Development of IoT-based machine learning application for data anomaly detection with in a smart manufacturing plant

Kukuni, TGMarkus, EKotze, BAbu-Mahfouz, Adnan MI2023-02-262023-02-262022-08Kukuni, T., Markus, E., Kotze, B. & Abu-Mahfouz, A.M. 2022. Development of IoT-based machine learning application for data anomaly detection with in a smart manufacturing plant. NeuroQuantology, 20(10). http://hdl.handle.net/10204/126061303-5150Doi: 10.14704/nq.2022.20.10.NQ55352http://hdl.handle.net/10204/12606The application of machine learning in resolving complex cyber-security challenges in smart manufacturing plant is growing. Network intrusion and anomaly detection is posing high risks in sensory data integrity and optimisation of processes leading to high efficiency and high profits within smart manufacturing plants. This research paper makes use of interquartile range algorithm for the detection of anomalies. The data is collected within a 5- hour period and is transmitted to Google sheets via WiFi connectivity. However, the data transfer requires the user to permit access to the google account for this process to take place. After the addition of errors for every 11th entry, the file is resent back to Raspberry PI for the execution of interquartile range algorithm. Once the results are obtained, the results file is transmitted via WiFi connectivity to the output monitor. This research results demonstrates that if the data collected is higher or lower than the required threshold (11- 14oC for temperature and 45-50% for humidity) the system will automatically detect and flag the anomaly. This paper therefore concludes that the use of interquartile range algorithm for anomaly detection based on sensory data is relevant and efficient for such an investigation.FulltextenIntrusion Detection SystemMachine learningCyber-securityInternet of ThingsIoTSensory dataSmart Manufacturing PlantDevelopment of IoT-based machine learning application for data anomaly detection with in a smart manufacturing plantArticleKukuni, T., Markus, E., Kotze, B., & Abu-Mahfouz, A. M. (2022). Development of IoT-based machine learning application for data anomaly detection with in a smart manufacturing plant. NeuroQuantology, 20(10), http://hdl.handle.net/10204/12606Kukuni, TG, E Markus, B Kotze, and Adnan MI Abu-Mahfouz "Development of IoT-based machine learning application for data anomaly detection with in a smart manufacturing plant." NeuroQuantology, 20(10) (2022) http://hdl.handle.net/10204/12606Kukuni T, Markus E, Kotze B, Abu-Mahfouz AM. Development of IoT-based machine learning application for data anomaly detection with in a smart manufacturing plant. NeuroQuantology, 20(10). 2022; http://hdl.handle.net/10204/12606.TY - Article AU - Kukuni, TG AU - Markus, E AU - Kotze, B AU - Abu-Mahfouz, Adnan MI AB - The application of machine learning in resolving complex cyber-security challenges in smart manufacturing plant is growing. Network intrusion and anomaly detection is posing high risks in sensory data integrity and optimisation of processes leading to high efficiency and high profits within smart manufacturing plants. This research paper makes use of interquartile range algorithm for the detection of anomalies. The data is collected within a 5- hour period and is transmitted to Google sheets via WiFi connectivity. However, the data transfer requires the user to permit access to the google account for this process to take place. After the addition of errors for every 11th entry, the file is resent back to Raspberry PI for the execution of interquartile range algorithm. Once the results are obtained, the results file is transmitted via WiFi connectivity to the output monitor. This research results demonstrates that if the data collected is higher or lower than the required threshold (11- 14oC for temperature and 45-50% for humidity) the system will automatically detect and flag the anomaly. This paper therefore concludes that the use of interquartile range algorithm for anomaly detection based on sensory data is relevant and efficient for such an investigation. DA - 2022-08 DB - ResearchSpace DP - CSIR J1 - NeuroQuantology, 20(10) KW - Intrusion Detection System KW - Machine learning KW - Cyber-security KW - Internet of Things KW - IoT KW - Sensory data KW - Smart Manufacturing Plant LK - https://researchspace.csir.co.za PY - 2022 SM - 1303-5150 T1 - Development of IoT-based machine learning application for data anomaly detection with in a smart manufacturing plant TI - Development of IoT-based machine learning application for data anomaly detection with in a smart manufacturing plant UR - http://hdl.handle.net/10204/12606 ER -26457