Adaptive Robot Framework: Providing Versatility and Autonomy to Manufacturing Robots Through FSM, Skills and Agents

Abstract

The main conclusions that can be extracted from an analysis of the current situation and future trends of the industry,in particular manufacturing plants, are the following: there is a growing need to provide customization of products, ahigh variation of production volumes and a downward trend in the availability of skilled operators due to the ageingof the population. Adapting to this new scenario is a challenge for companies, especially small and medium-sizedenterprises (SMEs) that are suffering first-hand how their specialization is turning against them.The objective of this work is to provide a tool that can serve as a basis to face these challenges in an effective way.Therefore the presented framework, thanks to its modular architecture, allows focusing on the different needs of eachparticular company and offers the possibility of scaling the system for future requirements. The presented platform isdivided into three layers, namely: interface with robot systems, the execution engine and the application developmentlayer.Taking advantage of the provided ecosystem by this framework, different modules have been developed in order toface the mentioned challenges of the industry. On the one hand, to address the need of product customization, theintegration of tools that increase the versatility of the cell are proposed. An example of such tools is skill basedprogramming. By applying this technique a process can be intuitively adapted to the variations or customizations thateach product requires. The use of skills favours the reuse and generalization of developed robot programs.Regarding the variation of the production volumes, a system which permits a greater mobility and a faster reconfigurationis necessary. If in a certain situation a line has a production peak, mechanisms for balancing the loadwith a reasonable cost are required. In this respect, the architecture allows an easy integration of different roboticsystems, actuators, sensors, etc. In addition, thanks to the developed calibration and set-up techniques, the system canbe adapted to new workspaces at an effective time/cost.With respect to the third mentioned topic, an agent-based monitoring system is proposed. This module opens up amultitude of possibilities for the integration of auxiliary modules of protection and security for collaboration andinteraction between people and robots, something that will be necessary in the not so distant future.For demonstrating the advantages and adaptability improvement of the developed framework, a series of real usecases have been presented. In each of them different problematic has been resolved using developed skills,demonstrating how are adapted easily to the different casuistic.