Whole-Body Joint Angle Estimation for Real-Time Humanoid Robot Imitation Based on Gaussian Process Dynamical Model and Particle Filter

Abstract

Real-time imitation enables a humanoid robot to mirror the behavior of humans, being important for applications of human–robot interaction. For imitation, the corresponding joint angles of the humanoid robot should be estimated. Generally, a humanoid robot comprises dozens of joints that construct a high-dimensional exploration space for estimating the joint angles. Although a particle filter can estimate the robot state and provides a solution for estimating joint angles, the computational cost becomes prohibitive given the high dimension of the exploration space. Furthermore, a particle filter can only estimate the joint angles accurately using a motion model. To realize accurate joint angle estimation at low computational cost, Gaussian process dynamical models (GPDMs) can be adopted. Specifically, a compact state space can be constructed through the GPDM learning of high-dimensional time-series motion data to obtain a suitable motion model. We propose a GPDM-based particle filter using a compact state space from the learned motion models to realize efficient estimation of joint angles for robot imitation. Simulations and real experiments demonstrate that the proposed method efficiently estimates humanoid robot joint angles at low computational cost, enabling real-time imitation.

Simulation research on the reappearance technology of manipulator trajectory based on Marker ball

“…The simulation experiment is the motion simulation of the manipulator when the manipulator moves under the motion capture system. Due to the deviation of the collected data or the excessive movement range of the human arm in the process of collection, these reasons may lead to the situation of singularity in the action repetition of the robotic arm, which may cause damage to the robotic arm [13] . Therefore, before using the actual robotic arm for trajectory repetition, simulation experiments need to be carried out on the simulation software to verify the trajectory acquisition and recurrence method adopted in this experiment, and whether the collected data information is feasible [14] ..…”

Action Planning and Design of Humanoid Robot Based on Sports Analysis in Digital Economy Era

“…(2) Inverse kinematics There are two methods to solve inverse kinematics, one is numerical method and the other is analytical method [24][25]. The analytic method is limited by the mechanical structure of the specific robot.…”

Analysis of Cost Functions for Reinforcement Learning of Reaching Tasks in Humanoid Robots

“…This approach empowers non-experts to program robots according to their needs. The first objective in LfD for humanoid robots is to generate robot motions that accurately reproduce the demonstrated motions while preserving the robot's stability, a process hampered by the correspondence problem arising from the differing kinematic and dynamic characteristics of humans and humanoid robots [6][7][8][9][10][11]. To facilitate the successful performance of the desired tasks, it is necessary to adapt the initially generated motions by also considering task constraints besides the robot's physical constraints [12].…”