Research on robot computed torque control based on DeLaN model
ZHANG Jiaheng
SHI Donghao
YANG Pu
LI Qinchuan
Abstract：[Objective]To address the problems of complex dynamic modeling and difficult parameter identification in robot trajectory tracking control,a computed torque control(CTC)method based on deep Lagrangian network(DeLaN)was proposed.[Methods]Firstly,a DeLaN model considering friction was designed to improve the identification accuracy of robot dynamic parameters.Secondly,a robot excitation trajectory optimization method was proposed,which provided a high-quality dataset for DeLaN model training.Finally,the DeLaN-CTC algorithm was proposed,and relevant derivation proofs were provided.Dynamic model identification tests and trajectory tracking tests of the DeLaN-CTC algorithm were conducted on the Unitree Z1 robotic arm platform.[Results]The test results show that the DeLaN model combined with friction can accurately identify the dynamic parameters of the Z1 robotic arm,and the DeLaN-CTC algorithm has higher trajectory tracking control accuracy than the DeLaN-PD[proportional-derivative(PD)control algorithm combined with DeLaN dynamic model feedforward],verifying its effectiveness in high-precision trajectory tracking tasks.
Keywords：Deep Lagrangian networkExcitation trajectoryComputed torque controlZ1 robotic arm
Publication Date：2026-07-15
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：8( 43-50 )