ABB Robotics has integrated vision-based artificial intelligence into its Fashion Inductor sorting system. ABB Robotics has integrated vision-based artificial intelligence into its Fashion Inductor sorting system.

ABB targets fashion logistics with Physical AI robotics system

ABB Robotics has integrated vision-based artificial intelligence into its Fashion Inductor sorting system, claiming the technology can cut engineering and commissioning effort by up to 75%.

The upgraded system combines the Fashion Inductor module with AI Robot Trainer, which enables system integrators to deploy, retrain and optimise robotic applications without specialist AI expertise. ABB is targeting fashion and apparel logistics, where variations in products, packaging and operating conditions have traditionally made automation more difficult.

Fashion Inductor can sort up to 1,300 items per hour per cell, with picking accuracy targets above 99.5%, according to ABB. The system uses AI-powered automatic path planning, motion control and application-specific vision technology.

AI Robot Trainer adds pre-trained vision models tailored to fashion logistics, alongside tools for retraining models as operating requirements change. ABB said the approach reduces the engineering work required to deploy applications and makes it easier to adapt systems after installation.

“Some of the greatest opportunities for physical AI are in applications that have traditionally been difficult to automate because of high variability,” said Craig McDonnell, Managing Director of the Industries business line at ABB Robotics.

The system also incorporates a digital twin of the Fashion Inductor module, linking ABB’s Application Controller Platform, automatic path planning software, AI Robot Trainer and RobotStudio simulation environment. The digital twin is designed to accelerate commissioning and allow greater flexibility when configuring applications, including integrating third-party components such as grippers.

Rather than treating AI models as static components, ABB’s approach is designed to improve them throughout a system’s operational life. AI Robot Trainer monitors robot performance, identifies anomalies, generates new training data and deploys updated models to the physical robot.

RobotStudio provides the virtual design, simulation and validation environment, while ABB’s OmniCore controller executes robot programmes in production.

Fashion Inductor is the first application in ABB Robotics’ wider rollout of vision-based Physical AI, which the company sees as part of its development of more autonomous and versatile robotics. Further applications using AI Robot Trainer are planned through the first quarter of 2027.