Predictive Knitwear

Bridging design and manufacturing with AI

The Knitwear Communication Gap

While advanced knitting machines can create highly functional, seamless garments, a significant gap exists between a designer’s creative vision and the technical programming required by the machine. The relationship between a fabric’s visual appearance, its physical form, and its technical performance is subtle and complex. This disconnect leads to a slow, costly development cycle that relies on a fragmented workflow: a designer’s concept is translated by an engineer, a physical sample is produced, and only then can its properties be evaluated. This process hinders rapid and informed creative exploration.

Overview of the steps normally taken in the design process of garments (photos by CRISP project, Eindhoven University of Technology).

An example of a highly personal and functional garment. This seamless sports bra combines distinct knit structures for tailored support, ventilation, and comfort. Project by Studio Eva x Carola.

Machine Learning for Functional Performance

This approach focused on predicting a fabric’s functional properties. We systematically produced 36 unique knit samples, varying key parameters like stitch type and structure. These samples then underwent rigorous lab testing to quantify performance metrics like compression, weight, comfort, and evaporation. We used this dataset to train machine learning algorithms, resulting in a prototype with a simple 2D interface. This tool allows designers to select knitting parameters and receive instant numerical feedback on the predicted performance, helping to optimize a garment’s functionality without creating physical samples.

The 36 systematically varied knit samples that form the physical dataset used to train our machine learning models. Each sample represents a unique combination of stitch parameters.

The prototype user interface. It allows a designer to manipulate knit parameters on the left and receive instant, predictive feedback on the fabric's functional properties on the right.

Parametric Simulation for 3D Form-Finding

This approach focused on predicting a fabric’s physical form and behavior under tension. We developed a novel method using parametric CAD software (Grasshopper) to simulate the tensile state of a tubular fabric. First, we built physical models to analyze the fabric’s elasticity. We then created a digital model that could accurately replicate these measured properties. By comparing the physical and digital results, we validated the simulation and discovered a clear linear relationship between the fabric’s deformation and a key simulation parameter called “rest length factor”.

This resulted in a 3D parametric simulation that gives designers an intuitive way to work with the complex forms that emerge from stretched textiles. This form-finding tool was successfully applied to design a 3D knitted lamp, which artfully uses the changes in density and light transmittance of the fabric under tension.

Validation of the parametric simulation. On the left, complex geometries are explored using a digital wireframe model. On the right, the final 3D knitted lamp is shown, confirming that the physical object accurately materializes the form predicted by the digital tool.

Physical analysis of fabric deformation. The silhouettes of different tubular knit samples are overlapped to visually compare how their underlying structure—utilizing Miss, Tuck, and Float stitches—affects the final 3D shape when tensioned between three rings.

My Role: Principal Investigator

As the project lead, I initiated and directed this interdisciplinary research, working at the intersection of industrial design, computer science, architecture, and textile engineering. My role involved:

Defining the research vision and core research questions.
Designing the experimental methodology for data collection, physical testing, and simulation.
Overseeing the development of machine learning and parametric models in collaboration with researchers and students.
Leading the user interface design process and managing industry collaborations with partners like Santoni Shanghai.
Co-authoring of papers (ten Bhömer et al., 2019; ten Bhömer et al., 2019)

Collaborators & Partners

Academic Collaborators (Xi’an Jiaotong-Liverpool University)

Computer Science: Prof. Hai-Ning Liang, Dr. Difeng Yu, Yanyan Meng
Architecture: Prof. Thomas Wortmann , Mingyu Wang
Industrial Design: Yuanjin Liu, Yifan Zhang, Tong Zhang

Industry Partners

Studio Eva x Carola: Eva de Laat & Carola Leegwater
Santoni Shanghai

Funding

This work was supported by the National Natural Science Foundation of China (NSFC, Grant No. 51750110497) and by Xi’an Jiaotong-Liverpool University (Grant No. RDF-13-02-19 and RDF-17-01-54).

References

2019

Designing Predictive Tools for Personalized Functionalities in Knitted Performance Wear

Martijn ten Bhömer, Hai-Ning Liang, Difeng Yu, and 4 more authors

Jun 2019

Abs DOI Bib PDF Website

Developments of advanced textile manufacturing techniques—such as 3D body-forming knitwear machinery—allows the production of almost finalized garments, which require little to no further production steps to finalize the garment. Moreover, advanced knitting technology in combination with new materials enables the integration of localized functionalities within a garment on a “stitch by stitch level.” There is potential in enhancing the design tools for advanced knitting manufacturing through the use of technologies such as data gathering, machine learning, and simulation. This approach reflects the potential of Industry 4.0, as design, product development, and manufacturing are moving closer together. However, there is still limited knowledge at present about how these new technologies and tools can have an impact on the creative design process. The case study presented in this paper explores the potential of predictive software design tools for fashion designers who are developing personalized advanced functionalities in textile products. The main research question explored in this article is: “How can designers benefit from intelligent design software for the manufacturing of advanced personalized functionalities in textile products?”. Within this larger research question three sub-research questions are explored: (1) What kind of advanced functionalities can be considered for the personalization process of knitwear? (2) How to design interactions and interfaces that use intelligent predictive algorithms to stimulate creativity during the fashion design process? (3) How will predictive software impact the manufacturing process for other stakeholders and production steps? These questions are investigated through the analysis of a Research Through Design case study, in which several predictive algorithms were compared and implemented in a user interface that would aid knitwear designers during the development process of high-performance running tights.
@article{designing-predictive-tools, title = {Designing Predictive Tools for Personalized Functionalities in Knitted Performance Wear}, author = {{ten Bhömer}, Martijn and Liang, Hai-Ning and Yu, Difeng and Liu, Yuanjin and Zhang, Yifan and de Laat, Eva and Leegwater, Carola}, year = {2019}, month = jun, volume = {35}, pages = {42--75}, doi = {10.46467/TdD35.2019.42-75}, journaltitle = {Temes de disseny}, issue = {1}, }
Machine Learning Enhanced User Interfaces for Designing Advanced Knitwear

Martijn ten Bhömer, Hai-Ning Liang, and Difeng Yu

In HCI International 2019 - Posters. HCII 2019. Communications in Computer and Information Science, vol 1033., Jul 2019

Abs DOI Bib PDF

The relationship between visual appearance and structure and technical properties of a knitted fabric is subtle and complex. This is an area that has been traditionally problematic within the knitting sector, understanding between technologists and designers is hindered which limits the possibility of dialogues from which design innovation can emerge. Recently there has been interest from the Human-Computer Interaction (HCI) community to narrow the gap between product design and knitwear. The goal of this article is to show the potential of predictive software design tools for fashion designers who are developing personalized advanced functionalities in textile products. The main research question explored in this article is: “How can designers benefit from intelligent design software for the manufacturing of personalized advanced functionalities in textile products?”. In particular we explored how to design interactions and interfaces that use intelligent predictive algorithms through the analysis of a case study, in which several predictive algorithms were compared in the practice of textile designers.
@inproceedings{machine-learning-enhanced-ui, title = {Machine Learning Enhanced User Interfaces for Designing Advanced Knitwear}, author = {{ten Bhömer}, Martijn and Liang, Hai-Ning and Yu, Difeng}, year = {2019}, month = jul, booktitle = {HCI International 2019 - Posters. HCII 2019. Communications in Computer and Information Science, vol 1033.}, doi = {10.1007/978-3-030-23528-4_30}, editor = {Stephanidis, Constantine}, }