Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Generative Model-based Synthesis of Four-bar Linkage Mechanisms with Target Conditions

Sumin Lee, Jihoon Kim, Namwoo Kang

TL;DR

This paper tackles the challenge of synthesizing crank-rocker four-bar linkage mechanisms that satisfy both kinematic and quasi-static requirements by employing a conditioned generative approach. A large, labeled dataset of 100,000 valid linkages is generated via Latin Hypercube Sampling with l1 fixed at 1.0, enabling the conditional GAN to learn mappings from target workspace and torque-transmission criteria to diverse, feasible linkages. A modified cGAN is proposed, incorporating a predictor for realized conditions and a diversity-promoting loss, along with a k-NN-based sampling strategy to reflect real condition distributions. Compared to a traditional cVAE and NSGA-II, the approach yields significantly more diverse yet compliant designs with competitive accuracy, highlighting its potential as a rapid design tool for mechanism synthesis in engineering practice.

Abstract

Mechanisms are essential components designed to perform specific tasks in various mechanical systems. However, designing a mechanism that satisfies certain kinematic or quasi-static requirements is a challenging task. The kinematic requirements may include the workspace of a mechanism, while the quasi-static requirements of a mechanism may include its torque transmission, which refers to the ability of the mechanism to transfer power and torque effectively. In this paper, we propose a deep learning-based generative model for generating multiple crank-rocker four-bar linkage mechanisms that satisfy both the kinematic and quasi-static requirements aforementioned. The proposed model is based on a conditional generative adversarial network (cGAN) with modifications for mechanism synthesis, which is trained to learn the relationship between the requirements of a mechanism with respect to linkage lengths. The results demonstrate that the proposed model successfully generates multiple distinct mechanisms that satisfy specific kinematic and quasi-static requirements. To evaluate the novelty of our approach, we provide a comparison of the samples synthesized by the proposed cGAN, traditional cVAE and NSGA-II. Our approach has several advantages over traditional design methods. It enables designers to efficiently generate multiple diverse and feasible design candidates while exploring a large design space. Also, the proposed model considers both the kinematic and quasi-static requirements, which can lead to more efficient and effective mechanisms for real-world use, making it a promising tool for linkage mechanism design.

Deep Generative Model-based Synthesis of Four-bar Linkage Mechanisms with Target Conditions

TL;DR

This paper tackles the challenge of synthesizing crank-rocker four-bar linkage mechanisms that satisfy both kinematic and quasi-static requirements by employing a conditioned generative approach. A large, labeled dataset of 100,000 valid linkages is generated via Latin Hypercube Sampling with l1 fixed at 1.0, enabling the conditional GAN to learn mappings from target workspace and torque-transmission criteria to diverse, feasible linkages. A modified cGAN is proposed, incorporating a predictor for realized conditions and a diversity-promoting loss, along with a k-NN-based sampling strategy to reflect real condition distributions. Compared to a traditional cVAE and NSGA-II, the approach yields significantly more diverse yet compliant designs with competitive accuracy, highlighting its potential as a rapid design tool for mechanism synthesis in engineering practice.

Abstract

Mechanisms are essential components designed to perform specific tasks in various mechanical systems. However, designing a mechanism that satisfies certain kinematic or quasi-static requirements is a challenging task. The kinematic requirements may include the workspace of a mechanism, while the quasi-static requirements of a mechanism may include its torque transmission, which refers to the ability of the mechanism to transfer power and torque effectively. In this paper, we propose a deep learning-based generative model for generating multiple crank-rocker four-bar linkage mechanisms that satisfy both the kinematic and quasi-static requirements aforementioned. The proposed model is based on a conditional generative adversarial network (cGAN) with modifications for mechanism synthesis, which is trained to learn the relationship between the requirements of a mechanism with respect to linkage lengths. The results demonstrate that the proposed model successfully generates multiple distinct mechanisms that satisfy specific kinematic and quasi-static requirements. To evaluate the novelty of our approach, we provide a comparison of the samples synthesized by the proposed cGAN, traditional cVAE and NSGA-II. Our approach has several advantages over traditional design methods. It enables designers to efficiently generate multiple diverse and feasible design candidates while exploring a large design space. Also, the proposed model considers both the kinematic and quasi-static requirements, which can lead to more efficient and effective mechanisms for real-world use, making it a promising tool for linkage mechanism design.
Paper Structure (25 sections, 9 equations, 12 figures, 6 tables)

This paper contains 25 sections, 9 equations, 12 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Numerical Methods
  4. Deep Learning-based Methods
  5. Variational Autoencoders (VAE)
  6. Generative Adversarial Networks (GAN)
  7. Generation of Linkage Mechanisms
  8. Crank-rocker Mechanism
  9. Conditions
  10. Kinematic Condition
  11. Quasi-static Condition
  12. Data Generation for Model Training
  13. Synthesis with Deep Generative Model
  14. Conditional Generative Adversarial Network
  15. Training
  16. ...and 10 more sections

Figures (12)

  • Figure 1: Deep generative model-based mechanism synthesis considering both kinematic and quasi-static conditions
  • Figure 2: A labeled example of a crank-rocker four-bar linkage mechanism
  • Figure 3: (a) The parameters needed for the calculation of $\eta_{min}$ and (b) the plot of $\eta$ along the path of a mechanism
  • Figure 4: (a), (b) Histograms and (c) a scatter plot of $d_{max}$ and $\eta_{min}$ of $X_{train}$
  • Figure 5: Visualization of few mechanisms with similar conditions generated for the model training
  • ...and 7 more figures