Ng R, Tam WP, Lau CY, Chan J, Ko WF and Chan A, Smart watchdog: intelligent virtual checker based on human factors, HKIE Transactions, Vol. 31, No. 1 (Regular Issue), Article 20220017, 2024, 10.33430/V31N1THIE-2022-0017
Many operations in the railway rely on manual operations, which means that there are human factors involved. Meanwhile, safety is always the first priority on the railways and is the cornerstone of MTR’s success, particularly during these various manual operations. Although there are many safeguards for addressing human factors, these traditional approaches can be complex, may require huge efforts, or may involve additional human resources to perform the checks and balances. Therefore, an innovative methodology called the Smart Watchdog was explored to automatically monitor the manual operations on display screens through the application of artificial intelligence to the videos captured from the display screens. Through a series of artificial intelligence learning, fine-tuning, and testing processes, Smart Watchdog enables accurate and robust detection of the real-time status available on the display screens and generates respective outputs as timely reminders to safeguard potential wrong manual operations. Four representative use cases in MTR are presented in this paper, which shows the significant improvement after the implementation of Smart Watchdog. With this innovative solution, train services can be enhanced and MTR, being one of the leading railway companies, can evolve to be a smart, safe, and efficient railway company.
Keywords:
Artificial intelligence; computer vision; video analytic; human factor; railway; safety
Reference List:
MTR (2024). MTR. [online]. Available at: . [Accessed on 8 August 2024].
Silva VL, Kovaleski JL, Pagani RN, Corsi A and Gomes MAS (2020). The human factor in the smart industry: A literature review. Future Studies Research Journal: Trends and Strategies, 12(1), pp. 87-111.
Dadashi N, Wilson JR, Golightly D and Sharples S (2014). A framework to support human factors of automation in railway intelligent infrastructure. Ergonomics, 57(3), pp. 387-402.
Sethu M, Kotla B, Russell D, Madadi M, Titu NA, Coble JB, Boring RL, Blache K, Agarwal V, Yadav V and Khojandi A (2022). Application of Artificial Intelligence in Detection and Mitigation of Human Factor Errors in Nuclear Power Plants: A Review. Nuclear Technology, 209(3), pp. 276-294.
Ristić-Durrant D, Haseeb MA, Franke M, Banić M, Simonović M and Stamenković D (2020). Artificial intelligence for obstacle detection in railways: Project SMART and beyond. In: Dependable ComputingEDCC 2020 Workshops. Munich: Springer International Publishing. pp. 44-55.
Tang R, De Donato L, Besinović N, Flammini F, Goverde RM, Lin Z, Liu R, Tang T, Vittorini V and Wang Z (2022). A literature review of Artificial Intelligence applications in railway systems. Transportation Research Part C: Emerging Technologies, 140, 103679.
Arslan B and Tiryaki H (2020). Prediction of railway switch point failures by artificial intelligence methods. Turkish Journal of Electrical Engineering and Computer Sciences, 28(2), pp. 1044-1058.
The Linux Kernel Archives (2016). Part I - Video for Linux API. [online]. Available at: . [Accessed on 8 August 2024].
Windows App Development Documentation (2021). Recommended 8-Bit YUV Formats for Video Rendering. [online]. Available at: . [Accessed on 8 August 2024].
OpenCV (2024). OpenCV. [online]. Available at: . [Accessed on 8 August 2024].
Sammut C and Webb GI (2011). Encyclopedia of Machine Learning. Boston: Springer.
Khan R, Rahmani H, Shah SAA and Bennamoun M (2018). A Guide to Convolutional Neural Networks for Computer Vision. Morgan & Claypool Publishers.
Lim J (1990). Two-dimensional signal and image processing (Prentice Hall signal processing series). Englewood Cliffs, N.J.: Prentice Hall.
Apostolico A and Galil Z (1997). Pattern matching algorithms. New York: Oxford University Press.
Albalate A and Minker W (2011). Semi-supervised and unsupervised machine learning: Novel strategies (ISTE). London: ISTE.
