An official website of the United States government
Abstract Large deflection modeling is a crucial field of study in the analysis and design of compliant mechanisms (CM). This paper proposes a machine learning (ML) approach for predicting the deflection of discrete variable stiffness units (DSUs) that cover a range from small to large deflections. The primary structure of a DSU consists of a parallel guide beam with a hollow cavity that can change stiffness discretely by inserting or extracting a solid block. The principle is based on changing the cross-sectional area properties of the hollow section. Prior to model training, a large volume of data was collected using finite element analysis (FEA) under different loads and various dimensional parameters. Additionally, we present three widely used machine learning-based models for predicting beam deflection, taking into account prediction accuracy and speed. Several experiments are conducted to evaluate the performance of the ML models that were compared with the FEA and analytical model results. The optimal ML model, multilayer perceptron (MLP), can achieve a 7.9% maximum error compared to FEA. Furthermore, the model was employed in a practical application for inverse design, with various cases presented depending on the number of solved variables. This method provides a innovative perspective for studying the modeling of compliant mechanisms and may be extended to other mechanical mechanisms.
more »
« less
Deep Learning for Strain Field Customization in Bioreactor with Dielectric Elastomer Actuator Array
In the field of biomechanics, customizing complex strain fields according to specific requirements poses an important challenge for bioreactor technology, primarily due to the intricate coupling and nonlinear actuation of actuator arrays, which complicates the precise control of strain fields. This paper introduces a bioreactor designed with a 9 × 9 array of independently controllable dielectric elastomer actuators (DEAs), addressing this challenge. We employ image regression-based machine learning for both replicating target strain fields through inverse control and rapidly predicting feasible strain fields generated by the bioreactor in response to control inputs via forward control. To generate training data, a finite element analysis (FEA) simulation model was developed. In the FEA, the device was prestretched, followed by the random assignment of voltages to each pixel, yielding 10,000 distinct output strain field images for the training set. For inverse control, a multilayer perceptron (MLP) is utilized to predict control inputs from images, whereas, for forward control, MLP maps control inputs to low-resolution images, which are then upscaled to high-resolution outputs through a super-resolution generative adversarial network (SRGAN). Demonstrations include inputting biomechanically significant strain fields, where the method successfully replicated the intended fields. Additionally, by using various tumor–stroma interfaces as inputs, the bioreactor demonstrated its ability to customize strain fields accordingly, showcasing its potential as an advanced testbed for tumor biomechanics research.
more »
« less
- Award ID(s):
- 2301509
- PAR ID:
- 10618570
- Publisher / Repository:
- Science Partner Journals
- Date Published:
- Journal Name:
- Cyborg and Bionic Systems
- Volume:
- 5
- ISSN:
- 2692-7632
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract BackgroundMagnetic resonance imaging (MRI) scans are known to suffer from a variety of acquisition artifacts as well as equipment‐based variations that impact image appearance and segmentation performance. It is still unclear whether a direct relationship exists between magnetic resonance (MR) image quality metrics (IQMs) (e.g., signal‐to‐noise, contrast‐to‐noise) and segmentation accuracy. PurposeDeep learning (DL) approaches have shown significant promise for automated segmentation of brain tumors on MRI but depend on the quality of input training images. We sought to evaluate the relationship between IQMs of input training images and DL‐based brain tumor segmentation accuracy toward developing more generalizable models for multi‐institutional data. MethodsWe trained a 3D DenseNet model on the BraTS 2020 cohorts for segmentation of tumor subregions enhancing tumor (ET), peritumoral edematous, and necrotic and non‐ET on MRI; with performance quantified via a 5‐fold cross‐validated Dice coefficient. MRI scans were evaluated through the open‐source quality control tool MRQy, to yield 13 IQMs per scan. The Pearson correlation coefficient was computed between whole tumor (WT) dice values and IQM measures in the training cohorts to identify quality measures most correlated with segmentation performance. Each selected IQM was used to group MRI scans as “better” quality (BQ) or “worse” quality (WQ), via relative thresholding. Segmentation performance was re‐evaluated for the DenseNet model when (i) training on BQ MRI images with validation on WQ images, as well as (ii) training on WQ images, and validation on BQ images. Trends were further validated on independent test sets derived from the BraTS 2021 training cohorts. ResultsFor this study, multimodal MRI scans from the BraTS 2020 training cohorts were used to train the segmentation model and validated on independent test sets derived from the BraTS 2021 cohort. Among the selected IQMs, models trained on BQ images based on inhomogeneity measurements (coefficient of variance, coefficient of joint variation, coefficient of variation of the foreground patch) and the models trained on WQ images based on noise measurement peak signal‐to‐noise ratio (SNR) yielded significantly improved tumor segmentation accuracy compared to their inverse models. ConclusionsOur results suggest that a significant correlation may exist between specific MR IQMs and DenseNet‐based brain tumor segmentation performance. The selection of MRI scans for model training based on IQMs may yield more accurate and generalizable models in unseen validation.more » « less
-
null (Ed.)Abstract Computational approaches, especially finite element analysis (FEA), have been rapidly growing in both academia and industry during the last few decades. FEA serves as a powerful and efficient approach for simulating real-life experiments, including industrial product development, machine design, and biomedical research, particularly in biomechanics and biomaterials. Accordingly, FEA has been a “go-to” high biofidelic software tool to simulate and quantify the biomechanics of the foot–ankle complex, as well as to predict the risk of foot and ankle injuries, which are one of the most common musculoskeletal injuries among physically active individuals. This paper provides a review of the in silico FEA of the foot–ankle complex. First, a brief history of computational modeling methods and finite element (FE) simulations for foot–ankle models is introduced. Second, a general approach to build an FE foot and ankle model is presented, including a detailed procedure to accurately construct, calibrate, verify, and validate an FE model in its appropriate simulation environment. Third, current applications, as well as future improvements of the foot and ankle FE models, especially in the biomedical field, are discussed. Finally, a conclusion is made on the efficiency and development of FEA as a computational approach in investigating the biomechanics of the foot–ankle complex. Overall, this review integrates insightful information for biomedical engineers, medical professionals, and researchers to conduct more accurate research on the foot–ankle FE models in the future.more » « less
-
Digital image correlation (DIC) is an increasingly popular and effective non-contact method for measuring full-field displacements and strains of deformable bodies under load. Current DIC methods applied to bodies undergoing large displacements and rotations require a large measurement area for both the reference (i.e., undeformed) image and the deformed images. This can limit the resulting resolution of the displacement and strain fields. To address this issue, we propose a two-stage dynamic DIC method capable of measuring displacements and strains under material convection with high resolution. During the first stage, the reference image is assembled from smaller, high-resolution images of the undeformed object obtained using a spatially-fixed or moving frame. Following capture, each sub-image is rigidly translated and rotated into its appropriate place, thereby producing a full, high-resolution image of the reference body. In stage two, images of the loaded and deformed body, again obtained using a small camera frame with high resolution, are aligned with matching regions of the undeformed composite image using BRISK feature detection before performing DIC.We demonstrate the method on a contact problem whereby an elastomeric roller travels along a rigid surface. In doing so, we obtain fine resolution measurements of the state of strain of the region of the roller sidewall in contact with the substrate, even as new material convects through the region of interest. We present these measurements as a series of images and videos capturing strain evolution as the roller transitions from static loads to a fully dynamic steady-state, documenting the effectiveness of the method.more » « less
-
Implicit neural representations (INRs) have recently advanced numerous vision-related areas. INR performance depends strongly on the choice of activation function employed in its MLP network. A wide range of nonlinearities have been explored, but, unfortunately, current INRs designed to have high accuracy also suffer from poor robustness (to signal noise, parameter variation, etc.). Inspired by harmonic analysis, we develop a new, highly accurate and robust INR that does not exhibit this tradeoff. Our Wavelet Implicit neural REpresentation (WIRE) uses as its activation function the complex Gabor wavelet that is well-known to be optimally concentrated in space–frequency and to have excellent biases for representing images. A wide range of experiments (image denoising, image inpainting, super-resolution, computed tomography reconstruction, image overfitting, and novel view synthesis with neural radiance fields) demonstrate that WIRE defines the new state of the art in INR accuracy, training time, and robustness.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.34133/cbsystems.0155
