Task-level time-optimal collision avoidance trajectory planning for grinding manipulators

Diao, Shipu; Chen, Xindu; Wu, Lei; Zhong, Zhanqi; Lin, Zhongqin

doi:10.1177/0954406218793660

Cited by 2 publications

(2 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Where ½N i;k ðuÞ� ðlÞ is the l derivative of N i;k ðuÞ, and B can be calculated by equation (4). From equation ( 9), ½d j 0 ; .…”

Section: Trajectory Constructionmentioning

confidence: 99%

“…Optimization algorithms are the core of solving trajectory optimization problems for manipulators. In recent years, some algorithms have been given by researchers to solve the trajectory optimization problems, [1][2][3][4][5][6][7][8][9] mainly including immune cloning algorithm (ICA), 1 differential evolutionary algorithm (DE), [2][3] simulated annealing algorithm (SA), 4 non-dominated sorting genetic algorithm (NSGA-II), [5][6][7] particle swarm optimization (PSO), [8][9] etc. MOPSO has the following advantages in solving optimization problems.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Trajectory optimization for manipulators based on external archives self-searching multi-objective particle swarm optimization

Liu

Zhang

2021

Proceedings of the Institution of Mechanical Engineers, Part C:

View full text Add to dashboard Cite

In order to improve the quality of the non-inferior solutions obtained by multi-objective particle swarm optimization (MOPSO), an improved algorithm called external archives self-searching multi-objective particle swarm optimization (EASS-MOPSO) was proposed and applied to a multi-objective trajectory optimization problem for manipulators. The position curves of joints were constructed by using quartic B-splines; the mathematical models of time, energy and jerk optimization objectives for manipulators were established; and the kinematic constraints of joints were transformed into the constraints of the control vertexes of the B-splines. A self-searching strategy of external archives to make non-inferior solutions have the ability to search the surrounding hyperspace was explored, and a diversity maintaining strategy of the external archives was proposed. The results of several test functions by simulation show that the convergence and diversity of the proposed algorithm are better than those of other 4 selected algorithms; the results of the trajectory optimization problem for manipulators by simulation show that the convergence, diversity and time consumption of the proposed algorithm are significantly better than those of non-dominated sorting genetic algorithm.

show abstract

“…Where ½N i;k ðuÞ� ðlÞ is the l derivative of N i;k ðuÞ, and B can be calculated by equation (4). From equation ( 9), ½d j 0 ; .…”

Section: Trajectory Constructionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Trajectory optimization for manipulators based on external archives self-searching multi-objective particle swarm optimization

Liu

Zhang

2021

Proceedings of the Institution of Mechanical Engineers, Part C:

View full text Add to dashboard Cite

show abstract

Cervical cell classification based on strong feature CNN-LSVM network using Adaboost optimization

Jia

Zhang

et al. 2023

IFS

View full text Add to dashboard Cite

Cervical cancer is one of the most common causes of death in women in the world, and early screening is an effective means of diagnosis and treatment, which can greatly improve the survival rate. Cervical cell classification model is an effective means to assist screening. However, the existing single model, including CNNs and machine learning methods, still has shortcomings such as unclear feature meaning, low accuracy and insufficient supervision. To solve the shortcomings of a single model, a novel framework based on strong feature Convolutional Neural Networks (CNN)-Lagrangian Support Vector Machine (LSVM) model is proposed for the accurate classification of cervical cells. Strong features extracted by hybrid methods are fused with the abstract ones from hidden layers of LeNet-5, then the fused features are processed with dimension reduction and fed into the LSVM classifier optimized by Adaboost for classification. Proposed model is evaluated using the augmented Herlev and private dataset with the metrics including accuracy (Acc), sensitivity (Sn), and specificity (Sp), which outperformed the baselines and state-of-the-art approaches with the Acc of 99.5% and 94.2% in 2&7-class classification, respectively.

show abstract

Task-level time-optimal collision avoidance trajectory planning for grinding manipulators

Cited by 2 publications

References 34 publications

Trajectory optimization for manipulators based on external archives self-searching multi-objective particle swarm optimization

Trajectory optimization for manipulators based on external archives self-searching multi-objective particle swarm optimization

Cervical cell classification based on strong feature CNN-LSVM network using Adaboost optimization

Contact Info

Product

Resources

About