Thesis supervisor: Tamás Ruppert
Location of studies (in Hungarian): University of Pannonia Abbreviation of location of studies: PE
Description of the research topic:
The Industry 5.0 which is the progression of industry 4.0 aims at bringing the humans back to the shopfloor. Under this new paradigm the robots or artificial intelligence and humans work in collaboration with each other. The idea is to leverage the accuracy of the machine or the robot with the innovative potential of the human worker.
However, a Human-AI or a Human- Cobot working in collaboration is faced with several challenges. The central challenge here is the difficulty in establishing reliable communication channel between Human to AI and vice versa. Conventional enablers of such communication include sensors such as Heart rate Variability (HRV), Electro encephalograms(EEG), Galvanic skin response(GSR) etc. The main operating mode of such sensors is to estimate the cognitive load of the operator based on the sensor readout values. Under high Cognitive Loads, the operator situational awareness drops which may result in sub optimal decisions on the part of the operator. As such the aim of an efficient collaborative AI which is hence forth referred as “virtual assistant” is to intervene under such high Cognitive load situations by providing necessary support to the operator to help improve KPIs in general.
The development of reliable sensor-based monitoring systems can be particularly challenging because of several inherent challenges related to variations caused by operator differences, sensor noise resulting from improper attachment of the sensor etc. At the same time different workers may react differently to different types of input modes. This corresponds to the fact that some workers may be more perceptive to visual instructions as against audio instruction. So, an in an ideal state a virtual assistant supported operator system should support multiple modes of information dissemination.
Login
