ABB Robotics on October 15, 2026, launched what it describes as its first end-to-end vision-based industrial-grade physical AI solution, integrating AI Robot Trainer into its Fashion Inductor functional module for fashion logistics. The company said the addition reduces engineering and commissioning effort by up to 75% for system integrators deploying AI-powered fashion logistics applications.
The upgrade, announced from Zurich, Switzerland, is intended to help integrators deploy, scale and continuously optimize fashion logistics applications with reduced effort and risk, according to the company. Craig McDonnell, Managing Director, Business Line Industries at ABB Robotics, said in the announcement that applications with high variability have traditionally been difficult to automate and present some of the greatest opportunities for physical AI. “Fashion Inductor with AI Robot Trainer is our first end-to-end vision-based physical AI solution, demonstrating how intelligent automation can be deployed, operated and continuously optimized at industrial scale,” McDonnell said.
Pre-Trained Vision Models and Closed-Loop Retraining
AI Robot Trainer provides pre-trained, fashion-specific AI vision models, application-specific tools and rapid retraining capabilities. ABB said the platform helps system integrators deploy, adapt and continuously optimize applications as requirements evolve, without needing specialist AI expertise.
The retraining workflow runs as a continuous closed loop that monitors robot performance, identifies anomalies, generates new model training data and deploys improvements back to the physical robot. According to ABB, the AI models are continually improved throughout the system lifecycle, measured against real performance targets and changing operating conditions, through structured, rapid cell-specific retraining.
The Fashion Inductor product page describes AI Robot Trainer as ABB Robotics’ vision-based Physical AI platform, purpose-built for the variability of fashion applications. ABB lists fast adaptation to changing products and packaging and continuous improvement through lifecycle retraining among its capabilities, and states that, unlike traditional automation systems, the module continuously improves performance through that closed-loop workflow.
Digital Twin and Control Architecture
The upgraded module introduces a new digital twin, an exact replica of the module software, connecting the Application Controller Platform, Automatic Path Planning, AI Robot Trainer and RobotStudio. ABB said the twin accelerates commissioning and gives greater flexibility in set-up, including the ability to integrate third-party components such as grippers.
RobotStudio provides the foundation for virtual design, simulation and validation across the system lifecycle, while the OmniCore robot controller serves as the execution layer in production. The product page states that the integrated digital twin connects all software elements, enabling faster commissioning, simplified validation and the flexibility to scale and adapt applications as requirements evolve.
Throughput, Accuracy and Related Hardware
Fashion Inductor is designed for the variability of fashion and apparel logistics, where products, packaging and operating conditions change continuously. ABB states throughput of up to 1,300 items per hour per cell and picking accuracy targets above 99.5%, enabled by AI-powered Automatic Path Planning, motion accuracy and application-specific vision technology. The product page’s technical data lists accuracy above 99.5% and throughput of up to 1,300 picks per hour.
Related offerings listed with the module include the Robotic Item Picker, the Robotic Parcel Inductor and the IRB 1300 and IRB 2600 robots. ABB also offers a downloadable five-step guide on apparel sorting and induction, which identifies labor shortages, seasonal peaks and changing product mixes as factors that can limit manual sorter induction.
Item Picking Family Background and Expansion Plans
ABB introduced the Fashion Inductor alongside the Parcel Inductor as AI-powered functional modules in its Item Picking family for logistics and e-commerce supply chains. ABB said both modules use its AI-based vision technology to handle unknown and randomly arranged items in unstructured environments, with picking accuracy of over 99.5%, and that the AI can be trained to detect and reject non-inducible items such as spherical or cylindrical products. The earlier release rated the Fashion Inductor at up to 1,300 picks per hour for polybagged apparel and accessories and the Parcel Inductor at up to 1,500 picks per hour for boxes, bags, envelopes and packages.
According to that release, both modules came pre-integrated, with commissioning possible in as little as one week, using ABB’s Application Controller Platform to integrate robots, grippers and cameras through a single computer and user interface.
ABB describes Fashion Inductor with AI Robot Trainer as the first step in a broader rollout of vision-based physical AI, a core part of its work on more autonomous and versatile robotics. On its Autonomous Versatile Robotics program page, ABB describes AVR robots as using Generative AI to combine six skill areas: robot-human interaction, sensing and perception, intelligence and autonomous reasoning, motion control and safety, navigation, and grasping and dexterity. The company said the same AI Robot Trainer-powered approach will expand across additional ABB Robotics vision-based applications through Q1 2027.
Credit: Source link
