SmartTestudo: Adaptive Monitoring for Cyber-Physical Systems A proof of concept applied in an industrial environment simulation with TurtleBot3

Abstract

With the advent of Industry 4.0, Cyber-Physical Systems (CPS) have started to become more adopted in different kind of industries, increasing productivity and efficiency. This rapid adoption has led various companies to develop their in-house solutions to stay competitive and try to dominate the market in selling robotic systems. This division in different ways of building CPS led to standardization problems and made making new innovations and discoveries harder. To solve this issue of a lack of a framework and solid bases, two researchers developed the Robot Operating System (ROS) to have a common baseline to develop CPSs. With new advancement in hardware technologies ROS had it’s bottlenecks, so a new version called ROS2 was developed. This brought new possibilities to improve CPSs and overall productivity.

As CPS became more adopted, new problems started to arise, mainly regarding safety, maintenance and monitoring. Each CPS produced big amounts of data that isn’t always useful and risks to clutter the system and robot performance. Solutions to this issue have been proposed in existing research, and the objective of this research is to analyze the existing literature to identify potential limitations and issues with the current state of the art, and propose an innovative way of monitoring CPS with a technique called adaptive monitoring.

To achieve this objective, a design science approach has been devised, with three iteration, one regarding the analysis and the second two regarding the development and verification of the artifact. After each iterations results are presented and discussed, with the artifact improved.

The proposed artifact consists in an adaptive monitoring system which inner workings are based on the MAPE-K loop, where the knowledge is specified by the user with the use of a Domain Specific Language (DSL) and the values of the frequency of each topic is optimized and adjusted with the adoption of the Quality Performance Model (QUPER).

The findings of this research highlight the limitations faced by existing literature and how the current artifact approach comes in aid.