An operational calibration approach of industrial robots through a motion capture system and an artificial neural network ELM

Gao, Tianchi; Meng, Fan; Zhang, Xiaoyu; Zhong, Tian; Song, Hanwen

doi:10.1007/s00170-023-10856-w

Cited by 4 publications

(2 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, with the rise of deep learning algorithms such as artificial intelligence (Gao et al , 2023), it provides an opportunity to improve the accuracy of robot torque prediction using its powerful nonlinear fitting ability (Wang et al , 2020). Deep learning is studying the internal rules and hierarchical representation of sample data, forming more abstract high-level representations of attribute categories or features by combining low-level features, to discover the distributed feature representations of data (Schmidhuber, 2014).…”

Section: Introductionmentioning

confidence: 99%

Joint torque prediction of industrial robots based on PSO-LSTM deep learning

Xiao,

Fu,

Wang

et al. 2024

View full text Add to dashboard Cite

Purpose Because of the key role of joint torque in industrial robots (IRs) motion performance control and energy consumption calculation and efficiency optimization, the purpose of this paper is to propose a deep learning torque prediction method based on long short-term memory (LSTM) recurrent neural networks optimized by particle swarm optimization (PSO), which can accurately predict the the joint torque. Design/methodology/approach The proposed model optimized the LSTM with PSO algorithm to accurately predict the IRs joint torque. The authors design an excitation trajectory for ABB 1600–10/145 experimental robot and collect its relative dynamic data. The LSTM model was trained with the experimental data, and PSO was used to find optimal number of LSTM nodes and learning rate, then a torque prediction model is established based on PSO-LSTM deep learning method. The novel model is used to predict the robot’s six joint torque and the root mean error squares of the predicted data together with least squares (LS) method were comparably studied. Findings The predicted joint torque value by PSO-LSTM deep learning approach is highly overlapped with those from real experiment robot, and the error is quite small. The average square error between the predicted joint torque data and experiment data is 2.31 N.m smaller than that with the LS method. The accuracy of the novel PSO-LSTM learning method for joint torque prediction of IR is proved. Originality/value PSO and LSTM model are deeply integrated for the first time to predict the joint torque of IR and the prediction accuracy is verified.

show abstract

Section: Introductionmentioning

confidence: 99%

Joint torque prediction of industrial robots based on PSO-LSTM deep learning

Xiao,

Fu,

Wang

et al. 2024

View full text Add to dashboard Cite

show abstract

“…Liu et al [26] established an error prediction model based on BPNN (Back propagation neural network), which can accurately and efficiently predict the pose errors over the entire workspace of the robot by a small number of measurement configurations. Gao et al [27] adopted the Extreme Learning Machine (ELM) neural network for compensation of non-geometric errors that are difficult or impossible to model correctly or completely. Ma et al [28] proposed an incremental extreme learning machine based robot error prediction model and improved the prediction accuracy of the prediction model by means of an improved sparrow search algorithm to improve the absolute positioning accuracy of the robot.…”

Section: Introductionmentioning

confidence: 99%

Positioning error compensation method for industrial robots based on stacked ensemble learning

Chen,

Zhang,

et al. 2024

Preprint

View full text Add to dashboard Cite

Due to the advantages of low cost, high flexibility and large workspace, industrial robot has been considered to be the most promising plan to replace traditional CNC machine tool. However, the low absolute positioning accuracy of robot is a key factor that restricts further application in high-precision metal cutting scenarios. In order to improve the absolute positioning accuracy of robot, a positioning error compensation method based on the stacked ensemble learning is proposed. Firstly, the sources of positioning errors and compensation strategies are clarified by analyzing the kinematic model and structural composition of industrial robot. Then, based on the stacked ensemble learning algorithm, robot positioning error prediction model containing multi-layer learners is constructed. And a discrete grid optimization method is presented for model hyper-parameters optimization calculation. Next, predicted positioning errors are adopted to the realize the positioning compensation by offline compensation method. Finally, by set up a robotic milling platform based on MOTOMAN ES165D robot, a series of error compensation experiments have been implemented to verify the proposed method. After compensation, the maximum absolute position error and average position error have decreased by 83% and 89% respectively in the compensation experiments of single point. Moreover, the error compensation of the end milling experiments has also brought significant accuracy improvement, which proved the effectiveness of the proposed method in robotic machining.

show abstract

A novel approach for robot calibration based on measurement sub-regions with comparative validation

Toquica,

S.T. Motta

2024

Int J Adv Manuf Technol

View full text Add to dashboard Cite

An operational calibration approach of industrial robots through a motion capture system and an artificial neural network ELM

Cited by 4 publications

References 32 publications

Joint torque prediction of industrial robots based on PSO-LSTM deep learning

Joint torque prediction of industrial robots based on PSO-LSTM deep learning

Positioning error compensation method for industrial robots based on stacked ensemble learning

A novel approach for robot calibration based on measurement sub-regions with comparative validation

Contact Info

Product

Resources

About