AI Assistance System

With its Smart Wearable ClickID, VOSS Automotive GmbH is developing a solution for the flawless installation of connectors and wiring harnesses at manual assembly workstations. The sensorized glove captures all relevant process data and processes it in real time to immediately detect faulty connections.

aiXbrain shall implement the AI-based processing of sensor data on the wearable's microprocessor to enable highly precise quality analysis in real time.

• Integration with embedded microprocessor system + on-premise
• AI framework for training and managing AI models
• Classification of assembly quality with AI

First, efficient integration and management of training data was ensured to create a solid foundation for developing the AI models.

Building on this, a comprehensive MLOps pipeline as an on-premise service was implemented, enabling automated training of various AI models and visualizing their performance. The models use supervised deep learning to recognize complex patterns across all sensor channels.

Trained AI models are seamlessly transferred as runtime models to the glove's microprocessor, where they provide real-time assessment of assembly quality through haptic and visual feedback.

Artificial intelligence powers the ClickID Smart Wearable to detect quality defects in the assembly process in real time, avoiding costly rework. Assembly personnel benefit from an improved work experience.

aiXbrain contributed with a customized version of its Dataray AI Framework and fine-tuned AI models. The framework enables flexible exchange of data and AI models, paving the way for customer-specific AI that continuously learns.