Heterogeneous aggregated data integration framework for the internet of things and industry 4.0 applications

Abstract

The emergence of the Internet of Things (IoT) empowered heterogeneous distributed systems development for various application domains using smart sensors, actuators, embedded devices, and diverse transmission protocols. Collaborative integration of such systems poses interoperability issues at different automation levels, requiring standardized unification and connectivity between the system components. The current industrial practice involves Open Platform Communications Unified Architecture (OPC UA) industrial standards for resolving these issues. However, OPC UA is encapsulated for commercial purposes and caters to a specific data transfer protocol only. This work aims to design and develop a unified framework for addressing the interoperability of heterogeneous systems and provide a simplified uniform interface to integrate the aggregated multi-source data with consumer platforms. Nine physical nodes were developed as the heterogeneous event generator through different standards such as Modbus Remote Terminal Unit (RTU) through RS232 and RS485 link, Modbus Transmission Control Protocol (TCP) in Raspberry Pi (RPi), NodeMCU, and Delta PLC, Wireless Sensor Node (WSN) in RPi through 433.4 MHz Radio Frequency link, Hypertext Transfer Protocol (HTTP), Representational State Transfer (REST), and Message Query Telemetry Transport (MQTT) protocol. The framework incorporates an IEC 625141 Compliant OPC UA aggregator module to aggregate data from these heterogeneous sources and a REST-capable middleware module to provide a uniform interface to share these aggregated data with end clients. A Standalone RESTful web and cloud services were implemented as the end client to acquire middleware data for remote supervision, perform Machine to Machine (M2M) communication, and store data in a remote cloud database for further processing. Performance assessments of the developed framework have been conducted by deploying in a windows-based Intel NUC PC and Linux-based RPi model 3B+ under two test case scenarios. Data from 63 sensors within nine heterogeneous physical nodes were transferred via the above protocols (Modbus RTU RS232 and RS485 link, Modbus TCP, WSN via RF 433.4 MHz, HTTP, RESTful HTTP, and MQTT) through the aggregator and middleware module to the web consumer platforms. The experimental results indicated a 100% successful data transfer rate at acceptable platform resource utilization with an average of 30.56% in NUC and 40.87% of CPU consumption in RPi, 72.61 MB (0.59%) in NUC, and 87.08 MB (9.94%) in RPi memory usage, 4.9185 MB/s in NUC and 1.4968 MB/s network bandwidth utilization in RPi at 2.9044W average power consumption. The developed system also exhibits an average of 1.05 s in NUC and 1.15 s in RPi data aggregation service time, 0.37 s in NUC and 0.49 s in RPi client data processing time, 0.43 s in NUC and 0.46 s in RPi cloud data transfer time. The characterization of the developed framework was validated by integrating with the proprietary and open-source infrastructures enabling affordable communication between the heterogeneous data sources with the end clients in an efficient and simplified manner. The developed framework shows salient suitability with IoT and I4.0 applications, which would be beneficial to integrate with small and medium industrial entities.