Cognitive Robotic Cell through Embedded ML Vision (COGNICELL)

Challenge 4: Manufacturing Industry

Summary: The project aims to develop a new solution for embedded vision using AI/ML technology in order to improve the productivity and quality of industrial robotic cells. The solution will leverage commercial-off-the-shelf (COTS) components and open source software libraries for cost effectiveness and flexibility. The aim is to assure early and real-time defect detection in manufacturing and enable vision-based control of a robotic arm in a variety of manufacturing operations and scenarios. An existing hardware platform is available which currently uses an expensive and constrained industrial vision camera (COGNEX 7010) which should be replaced by the new system. The system will use existing StairwAI libraries for image-based defect detection in manufactured parts which will be adapted in accorandance to the particular context of the robotic cell, collected training datasets and domain knowledge from experienced automation engineers. The system will be benchmarked according to the time of detection, detection accuracy and complexity. Implementation will be carried out on an existing prototype platform that includes both an ABB IRB120 6-DoF robot arm and a conveyor based drive with frequency converter and Siemens S7-1200 PLC control system. This will enhance the feasibility of the solution towards the industrial partners and customers of Asti Automation. Bridging between commercial proprietary technologies and the new open source solution will be done through open standards interfaces and communication protocols such as Node-RED, OPC UA and MQTT. The project will be able to be replicated at scale and deployed for various industry verticals that increasingly rely on the adoption of robots in their manufacturing/production workflows such as logistics, food industry, automotive and electronics. Implementation of the solution will have both economic and environmental benefits through reduction of re-manufacturing operations, waste avoidance and increased throughput.

StairwAI project has received funding from the European Union’s Horizon 2020 research and innovaon programme under grant agreement No. 101017142