**1. Introduction**

Lean manufacturing is a collection of synchronized methods and principles for organizing and controlling production sites in a technology-independent way to reach shortest lead time with minimum costs and the highest quality [1]. Due to the evolution of technology and its introduction in the factories, industry and manufacturing processes have changed and evolved throughout the industrial revolutions, from the introduction of mechanical production facilities powered by water and steam to the current cyber-physical production systems (CPPSs) and intelligent automation, keys of the Fourth Industrial Revolution, also known as "Industry 4.0" [2]. The integration of automation technologies and lean manufacturing is called "lean automation" [3]. These CPPSs monitor the physical processes, make decentralized decisions, and trigger actions, communicating and cooperating with each other and with humans in real time. Networked machines perform more efficiently, collaboratively, and resiliently [4].

The intelligent automation also requires the use of autonomous machines or robots, which are controlled by the CPPS and the humans. The main objective is to increase productivity and safety, which were traditionally limited by manual processes. The safety conditions have limited the use of robots in industrial environments, as they were isolated from people and some tasks were not affordable. The introduction of process automation and intelligent collaborative robots results in a rapid increase in productivity, major material and energy savings, and safer working conditions (repetitive and dangerous tasks can be done by robots). Robots provide versatility and flexibility; thus they are perfect substitutes for a skilled workforce for some repeatable, general, and strategically-important tasks [5]. Moreover, if robots' capacities are combined with humans' qualities, cost-effective productivity can be guaranteed [6].

Up until fairly recently, the information about any physical object or process was relatively inseparable from the physical object or process itself [7]. Digital data and artificial intelligence allow the dematerialization and the coexistence of the real factory with digital twins [8], where manufacturing processes are virtually simulated, monitored, and controlled. While at first, this digital twin was merely descriptive and static, in recent years it has become actionable and experimentable [9]. The digital information related to a physical system can be created as an entity itself. This means that there is a mirroring or twinning of systems between what exists in real space to what exists in virtual space and vice versa [7]. For this purpose, access to very realistic models of the current state of the process is necessary [10]. The use of sensors and 3D visualization technologies, such as virtual and augmented reality, makes this connection possible and facilitates the interaction with humans [11,12]. The digital twin is fed with the data from the sensors, PLCs, controllers, etc., and the 3D environment is visualized using the 3D glasses. A digital twin not only allows a static perspective at design stage, but also a real-time synchronization and optimization of the virtual object [13].

At this point, robotized processes can be mirrored using digital twin models during the design, implementation, and operation steps. Each robot manufacturer has its own simulation environment for cell design and program testing, but the challenge arises when the same cell contains robots from different manufacturers working collaboratively with humans. Multiple robots can perform tasks in parallel, speed up the execution time, and improve system performance [14]. Thus, this work presents a novel methodology for process automation design, enhanced implementation, and real-time monitoring in operation. The proposed approach is based on creating a digital twin of the manufacturing process with an immersive virtual reality interface to simulate and analyze the layout (the physical location of the elements) and to determine whether robots and other components are suitable. The results in this virtual testbed will easily permit modifications in the original design before the physical implementation, presenting a far more cost-efficient solution. In addition, once the new process has been implemented, the digital twin can be efficiently used for operator training, real-time process monitoring, and feasibility study of future optimizations, resulting in a novel and intelligent mirror with high potential benefits. As it is not a simple replica, it is able to process and understand data, and automatically react to changes according to them. The innovation of the proposed approach is that it combines design, feasibility studies, virtual and real commissioning, training, and monitoring of multi-robot cells in a unique application, which implies a cost-effective and affordable solution for manufacturing industries of all types. The theoretical outcome is the proposed methodology for robot-based automation, while the practical is the digital twin framework with an immersive virtual reality interface which is used as a testbed environment.

The proposed approach is validated in a real-life case study that provides a solution for the assembly of parts, demonstrating that the use of the digital twin-based methodology is feasible. Not only is the result of the proposed approach more visual, thanks to the integration of virtual reality, it also reduces costs and increases the productivity, as proven with real operations. Therefore, it is very likely to find potential applications in a number of different areas and multiple industries to create flexible and easily reconfigurable production lines.

The rest of the paper is organized as follows: Section 2 reviews the state of the art and previous work, Section 3 presents the proposed approach, Section 4 contains the use case, and main conclusions are found in Section 5.
