A convolution neural network based semi-parametric dynamic model for industrial robot

2021 ◽  
pp. 095440622110398
Author(s):  
Chungang Zhuang ◽  
Yihui Yao ◽  
Yichao Shen ◽  
Zhenhua Xiong
Keyword(s):  
Neural Network ◽  
Dynamic Model ◽  
Dynamic Models ◽  
Industrial Robot ◽  
Friction Model ◽  
Rigid Body Dynamics ◽  
Convolution Neural Network ◽  
Joint Torque ◽  
Average Error ◽  
Proprioceptive Sensors

Robot dynamic model is widely applied to control, collision detection and motion planning. Accurate dynamic model can achieve better performance for the above applications. Traditional dynamic models have several limitations, such as the complex hypotheses for friction model and the requirement of additional joint torque sensors. This article constructs a convolution neural network (CNN) based semi-parametric dynamic (SPD) model by only using the motor encoder signals and motor currents. The SPD model not only contains the physically feasible parameters but also compensates the dynamic model by CNN. The parametric and non-parametric parts constitute the SPD model. A lightweight CNN is proposed to simultaneously ensure the accuracy and computational efficiency. To effectively train the CNN model, a dataset generation method, which expands the excitation trajectory and only uses a continuous trajectory to record data, is proposed. The CNN-based SPD model is verified on a 6-DoF laboratory-developed industrial robot only with the proprioceptive sensors. Compared with the traditional rigid body dynamics (RBD) model, the average error of the CNN-based SPD model is reduced by 9.23% in terms of the experimental results. Meanwhile, the proposed CNN-based method achieves better performance than other supervised methods.

Research on mechanism configuration and dynamic characteristics for multi-split transmission line mobile robot

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Wei Jiang ◽  
Yating Shi ◽  
Dehua Zou ◽  
Hongwei Zhang ◽  
Hong Jun Li
Keyword(s):  
Mobile Robot ◽  
Transmission Line ◽  
Dynamic Model ◽  
Dynamic Characteristics ◽  
Transmission Lines ◽  
Dynamic Models ◽  
Working Environment ◽  
Joint Torque ◽  
Content Type ◽  
Working Space

Purpose The purpose of this paper is to achieve the optimal system design of a four-wheel mobile robot on transmission line maintenance, as the authors know transmission line mobile robot is a kind of special robot which runs on high-voltage cable to replace or assist manual power maintenance operation. In the process of live working, the manipulator, working end effector and the working environment are located in the narrow space and with heterogeneous shapes, the robot collision-free obstacle avoidance movement is the premise to complete the operation task. In the simultaneous operation, the mechanical properties between the manipulator effector and the operation object are the key to improve the operation reliability. These put forward higher requirements for the mechanical configuration and dynamic characteristics of the robot, and this is the purpose of the manuscript. Design/methodology/approach Based on the above, aiming at the task of tightening the tension clamp for the four-split transmission lines, the paper proposed a four-wheel mobile robot mechanism configuration and its terminal tool which can adapt to the walking and operation on multi-split transmission lines. In the study, the dynamic models of the rigid robot and flexible transmission line are established, respectively, and the dynamic model of rigid-flexible coupling system is established on this basis, the working space and dynamic characteristics of the robot have been simulated in ADAMS and MATLAB. Findings The research results show that the mechanical configuration of this robot can complete the tightening operation of the four-split tension clamp bolts and the motion of robot each joint meets the requirements of driving torque in the operation process, which avoids the operation failure of the robot system caused by the insufficient or excessive driving force of the robot joint torque. Originality/value Finally, the engineering practicability of the mechanical configuration and dynamic model proposed in the paper has been verified by the physical prototype. The originality value of the research is that it has double important theoretical significance and practical application value for the optimization of mechanical structure parameters and electrical control parameters of transmission line mobile robots.

scholarly journals Establishing an Improved Kane Dynamic Model for the 7-DOF Reconfigurable Modular Robot

2017 ◽  
Vol 2017 ◽  
pp. 1-13 ◽  
Author(s):  
Xiao Li ◽  
Hanxu Sun ◽  
Linjing Liao ◽  
Jingzhou Song
Keyword(s):  
Dynamic Model ◽  
Dynamic Models ◽  
Simulation Experiment ◽  
Industrial Robot ◽  
Coefficient Matrix ◽  
Modular Robot ◽  
Analysis Model ◽  
Explicit Formulation ◽  
Undetermined Coefficient ◽  
Improved Function

We propose an improved Kane dynamic model theory for the 7-DOF modular robot in this paper, and the model precision is improved by the improved function T′it. We designed three types of progressive modular joints for reconfigurable modular robot that can be used in industrial robot, space robot, and special robot. The Kane dynamic model and the solid dynamic model are established, respectively, for the 7-DOF modular robot. After that, the experimental results are obtained from the simulation experiment of typical task in the established dynamic models. By the analysis model of error, the equation of the improved torque T′it is derived and proposed. And the improved Kane dynamic model is established for the modular robot that used T′it. Based on the experimental data, the undetermined coefficient matrix is five-order linear that was proved in 7-DOF modular robot. And the explicit formulation is solved of the Kane dynamic model and can be used in control system.

Analysis of Rigid-Body Dynamic Models for Simulation of Systems With Frictional Contacts

2000 ◽  
Vol 68 (1) ◽  
pp. 118-128 ◽  
Author(s):  
P. Song ◽  
P. Kraus ◽  
V. Kumar ◽  
P. Dupont
Keyword(s):  
Rigid Body ◽  
Coulomb Friction ◽  
Dynamic Models ◽  
Stable Solution ◽  
Friction Model ◽  
Rigid Body Dynamics ◽  
Frictional Contacts ◽  
Friction Law ◽  
Rigid Body Dynamic ◽  
Body Dynamic

The use of Coulomb’s friction law with the principles of classical rigid-body dynamics introduces mathematical inconsistencies. Specifically, the forward dynamics problem can have no solutions or multiple solutions. In these situations, compliant contact models, while increasing the dimensionality of the state vector, can resolve these problems. The simplicity and efficiency of rigid-body models, however, provide strong motivation for their use during those portions of a simulation when the rigid-body solution is unique and stable. In this paper, we use singular perturbation analysis in conjunction with linear complementarity theory to establish conditions under which the solution predicted by the rigid-body dynamic model is stable. We employ a general model of contact compliance to derive stability criteria for planar mechanical systems. In particular, we show that for cases with one sliding contact, there is always at most one stable solution. Our approach is not directly applicable to transitions between rolling and sliding where the Coulomb friction law is discontinuous. To overcome this difficulty, we introduce a smooth nonlinear friction law, which approximates Coulomb friction. Such a friction model can also increase the efficiency of both rigid-body and compliant contact simulation. Numerical simulations for the different models and comparison with experimental results are also presented.

An Extended Kalman Filter for Collision Detection During Manipulator Contact Tasks

2014 ◽  
Author(s):  
Andy Zelenak ◽  
Mitch Pryor ◽  
Kyle Schroeder
Keyword(s):  
Kalman Filter ◽  
Dynamic Model ◽  
Extended Kalman Filter ◽  
Collision Detection ◽  
Force Feedback ◽  
Industrial Robot ◽  
Detection Algorithm ◽  
Industrial Robots ◽  
Joint Torque ◽  
Contact Tasks

The development of control strategies that allow stiff industrial robots to operate safely in unstructured environments is a significant challenge. This paper integrates two strategies that improve safety for industrial manipulators in uncertain conditions. First, software compliance in the task space is implemented using force feedback. End-effector compliance is vital for many tasks, such as interacting with humans or manipulating uncertain payloads. Beyond compliance, a collision detection algorithm detects collisions based on joint torque deviation from a dynamic model. Collisions can be detected at any point along the manipulator via loading or impulse anomalies. Joint torque data is typically noisy, and the accuracy of the robot dynamic model is limited, so an Extended Kalman Filter (EKF) was developed to improve the torque estimates. Experiments and demonstrations were performed using a commercially available 7DOF industrial robot. The EKF improved collision detection during unplanned contact tasks, and the method described here is hardware agnostic and extensible.

A Predictive Control Based on Neural Network for Dynamic Model of Proton Exchange Membrane Fuel Cell

2013 ◽  
Vol 10 (3) ◽  
Author(s):  
M. Rezaei ◽  
M. Mohseni
Keyword(s):  
Neural Network ◽  
Fuel Cell ◽  
Dynamic Model ◽  
Proton Exchange Membrane ◽  
Dynamic Models ◽  
Adaptive Controller ◽  
Proton Exchange ◽  
Predictive Controller ◽  
Exchange Membrane ◽  
Electrochemical Model

This paper presents the development of dynamic models for proton exchange membrane fuel cells (PEMFC). The PEMFC control system has an important effect on operation of cell. Traditional controllers could not lead to acceptable responses because of time-change, long-hysteresis, uncertainty, strong-coupling and nonlinear characteristics of PEMFCs, This paper presents a dynamic model for PEMFC system, so an intelligent or adaptive controller is needed. In this paper, a neural network predictive controller have been designed to control the voltage of at the presence of fluctuations of temperature. The results of implementation of this designed NN Predictive controller on a dynamic electrochemical model of a small size 5 KW, PEM fuel cell have been simulated by matlab/SIMULINK.

scholarly journals 3D Large-Pose Face Alignment Method Based on the Truncated Alexnet Cascade Network

2020 ◽  
Vol 2020 ◽  
pp. 1-8
Author(s):  
Qian Zhang ◽  
Hao Zheng ◽  
Tao Yan ◽  
Jiehui Li
Keyword(s):  
Neural Network ◽  
Convolution Neural Network ◽  
Average Error ◽  
Face Alignment ◽  
Alignment Method ◽  
Deep Convolution Neural Network ◽  
A Current

Aiming at the low accuracy of large-pose face alignment, a cascade network based on truncated Alexnet is designed and implemented in the paper. The parallel convolution pooling layers are added for concatenating parallel results in the original deep convolution neural network, which improves the accuracy of the output. Sending the intermediate parameter which is the result of each iteration into CNN and iterating repeatedly to optimize the pose parameter in order to get more accurate results of face alignment. To verify the effectiveness of this method, this paper tests on the AFLW and AFLW2000-3D datasets. Experiments on datasets show that the normalized average error of this method is 5.00% and 5.27%. Compared with 3DDFA, which is a current popular algorithm, the accuracy is improved by 0.60% and 0.15%, respectively.

A fuzzy neural network approach for modeling the growth kinetics of FeB and Fe2B layers during the boronizing process

2018 ◽  
Vol 106 (6) ◽  
pp. 603 ◽  
Author(s):  
Bendaoud Mebarek ◽  
Mourad Keddam
Keyword(s):  
Neural Network ◽  
Experimental Data ◽  
Fuzzy Neural Network ◽  
Treatment Time ◽  
Calculation Model ◽  
Average Error ◽  
Network Approach ◽  
Neural Network Approach ◽  
Fuzzy Neural ◽  
Kinetics Of

In this paper, we develop a boronizing process simulation model based on fuzzy neural network (FNN) approach for estimating the thickness of the FeB and Fe2B layers. The model represents a synthesis of two artificial intelligence techniques; the fuzzy logic and the neural network. Characteristics of the fuzzy neural network approach for the modelling of boronizing process are presented in this study. In order to validate the results of our calculation model, we have used the learning base of experimental data of the powder-pack boronizing of Fe-15Cr alloy in the temperature range from 800 to 1050 °C and for a treatment time ranging from 0.5 to 12 h. The obtained results show that it is possible to estimate the influence of different process parameters. Comparing the results obtained by the artificial neural network to experimental data, the average error generated from the fuzzy neural network was 3% for the FeB layer and 3.5% for the Fe2B layer. The results obtained from the fuzzy neural network approach are in agreement with the experimental data. Finally, the utilization of fuzzy neural network approach is well adapted for the boronizing kinetics of Fe-15Cr alloy.

Sign in / Sign up

Export Citation Format

Share Document