Deep Tech and Artificial Intelligence for Worker Safety in Robotic Manufacturing Environments

The main objective of the work is to prevent hazardous situations for workers in automated and robotized agile-production environments by means of a new open-source FIWARE-based Smart Factory Service. New system will implement Deep Learning models based on Convolutional Neural Networks, Deep Reinforcement Learning (Double Deep Q Networks) and Recurrent Neural Networks for the detection of anomalous behavior patterns and the prediction of the workers’ level of attention, fatigue and distraction. For this purpose, FIWARE IoT Agents will gather video, audio and ECG data coming from camera and microphone arrays deployed in the environment, as well as ECG incorporated in wearables.