Article: ST-HGCN-enhanced real-time compensation for industrial robot positioning errors Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.39 pp.37 - 51 Abstract: Industrial robots frequently exhibit degraded positioning accuracy under dynamic coupling effects and environmental perturbations. To mitigate this, we introduce a real-time compensation framework powered by a spatio-temporal hybrid graph convolutional network (ST-HGCN). The methodology constructs a unified model integrating spatial sensor dependency graphs with temporal error propagation chains, utilising high-precision ground truth from the European Robotics Challenge (EUROC) micro aerial vehicle (MAV) dataset combined with inertial measurement unit (IMU) data. Experimental validation demonstrates a 62.3% reduction in root mean square positioning error (RMSE) relative to conventional graph-convolutional long short-term memory (LSTM) networks during complex multi-axis trajectories, while sustaining compensation latency under 2 ms. This work establishes a novel data-driven paradigm for high-precision robotic control, with direct applicability to precision manufacturing and flexible assembly operations requiring micron-level accuracy. Inderscience Publishers - linking academia, business and industry through research

Title: ST-HGCN-enhanced real-time compensation for industrial robot positioning errors

Authors: Mingxiong Wu

Addresses: College of Intelligent Manufacturing Engineering, Zhanjiang University of Science and Technology, Zhanjiang 524086, China

Abstract: Industrial robots frequently exhibit degraded positioning accuracy under dynamic coupling effects and environmental perturbations. To mitigate this, we introduce a real-time compensation framework powered by a spatio-temporal hybrid graph convolutional network (ST-HGCN). The methodology constructs a unified model integrating spatial sensor dependency graphs with temporal error propagation chains, utilising high-precision ground truth from the European Robotics Challenge (EUROC) micro aerial vehicle (MAV) dataset combined with inertial measurement unit (IMU) data. Experimental validation demonstrates a 62.3% reduction in root mean square positioning error (RMSE) relative to conventional graph-convolutional long short-term memory (LSTM) networks during complex multi-axis trajectories, while sustaining compensation latency under 2 ms. This work establishes a novel data-driven paradigm for high-precision robotic control, with direct applicability to precision manufacturing and flexible assembly operations requiring micron-level accuracy.

Keywords: ST-HGCNs; industrial robotics; localisation error compensation; real-time control; sensor fusion; inertial measurement unit; IMU.

DOI: 10.1504/IJICT.2025.149790

International Journal of Information and Communication Technology, 2025 Vol.26 No.39, pp.37 - 51

Received: 17 Jul 2025
Accepted: 08 Sep 2025
Published online: 12 Nov 2025 *

Title: ST-HGCN-enhanced real-time compensation for industrial robot positioning errors

Keep up-to-date