Table 5: Summary of performed experiments and results.

The testHow it is tested?The main results

Rotation (Section 4.1)Rotations and movement in 6D using preprogrammed path performed at different velocities.All algorithms perform similarly well in the standard test without any bias. DCM-IMU was slightly precise compared to other algorithms.

Acceleration (Section 4.2)Linear movement separately to , , and directions at different velocities.DCM-IMU is robust against rapid nongravitational accelerations compared to state-of-the-art algorithms.

Biases with rotation (Section 4.3)Rotation test is performed with an artificially added gyroscope bias in the measurement data.Only DCM-IMU can accurately estimate gyroscope biases independent of the amount of added bias.

Biases with linear acceleration and no rotations (Section 4.3)Acceleration test is performed with an artificially added gyroscope bias in the measurement data.DCM-IMU can estimate gyroscope biases for and axes. The gyroscope bias around -axis is unobservable in this case, but the proposed filter can handle the observability issue.

Temperature (Section 4.4)Gyroscope and accelerometer are measured as a function of temperature while it is changed.Gyroscope bias estimates change nearly by 0.5 deg./s as the temperature is changed by 20°C. Similarly, accelerometer readings change considerably.