Optimizing compliant gripper mechanism design by employing an effective bi-algorithm: fuzzy logic and ANFIS

TECHNICAL PAPER

Optimizing compliant gripper mechanism design by employing

an effective bi-algorithm: fuzzy logic and ANFIS

Tien V. T. Nguyen1,2 • Ngoc-Thai Huynh3 • Ngoc-Chien Vu1,4 • Vu N. D. Kieu1,5 • Shyh-Chour Huang1

Received: 22 October 2020 / Accepted: 18 November 2020

Springer-Verlag GmbH Germany, part of Springer Nature 2021

Abstract

This investigation confronts the note-worthy improvement configuration gap in which such a design method could be better

focused on the multi-objective optimization design of a compliant gripper mechanism as a robot arm through an effective

hybrid algorithm of fuzzy logic and adaptive neuro-fuzzy inference system (ANFIS). We found that the proposed bi￾algorithm approach is more compelling than theoretical ideas like auxiliary shape changes, materials, and directors of

mechanisms when designing the compliant gripper mechanism with a set of novel multi-objective optimization design

recently. In particular, it explores whether the compliant gripper mechanism shapes affect picking things up. In this unique

study, we considered displacement values and the frequency values as response parameters during the simulation and the

optimization design process. To test the effectiveness of the optimal design method, we proposed an initial compliant

gripper mechanism carried out through the numerically experimental matrix—the Box–Behnken design. After that, we

simulated the numerical model by utilizing the finite element method incorporating the approaches of desirability function,

fuzzy logic system, and ANFIS. The results turn larger than those of the previous approaches. Moreover, numerical results

reveal that the suggested hybrid method has a computational exactness more conspicuous than that of the Taguchi method.

In short, the principle accomplishments with variables to the compliant gripper mechanism optimization design can be

summarized up as follows: (i) the promising and potential proposed approach could meet the clients’ prerequisite, (ii) the

idea of multi-objective optimization design ought to be re-considered when designing compliant gripper mechanism as

well as applying related designing fields at the diminished expenses and the shortage time.

Abbreviations

2D Two-dimension

3D Three-dimension

ANFIS Adaptive neuro-fuzzy inference system

ANOVA Analysis of variance

BCM Beam constraint model

CGM Compliant gripper mechanism

DOF Degree of freedom

FEA Finite element analysis

FEM Finite-element-method

FIS Fuzzy inference systems

DFA Design for assembly

FLS Fuzzy logic system

& Shyh-Chour Huang

[email protected]

Tien V. T. Nguyen

[email protected]

1 Mechanical Engineering Department, National Kaohsiung

University of Science and Technology, No. 415, Jiangong

Rd., Sanmin Dist., Kaohsiung 80778, Taiwan

2 Faculty of Mechanical Technology, Industrial University of

Ho Chi Minh City, No. 12 Nguyen Van Bao St., Go Vap

Dist., Ho Chi Minh City 70000, Vietnam

3 Faculty of Automobile Engineering Technology, Industrial

University of Ho Chi Minh City, No. 12 Nguyen Van Bao St.,

Go Vap Dist., Ho Chi Minh City 70000, Vietnam

4 Mechanical Engineering Faculty, Nha Trang University, No.

2 Nguyen Dinh Chieu St., Nha Trang City,

Khanh Hoa Province 65000, Vietnam

5 Faculty of Automotive Mechanical and Electrical &

Electronics Engineering, Nguyen Tat Thanh University,

No.300A, Nguyen Tat Thanh St., Ward 13, Dist. 4,

Ho Chi Minh City 70000, Vietnam

123

Microsystem Technologies

https://doi.org/10.1007/s00542-020-05132-w(0123456789().,-volV)(0123456789().,-volV)

GRA Grey relational analysis

TMFL Taguchi method and fuzzy logic

LAPO Lightning attachment procedure optimization

MOO Multi-objective optimization

MPCI Multi-characteristic performance index

SR Scott-Russell mechanism

TLBO Teaching learning-based optimization

algorithm

1 Introduction

The improvement of widespread grippers ready to pick

diverse objects of generally differing shapes and surfaces is

an extremely practical task (Huang et al. 2019; Liu et al.

2018). An inactively hybrid algorithm including fuzzy

logic and ANFIS is to acquire the gripper which could

oblige to any sporadic and delicate getting a handle on

objects which was never deeply investigated in the last

several decades (Chau et al. 2018, 2019; Le Chau et al.

2020). The motivation behind the under-actuation is to

utilize the intensity of one actuator to drive the open and

close movement of the gripper. Even though the com￾pletely compliant gripper has one or numerous freedom

degrees under consideration and application of the multi￾objective optimization design (Nguyen et al. 2019; Dao

et al. 2018; Ho et al. 2019), there are many drawbacks and

no effects. This paper presents another approach (Ai and

Xu 2014; Chattaraj et al. 2016) of the versatile compliant

gripper with appropriated consistency. The kinematic￾based technique (Ai and Xu 2014; Chen and Fung 2009) is

applied to incorporate physical body practices (Kyung

et al. 2008), and simulation (Huang et al. 2019; Sueb￾somran 2019). The delayed consequences of past investi￾gations showed the most huge parameters of compliant

mechanisms such as displacement (Dao and Huang 2015;

Le Chau et al. 2020), frequency (Dao et al. 2017b; Huang

et al. 2012), stress (Cha et al. 2018; Chau et al. 2018), and

strain(Bhattacharya et al. 2015a). Regardless, their struc￾tures were built up for testing on account of the duplicates

of both unique mechanics rehearses and enormous non￾direct redirections conversely with the unbendable casing

accomplices. Observational showing can look at compliant

mechanisms (Dao and Huang 2017b) with high exactness;

nonetheless, it is extreme and repetitive. It is not simple to

separate a giant contorting of flexure turns using the basic

model. The finite element method (FEM) (Zhang et al.

2015) explores complex numerical conditions of compliant

mechanisms by discretizing reliable frameworks into

mechanisms and hubs (Liu et al. 2019b; Masory et al.

1989; Petkovic et al. 2014). Even though the previous

methods are up still now generous, they are so far testing to

separate the complex compliant mechanisms. The evalu￾ating precision (Bhattacharya et al. 2015a; Chattaraj et al.

2016) of exhibiting relies upon an unconventionality level

of designing mechanics coupling of suitable parts. Espe￾cially, to non-straight redirections with irregular structures,

the recently surface response is restricted. For the optimal

shape design; Babaei and Sheidaii (2017) of various

models of compliant mechanisms, the problem is getting

more eccentric. Likewise, a data-driven strategy, preparing

data, and AI were operated concerning the display of

complex systems. Data-driven techniques coupled with

enlisting feedbacks (Deaconescu and Deaconescu 2017;

Dearden et al. 2017; Dhelika et al. 2016) were a feasible

mechanism that has not loosened up to the field of com￾pliant mechanisms yet. From reviewing the composing

review, the motivation of this article is to develop another

hybrid numerical bi-algorithm to comprehend the optimal

structure for predictable parts. The proposed method is

contacted tackle for most accepted frameworks from a

basic shape to the compliant gripper.

A notable problem with getting the numerical model for

a beam constraint model (BCM) (Chang et al. 2020) is that

it is not considered through the referred informative liter￾ature. The BCM has nonlinear re-directions (Chen et al.

2017; Liu et al. 2018; Petkovic et al. 2012). In some case

studies, the predicted results probably will not be correct.

Appropriately, a later optimal result for the part may be

free. In such conditions, a few data-driven systems may be a

practical method to handle the MOO problem for the BCM

model. This system benefits by clearly designing the well￾springs of data and outputs through a numerical method. A

couple of standard systems consolidate the appealing

quality limit approaches of design for assembly (DFA)

(Tran et al. 2020b; Le Chau et al. 2020; Liu et al. 2019a)

grey-relational-analysis (GRA) (Chatterjee and Das 2020;

Li et al. 2020) Taguchi method-based Fuzzy logic (TMFL)

(Sahin and Erol 2018). Both the DFA and GRA require a

consigning of weight factors for each simulation time (Dao

et al. 2017a). These factors are unequivocally liable to

comprehension or customers. At that point, the TMFL is a

system without the need for any weighting factors. It

prompts optimal characteristics at discrete focuses (Liu

et al. 2017). Notwithstanding, these past investigations had

no much more mentioned findings of the concepts of TMFL

replying upon the basic knowledge of the Taguchi method

(Le Chau et al. 2020; Pattnaik et al. 2014; Salmasnia et al.

2012; Tran et al. 2020a, b) that glances through discrete

upgraded courses of action. The discrete spotlights are on

the local optimum design. To be more specific and detailed

for the case study, there is a need to collect intermediary

models, response surface method (Hauser et al. 2018; Issa

et al. 2013; Petkovic et al. 2013, 2014; Salmasnia et al.

2012; Wang et al. 2016), Kriging procedure, logic- neural

Microsystem Technologies

123