Efficient and accurate monitoring of production in dynamically changing environments is one of important requirements for effective management and control of the workshop. Thesis describes development of distributed vision system for the monitoring purposes. The generic concept consists of network cameras and PC computer and appropriate software for image processing. Vision systems have been widely exploited, usually for quality control and inspection applications, detecting presence or absence, measuring distances and angles, optical sorting and counting, but their application for shop floor control and management is still at an early stage. So far, reasons for that were in their high complexity and high costs. MVSfMWS (Machine vision system for manufacturing work system surveillance) developed in this work offers the ability to view live images with automatic identification of their characteristics. Also MVSfMWS allows the shop floor manager to view the current state of the producton line, enabling him to monitor staffing levels, maintain health and safety standards and visually assess production problems before stoppages occur. Another advantage of this system is in its mobility and reconfigurability, which could open new areas of implementation, as for example distributed assembly monitoring, distributed visuel communication, exchange visual data, etc.

Currently, Engineer-To-Order manufacturing companies are under high pressure from global competition and turbulent market excitations, resulting in increasing product variety, individualization, and shortened time frames. Additionally, manufacturing companies are faced with the increasing complexity of products, production structures, and processes. It is obvious that manufacturing companies with structural rigidity, a deterministic approach to decision-making in a stochastic environment, hierarchical allocation of competencies, and insufficient communication and exploitation of expertise cannot achieve these requirements at an appropriate level. The thesis builds on a hypothesis that better management of complexity in the ETO manufacturing environment can be achieved through real-time management methods and tools such as ubiquitous and calm technologies, sensor networks, and real-time traceability, visibility and decision support systems. To address the hypothesis, the concept of a ubiquitous autonomous manufacturing system is developed, which represents an advanced structure for further optimization of ETO manufacturing environment while embracing emerging information technologies and embedded system technologies to build such structure.