Abstract
The uncertainty in the end-effector acceleration and angular velocity estimation from measured joint angles of a robotic manipulator is investigated by means of Monte Carlo simulations. The process measurement model includes various sources of errors such as non-uniform sampling rate, additive measurement noise and quantization associated with relative encoders. It is shown that, if some precautions are not taken, the idea of using a robotic manipulator to characterize Inertial Measurement Units (IMUs) can be jeopardized due to excessive uncertainty. A methodology is proposed aiming to minimize the influence of these error sources in the final estimation.