Search

This paper propose a non-linear estimator for humanoid robot. The accurate estimation of the center of mass (CoM) of the robot plays an important role to enhance the dynamic locomotion and manipulation of a humanoid robot. However, the estimation of the CoM is difficult because of the non-linear dynamics. To deal with the nonlinearity, the Extended Kalman Filter (EKF) is a common solution used in real time control application. However, one of the disadvantages of EKF is the linearization of the transition and observation function. The Unscented Kalman Filter (UKF) is a derivative-free filter without linearization error. In this paper, a novel method is presented to estimate humanoid robot locomotion with an Unscented Kalman Filter for data fusion with encoders, Inertial Measurement Units (IMU) and Force/Torque (F/T) sensors. The method relies on forward kinematics, Zero Moment Point (ZMP) with the variation of the angular momentum and 5 mass preview control/time-varying divergent component of motion to estimate the locomotion of a robot. By the use of the humanoid robot Renbo in our lab, we indicate the efficiency and the efficacy of the proposed estimator.

In the industrial manufacturing processing, peg-in-hole assembly is the core technique for automation, which usually involves two stages. The first stage is the search phase, which is used to find the center of the hole. Then, the peg is aligned with the hole on the plane in the second stage, known as the insertion phase. However, the slight orientational misalignment between peg and hole makes the geometrical analysis of the peg-in-hole assembly process inaccurate. The peg would be stuck because of orientation errors. To deal with this issue, we present a strategy with a standard process to obtain both the location and orientation information in one stage. Compared with the general blind search approaches, the combination of the two targets can improve the efficiency and robustness for the task. The experiments are tested with a cylindrical peg, which is 6 millimeters in diameter, and the hole having a clearance of H7/g6, which means the maximum clearance is about 30 µm.