A deep learning based robot positioning error compensation

Sellami, Sami; Klimchik, Alexandr

doi:10.1109/nir52917.2021.9666097

Cited by 2 publications

(1 citation statement)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They also used sensitivity analysis to verify the effect of the design variables. Sellami and Klimchik (2021) introduced deep learning methods to high-precision positioning and control of robots. They first modeled identifiable error sources using traditional methods and then used deep learning methods to identify non-geometric errors that are difficult to model, such as linkage flexibility and gear clearance, to further extend the applicability of the algorithm and thus reduce the absolute robot position error.…”

Section: Introductionmentioning

confidence: 99%

Assembly precision design for parallel robotic mechanism based on uncertain hybrid tolerance allocation

Xu¹,

Tao²,

Gao³

et al. 2023

RIA

View full text Add to dashboard Cite

Purpose The coupling impact of hybrid uncertain errors on the machine precision is complex, as a result of which the designing method with multiple independent error sources under uncertainties remains a challenge. For the purpose of precision improvement, this paper focuses on the robot design and aims to present an assembly precision design method based on uncertain hybrid tolerance allocation (UHTA), to improve the positioning precision of the mechanized robot, as well as realize high precision positioning within the workspace. Design/methodology/approach The fundamentals of the parallel mechanism are introduced first to implement concept design of a 3-R(4S) &3-SS parallel robot. The kinematic modeling of the robot is carried out, and the performance indexes of the robot are calculated via Jacobian matrix, on the basis of which, the 3D spatial overall workspace can be quantified and visualized, under the constraints of limited rod, to avoid the singular position. The error of the robot is described, and a probabilistic error model is hereby developed to classify the hybrid error sensitivity of each independent uncertain error source by Monte Carlo stochastic method. Most innovatively, a methodology called UHTA is proposed to optimize the robot precision, and the tolerance allocation approach is conducted to reduce the overall error amplitude and improve the robotized positioning precision, on the premise of not increasing assembly cost. Findings The proposed approach is validated by digital simulation of medical puncture robot. The experiment highlights the mathematical findings that the horizontal plane positioning error of the parallel robotic mechanism can be effectively reduced after using UHTA, and the average precision can be improved by up to 39.54%. Originality/value The originality lies in UHTA-based precision design method for parallel robots. The proposed method has widely expanding application scenarios in industrial robots, biomedical robots and other assembly automation fields.

show abstract

Section: Introductionmentioning

confidence: 99%