Robots simply do not have the dexterity of skilled human fingers to accurately manipulate materials through sewing machines – yet.
Fabrics are not like the steel, plastic or other rigid parts that have enabled the full automation of other industries, most notably automotive and aerospace – fabrics bunch and stretch and their properties can vary significantly in how they respond to being processed.
As a result, the various stages of garment manufacturing – from pattern making through to quality control and packaging – involve more people at nearly every stage compared to fabric production.
Sewing remains the most labour-intensive of all production stages, and depending on the complexity of the garments being produced, typically involves between 30-50 per cent of the total workforce of a mill that is vertically integrated from yarn spinning through to finished garments.
As such, the final assembly of garments from their various components secures regular work for millions of people, and especially women, often in regions where other job opportunities can be limited.
Pattern and sample making
Designing and creating garment patterns, as the starting point for the entire garment making process, has in the past required skilled designers who are not only highly creative, but also adept at handling often complex CAD tools.
Garment design, however, is now certainly undergoing a transformation, with traditional methods being significantly replaced by digital tools, AI and 3D modelling, which offer greater simplicity, efficiency and precision.
3D pattern-making software platforms such as Shima Seiki’s APEXFiz are accelerating digital fabric sourcing prior to 3D design and production and also eliminating previously time-consuming and costly fabric sampling.
APEXFiz enables fabric data, including texture maps, to be created without fabric samples, and colourways and patterns to be freely edited, in combination with the creation of all information for automatic production.
For flat knitting, the programme incorporates Shima Seiki’s Yarnbank, consisting of extensive digitised yarn data supplied by all partner yarn manufacturers, and the company’s Datamall library of over 1,500 default knitting structure patterns. All yarns and structures can be freely edited.
“As 3D creation increases across the industry, accelerating the procurement of digital fabrics, APEXFiz can create fabric designs based on production data and simulate virtual fabrics using digitally-sampled yarns,” explains Hiyato Nishi, of Shima Seiki USA. “In recent years new functions have been added to create even more realistic digital fabrics while being able to export texture map data to be used in other 3D software.”
Create Design
In one of the latest developments in this field, apparel software developer Style3D/Assyst has teamed up with KM.ON, the automation arm of knitting technology leader Karl Mayer, to introduce Create Design.
This combined programme enables apparel companies to exploit both 3D fashion and knitwear design in their collection creation processes, significantly reducing time-to-market for knitwear and saving costs.
As a result of the integration of the programmes of the two specialists, designers and product developers can design and visualise fully fashioned knitwear patterns in the KM.ON software, including complex knitting techniques such as cable, Aran, pointelle, jacquard or intarsia, work with digital yarns and directly transfer them to Style3D Studio with a jointly developed plugin.
In Style3D Studio, the knitwear designs can be simulated in 3D on an avatar, including styling, animations, and style sharing by QR code or link. The AI rendering module in Style3D Studio also enables digital photoshoots for e-commerce, showrooms and other promotions.
Colour matching and calibration
Automation in colour matching is another area where technology is making a big impact. In traditional fabric sampling, colour discrepancies between fabric batches can lead to costly mistakes and delays. Automated colour matching systems generally use spectrophotometers to ensure that colours are consistent across fabric samples and production runs.
The DigiEye system for advanced non-contact colour measurement and digital imaging marketed by the UK’s VeriVide is said to overcome certain limitations of spectrophotometers, such as being restricted to products with a solid colour appearance and only measuring ‘average’ colour, assessing the limited area of the product exposed in the instrument’s aperture.
These limitations of spectrophotometers are further compounded if the product has any form of multi-coloured characteristics or appearance, an uneven or inconsistent surface, a colour area too small to be captured by the instrument or any sort of surface effect, such as carpet pile and the sheen of satin and silk.
Best practice
“Unlike a spectrophotometer, DigiEye provides a ‘best practice’ combination of lighting and visual assessment principles, together with non-contact digital measurement for quantifiable and objective quality control,” explains VeriVide sales manager Adam Dakin. “The system allows the entire product to be assessed, ensuring that all colours can be analysed and measured in context with all other visual and surface effect elements of a product. It can also accommodate items which are integral components of the apparel industry, such as lace, lingerie, patterns, prints and home furnishings, that are non-solid in colour and are heavily textured.”
DigiEye collects and processes data that can be communicated and shared instantly, enabling manufacturers, producers and processors to speed up quality control, sustain product integrity and reduce waste.
Cutting
A more well-established process is the use of CAD systems which enable precise pattern drafting, grading, nesting and automated garment part cutting.
Some of the most widely used systems in this field are supplied by Gerber and Lectra, which since June 2021 have been operating as a single company.
Lectra acquired all outstanding shares of Gerber for €175 million, bringing the two rivals in automated textile cutting, plotting and associated assembly software together to create a leading global Industry 4.0 player for the fashion, automotive and furniture markets.
Automated cutting machines such as those from Gerber and Lectra are equipped with advanced sensors and often also robotic arms that can cut fabric samples precisely, based on digital patterns.
Lectra’s Vector multi-ply cutting solutions, for example, enable just-in time and mass production processes while fulfilling last-minute orders with ease. Regardless of the production volume, minimal cost per cut part is the result, with little need for manual intervention.
During 2023, Lectra introduced two multi-ply fabric-cutting systems specifically tailored to the needs of the Asian market and compliant with Industry 4.0 standards – VectorFashion iC70 and Gerber Atria 70.
VectorFashion iC70 achieves the optimal balance between quality and efficiency while delivering stable equipment runtime. In addition, its operational platform leverages multiple digital technologies to enhance automation, optimise workflows and heighten operational efficiency. The Gerber Atria 70 meanwhile draws from multiple connectivity technologies for seamless integration with other Gerber solutions, ensuring streamlined and optimised digital workflows.
Nesting
Given that fabric is the biggest production expense for fashion brands (around 60-70 per cent of costs) and that making a single garment can generate up to 20 per cent waste, optimised nesting – the process of marking and laying out cutting patterns on fabrics – also plays a critical role.
Again, this is something that is still done manually by skilled workers in many mills throughout the world, but advanced programmes such as Lectra’s Cloud-based Flex Offer automatically calculate the ideal distribution of cutting patterns, analysing the parts to be produced for each cutting job and using algorithms to determine how to lay them out to best produce the required quantities, while minimising the amount of space between the patterns.
This improves marker efficiency, speeds up creation and makes a significant contribution to reducing operational costs and inventory expenses, while also enabling full deliveries and shortening response times. Previous markers and nest plots can also be retrieved for new collections.
Additionally, with Flex Offer, nesting can be made remotely and during the cutting, while precious in-house know-how can be replicated.
On-demand production
On-demand apparel manufacturing – focused on making garments only when someone needs them within a short time and at a reasonable price – has also recently been exploding, enabling many lean companies to enter the retail apparel market.
The on-demand model makes sense in numerous ways, and is being accelerated by numerous new technologies, not least digital printing.
At the ITMA 2023 textile machinery show in Milan last June, for example, Israel-based Kornit hosted a showcase under the slogan ‘Digital Production Goes Mainstream’. The company highlighted its end-to-end production facilities for on-demand manufacturing with the latest manufacturing technologies integrated into a T-shirt production line in Milan.
This model begins with user-friendly design software enabling the 3D simulation of a virtual garment, allowing customers to evaluate and finalise their design decisions. The customised designs are digitally printed using a direct-to-garment printer, with curing also administered seamlessly. The finished products are automatically packaged and shipped out for delivery and during the entire process every resource – such as print designs, fabric substrate decisions and information on size and fit – is managed by barcodes.
Monster Digital
Kornit celebrated countless digital printing machine sales at ITMA 2023, and online businesses and digital print companies are continuing to opt for its technologies and services in 2024, including most recently Custom Ink, DO Apparel, OGAT, Print Palace Textildruck and T-Formation.
“The apparel industry is evolving, and companies must evolve with it,” says Kornit CEO Ronen Samuel. “Today’s businesses and consumers value the ability to have unlimited design capabilities with fast delivery times and products that are made sustainably. While screen printing has been the standard for years, Kornit’s advanced digital solutions are revolutionising the market.”
Monster Digital, headquartered in Miami, Florida, for example, plans to install ten additional Kornit Apollo printing systems during 2025 as it accelerates its ongoing transition from screen printing to digital textile production.
Traditionally a private label screen printer, Monster has significantly expanded its digital production during the past few years, building a fleet of seven Apollos and 47 Kornit Atlas Max systems.
The additional Apollos to be installed during 2025 will be executed in a mix of capital purchases and Kornit’s new All-Inclusive Click (AIC) model. This major expansion will enable the company to capture massive new opportunities and better meet existing customer demands by shifting significant bulk production impressions from screen to digital production.
“Kornit’s innovation with the latest Apollo system has been a true game-changer for our business,” says Scott Valancy, COO at Monster Digital. “Now we can move more screen volume to digital with no compromises. Apollo is changing the way we manage order flow – with the ability to produce digitally at levels of speed and quality the industry has never seen. We are excited with the growth opportunities Apollo opens for us and look forward to providing the ultimate customer experience.”
Mad Engine Global
As an early Kornit Apollo beta customer, Mad Engine Global is also expanding its investment in Kornit Apollo production lines and its Atlas MAX armoury of machines for growing its on-demand digital production capacity – with additional investments expected for 2025.
Offering the exclusive and non-exclusive apparel of major brands such as Disney and Marvel, the San Diego, California-based company is shifting more of its apparel production from legacy screen-printing processes and capitalising on the performance, unit economics and quality of the Kornit technology.
“Our business is built on delivering high quality merchandise to customers year-round, which includes producing volumes of all sizes and types,” observes Jaymes Clements, executive vice president at Mad Engine Global. “As the market demand shifts strongly to high quality mid-sized runs, our continued adoption of the Kornit Max technology across platforms is a perfect fit for our needs. Kornit’s recently introduced AIC model also improves our predictability and allows us to obtain faster and better-informed decisions when it comes to the type of production capacity we add.”
CNC machines
Kornit’s vision for such digital apparel supply chains, however, makes no mention of the labour and time-intensive cut and sew processes and attempts to automate the sewing process that are ongoing.
Seamless knitting machines such as those that have been pioneered by Japan’s Shima Seiki and Italy’s Stoll, of course, seek to eliminate seams altogether, and hence the sewing bottleneck, but the time it takes to manufacture a single garment via seamless knitting is much longer than simply cutting individual fabric components and manually sewing them together.
Brands including, Brother, Dürkopp Adler, Juki and Pfaff all supply CNC (computer numerical control) sewing machines equipped with microprocessors and programmable control systems that allow them to execute complex sewing patterns with high precision.
These machines can store multiple stitching patterns, adjust thread tension automatically, and ensure consistent stitch quality. They also allow for automatic fabric feeding, reducing operator involvement and increasing output.
Juki Central Europe’s AW-3S automatic bobbin thread winding and feeding device, for example, not only automatically calculates the amount of bobbin thread required when loading a sewing pattern but also automatically changes the sewing thread. This not only increases productivity, but also reduces operator fatigue.
Pfaff’s intelligent Vision System, which can be equipped with up to four cameras, has been developed to compensate for tolerances that arise in seam contours as a result of material shrinkage and other imperfections during sewing operations. The Vision System compares the image of a master part with the image of a part clamped in the sewing template. The complete seam design is then automatically corrected in the sewing programme based on the individual measurement points.
Infinite rotation
In other notable automation developments, Dürkopp Adler, headquartered in Bielefeld, Germany, has developed the 911Revolve CNC machine with a rotating sewing mechanism which enables perfect seams to be achieved in all directions, opening up new design possibilities.
The unique rotation mechanism in the sewing head makes it possible to create visually perfect decorative seams that are difficult to achieve with conventional sewing units.
Dürkopp Adler’s Delta e-con platform consists of a large variety of flat bed, cylinder arm and post bed machines with different types of transport for a wide range of sewing applications.
For overall control of sewing machine teams, the company’s Qondac Guided Working software module for programmable sewing machines provides a seamless link between planning, operating personnel and machine by supporting processes from production planning to the last work step at the final workstation.
For further automation, fully digitised handling solutions for finished garments are designed by Eton Systems of Sweden to eliminate manual transportation and minimising handling.
Eton’s individually addressable product carriers are fully managed and controlled by the latest ETONingenious software, a web based real-time data collection and information system that continuously accumulates, processes and makes all production information instantly available to supervisors, quality control personnel and management.
Despite all these new benefits, however, sewing machines still need a lot of individual machine operators.
Sewbots
Software Automation, headquartered in Austell, Georgia, generated huge media attention with its introduction of Sewbots – fully robotic sewing machines – back in 2017. The launch followed five years of development work with assistance from the Georgia Tech Advanced Technology Development Center and a contract with DARPA – the US Defense Advanced Research Projects Agency.
The company’s Sewbots are powered by smart vision technology that consists of cameras and sensors which steer and adjust them deftly – as skilled human hands do – in real time, as it travels through the sewing head. This patented technology is at the core of every Sewbot, giving it the ability to achieve full automation of sewn goods.
In a further announcement intensifying the buzz around the technology back in 2017, Chinese manufacturing giant Tian Yuan Garments, the largest producer of sportswear for Adidas, announced that it was planning to purchase 21 Sewbot lines for its plant in Little Rock, Arkansas, capable of producing 800,000 cotton T-shirts per day – one approximately every 50 seconds.
Software Automation’s deal with Tian Yuan, however, fell through and in a frank interview with Harris Quinn in Wired magazine four years later in 2021, the company’s CEO Palaniswamy Rajan outlined some of the hurdles faced in commercialising the technology.
Challenge
“The biggest challenge we have faced is the requirement of being able to operate 24/7 at high speeds and greater than 98 per cent quality,” he said, adding that no two cotton batches are alike and variations in fabrics and dyes further complicate matters, leading to the Sewbots having to be recalibrated, interrupting operations.
Having raised $30 million in venture investments and grants by 2021, Rajan said that it would still take tens of millions more in financing to get to a target of an annual one billion T-shirts – between 1,600 and 2,000 Sewbots at multiple facilities, each with its own skilled workers to maintain them.
The real problem is that with 200 billion T-shirts currently produced every year, largely in developing countries, labour is simply too cheap.
Software Automation, however, still operates the first and only machine vision robotic workline in the world which can automatically sew T-shirts at its plant in Atlanta, Georgia.
“Our fully automated Sewbots enable on-demand manufacturing by moving supply chains local and closer to the customer, while creating higher quality products in the most sustainable way at a competitive cost,” the company says on its website.
Rigid manipulation
Another promising and patented concept for robotic sewing that is yet to be industrialised comes from San Francisco-based Sewbo.
Sewbo avoids the difficulties robots face when trying to manipulate limp, flexible fabrics by simply temporarily stiffening them.
This allows off-the-shelf industrial robots to easily build garments from rigid cloth, just as if they were working with sheet metal. The fabric panels can be easily moulded and welded before being permanently sewn together.
The water-soluble stiffener (polyvinyl alcohol) is removed at the end of the manufacturing process with a simple rinse in hot water, leaving a soft, fully assembled piece of clothing. The stiffener can then be recovered for reuse. Polyvinyl alcohol is a non-toxic polymer that is already used elsewhere in the textile production process, for example, as a sizing agent that temporarily strengthens yarn during weaving.
Conclusion
Despite promising developments, the sewing process remains the major bottleneck in almost every garment supply chain – be that mass production, online and on-demand, fast fashion or ultra-fast fashion.
Electric sewing machines have served the fashion industry for more than a century and are likely to continue doing so for the foreseeable future, as these machines still rely on human dexterity and experience.
