OptiStrick – Development of an AI-supported inline quality assurance system for optimizing highly flexible knitting technologies

Project description and goals


German textile market: small to medium series, very frequent item changes, high quality materials

An essential aspect of the qualitative growth and competitive positioning of the German textile industry, which is dominated by small and medium-sized enterprises, is to meet customer demands for cost-effective, high-quality, and individualized products. Flat knitting technology, with its enormous design possibilities and short set-up times, is particularly well suited to the production of individualized textile products with customer-specific properties. The process-integrated, sensor-supported, real-time quality assurance systems that have become established in industrial mass production in other sectors, such as automotive, processing machinery and semiconductors, are not easily transferable to textile technology due to the frequent product and material changes in small to medium batch sizes.

To meet the individual requirements of the textile industry, especially in flat knitting, an AI-based inline QA system is needed to optimize highly flexible knitting technologies.

The flat knitting technology challenges

Despite the known problems, there are still no technical systems for automatic determination and, at best, automatic feedback of machine setting values (mainly gauge depth, yarn tension, carriage, and take-down speed) in response to detected yarn, process or product changes or faults. Machine setting and monitoring to ensure high product quality and machine availability is currently based on the accumulated experience of a few highly qualified and experienced employees within a company. Based on the extensive knowledge and experience of these employees, the necessary adjustments to the settings on each machine are derived separately, via several optimization loops, from the evaluation of the look and feel of the knitted fabric as well as the acoustic emissions of the machines.

Acoustic analysis with AI to mimic the experience of professionals

The aim of this interdisciplinary and cross-industry project is to develop a novel, flexible, intelligent, and inline measuring quality assurance system. Using a flat knitting machine as an example, it will enable real-time analysis of the current material and operating conditions and the resulting textile properties. The findings from the previously investigated areas of application of acoustic monitoring will be transferred to textile production. To this end, existing approaches will be evaluated and further developed specifically for complex flat knitting, in order to be able to research and implement adapted AI solutions. The transfer to related types of textile processing such as weaving, embroidery, knitting and sewing will also be considered.

Responsibility of Fraunhofer IDMT

  • Development of a concept for an acoustic or vibro-acoustic sensor system
  • Development of machine learning based algorithms on the measured audio data to determine yarn properties, machine conditions and fabric quality


  • TU Dresden, Institute of Textile Machinery and High Performance Material Technology

The project is supported by a project monitoring committee made up of knitting manufacturers and machine builders.


Federal Ministry of Economics and Climate Protection BMWK


July 2023 to November 2025

Industrial Sound Analysis – Research and Practice