Multicriteria task classification in human-robot collaborative assembly through fuzzy inference

Abstract

The advent of new technologies and their implementation in manufacturing is accelerating the progress of Industry 4.0 (I4.0). Among the enabling technologies of I4.0, collaborative robots (cobots) push factory reconfiguration and prompt for worker empowerment by exploiting the respective assets of both humans and robots. Indeed, human has superior dexterity, flexibility, problem-solving ability. Robot excels in strength, endurance, accuracy and is expendable for risky activities. Therefore, task assignment problem in a production line with coexisting humans and robots cannot limit to the workload balancing among workers but should make the most of everyone respective abilities. The outcomes should not be only an increased productivity, but also improved production quality, human safety and well-being. Task assignment strategy should rely on a comprehensive assessment of the tasks from the viewpoint of suitability to humans or robots. As there are several conflicting evaluation criteria, often qualitative, the study defines the set of criteria, their metrics and proposes a method for task classification relying on Fuzzy Inference System to map each task onto a set of collaboration classes. The outcome of the study is the formal description of a set of evaluation criteria with their metrics. Another outcome is a Fuzzy Classification procedure that support production managers to properly consider all the criteria in the assignment of the tasks. The proposed methodology was tested on a case study derived from a manual manufacturing process to demonstrate its application during the process planning.