A “Look-Backward-and-Forward” Adaptation Strategy for Assessing Parameter Estimation Error of Human Motion Prediction Model

Silu Chen,Yufan Zhu,Yisha Liu,Chi Zhang,Zhongyu Piao,Guilin Yang
DOI: https://doi.org/10.1109/lra.2021.3139952
IF: 5.2
2022-04-01
IEEE Robotics and Automation Letters
Abstract:The prediction of human motion is essential for safe human-robot collaboration (HRC). For existing prediction methods based on adaptive neural network (NN) models, estimation errors (EEs) of model parameters are directly coupled with prior EEs of trajectories. This results in poor assessment of the mean square estimation error (MSEE) of model parameters, which is a potential danger to the safety HRC. In this paper, a novel look-backward-and-forward method is proposed to assess the EEs of model parameters. This method firstly computes EEs of model parameters in the past few time instants based on the actual error of trajectory predicted in the past. Subsequently, the EEs of model parameters are recursively updated till the current time instant. In this way, the trajectory of a single particle and its MSEE are accurately accessed. The direct coupling between the MSEE of the trajectory and that of model parameters is cut off and the mean estimation errors (MEEs) of parameters are only affected by the EEs of the past trajectory at the recent several time instants, rather than accumulating as the time propagates. This method is further extended to estimate the pose trajectory of a rigid body based on fitting of an ellipsoids parameters. The experimental results of predicting 3-D trajectories of human motions show the effectiveness of the proposed method.
robotics
What problem does this paper attempt to address?