 210◦ and 330◦ the orientation does not vary, which coincides with the vertices of the soft joint morphology. For these angles, the inclination is maximum. Moreover, when one of the vertices is passed, the opposite tendons cause the variation of orientation, and it takes a little time to change from unwinding to rewinding. This can be seen in the attached videos for this test.
