Session: CIE-05-01CIE Graduate Student Poster Symposium

Paper Number: 74690

Start Time: August 18, 10:00 AM

74690 - A Data-Driven Approach of Detecting Human Fatigue for Adaptation in Human-Robot-Collaboration 

Industrial Revolution 4.0 has witnessed the advancement of industrial robots to a collaborative robot with cognitive capabilities. Although, there is a rise in automation there are still tasks that are labour demanding. Human fatigue is induced in a human operator after several hours of work due to physical exertion, repetitive work, along other factors. The effects of these can cause lowered productivity, reduced quality of work also can lead to accidents or mishap. In the long run, the effects can escalate and develop ‘Work-related musculoskeletal disorder’ (WMSD). It makes it significant to detect fatigue as this not only hinders socially but has financial effects as well. In perspective to the human operator, it is noteworthy for cobot to adapt according to the human state. Fatigue is a sensation based according to an individual and to quantify the level of human fatigue to adapt precisely is a challenging and open research question. To achieve a seamless interaction between human and robot, it is important for the robot to analyze the human state accurately in real-time. To overcome the challenges an understanding of fatigue development and the modelling of fatigue detection to set a benchmark for multisensory input. An analysis of different multisensory input will collect the data. After the data collection, data exploration will include data cleaning, feature extraction and analysis. To select the optimum variables with help of different dimensionality reduction algorithm. Further, the functional relationship between the different variables was divulged with the help of different machine learning algorithm as well as to predict the human fatigue level. The detection of fatigue through the cobot will help in dynamically adapting to the human state. An experimental analysis to validate which will include a collaborative work environment. The cobot will be modelled to adapt to the human state by considering different demographic variables of an individual in an experiment. Thus, the research will help the operators in a collaborative environment to develop a seamless human-robot interaction (HRI) or Human-Robot Collaboration (HRC).

Presenting Author: Arsalan Lambay Cardiff University

Authors:

Arsalan Lambay Cardiff University