Highly accurate contour control of an articulated robot manipulator using a Gaussian neural network

Purpose – Aims to realize the high accurate contour control with high‐speed motion of articulated robot manipulator (ARM) with interference. Design/methodology/approach – Proposes a new contour control method by using Gaussian neural network (GNN) to solve the problem of the deterioration of the contour control performance due to the interference between robot links. The construction of the GNN controller and the approximation of the interference are based on the Euler‐Lagrange model of ARM. The actual input/out data about the motion of ARM are used for training the GNN to accurately represent the inverse dynamics of ARM with interference. With the Lyapunov function, the stability and the robustness of the GNN controller are discussed. Through the simulation and experiment, it verified that the precision of the contour control has been improved, and illustrated the good features of the proposed method. Findings – Finds that the actual data about the motion of ARM, which is easily obtained from the working field, can express the real features of ARM, and the GNN controller can improve the precision of the contour control with good features. Practical implications – The proposed method provides an effective method for realizing high accurate contour control of ARM with interference. It can be extended to the ARMs with more than two links and concerning more factors affecting the precision of the contour control, such as friction or gravity. Originality/value – Proposes a new GNN controller for realizing high accurate contour control of ARM with interference, which is significant for industry.