BSc/MSc thesis

Industrial development is evolving towards an increased focus on human-machine collaboration, exemplified by the emergence of Industry 5.0. This new industrial paradigm prioritizes solutions centered around humans and emphasizes resilience and sustainability. A critical aspect of this collaboration is the need for robots to possess more advanced cognitive abilities, allowing for safe co-working environments with humans. Advancements in technology have enabled various methods for monitoring human behavior. Analyzing these behavioral patterns enhances efficiency and collaboration. 

My thesis introduces a setting where monitoring the operator and the robot is possible. It is achieved through a camera, an indoor positioning system, and, in the future, wearable sensors. Indicators, for example, the utilization of participants or waiting times, can represent the collaboration.

Measurements were taken with a specific focus, and the results were analyzed comprehensively—the thesis aimed to establish a connection between human activities and collaborative robots while considering human needs. The process used to achieve this objective is demonstrated through two complex games, and the design of experiments, which were then used to create and do more comprehensive experiments, is also presented.

To plan production and estimate the costs of the production process, manufacturing companies use so-called process or base activity times. These activity times are often estimated or based on field measurements and are dependent on the subjectivity of the engineer who estimates the time effort of the processes. The main subject of the thesis is the development of a data collection and processing system that fits into the Operator 4.0 concept and can be adapted to a brownfield Industry 4.0 manufacturing systems.

Due to the stochastic nature of the problem - different workplace layouts, a wide spectrum of production processes, different work environments, different work habits, additional stress due to measurement and data collection - the objectives include the creation of a generally applicable flexible system.

After reviewing the applicable technologies with a view to making them as adaptable as possible, I concluded that visual information gathering is the easiest way to monitor existing production systems, because it allows both workers and machines to be observed and evaluated.

Optical flow-based displacement detection, clustering of displaced points, tracking of displaced objects with object tracking algorithms and machine learning-based technologies are the most suitable for modelling workers activity and for qualitative and quantitative performance evaluation.

In this study, the analysis of the manufacturing of a product in a one-operator and a two-operator work model at the investigated workplace is presented. By processing the data, the activity performed by the workers and their performance can be measured using indicators assigned to job roles. In addition to human performance, machine utilisation, number of pieces produced and cycle times can also be measured using the framework.