To enhance the effectiveness of railway safety regulation, the Railways Branch (RB) of EMSD has rolled out the following innovation and technology projects:
In July 2020, the RB launched a Light Detection and Ranging (LiDAR) detection system as a pilot project in a MTR station with the collaboration of MTRCL. This project aims to reduce the escalator accident caused by passengers using escalator improperly. When the system detects passengers carrying bulky items, bicycles or prams who are about to use escalators, it will issue specific broadcast messages to remind them to use lifts. The system has been installed at MTR Choi Hung Station in March 2021. When the system is put into operation, passengers’ safety will be further enhanced in using railway facilitates.
Besides, in order to strengthen the security and security of passengers who carrying dangerous goods (eg. balloon, long items and liquefied petroleum gas (LPG)) entering station premises, the aforesaid LiDAR detection system was installed at Jordan Station for trial test in Oct 2021. When the system detects the target passengers, it will issue specific broadcast messages and alert station personnel to provide suitable assistance in a bid to keep passengers safe.
To early identify incident trends and potential risks of the railway systems for enhancing the railway safety, Artificial Intelligence (AI) technology is applied to analyze the historical incident data.
In November 2021, the RB launched a pilot AI analytic data system, which was established a centralized railway incident database that can understand the semantic meaning of railway incidents by Natural Language Processing (NLP) for incident enquiry and identify the incident trend and correlations of incidents to analyze the cause of incidents more efficiently. The system would be further enhanced to analyze variety of railway data, such as maintenance records, asset information and real-time IoT sensor data, etc., so that potential risks and possible failures of railway systems can be predicted by the AI system.
Smart Railway Track Intrusion Detection System using the Artificial Intelligence and Video Analytics technologies for the real-time detection of intrusion objects infringing into railway track areas. Instant SMS (Short Message Service) will be sent out from the on-site edge device to the MTRCL’s Traffic Controller at Tsing Yi Operations Control Centre to take follow-up actions promptly. It will also provide the snapshot image of intrusion objects and the live video stream of the incident site.
The system was commissioned at Tai Wai Station (East Rail Line) and Che Kung Temple Station (Tuen Ma Line) in December 2021. It has been used for monitoring any object has incidentally fallen to the railway track areas due to the demolition works of Pavillia Farm III. Besides, the system was installed at Shatin (near Sha Tin Rural Committee Road flyover) and Hung Hom (near Chatham Road North flyover) in December 2021 to monitor any vehicle has accidentally fallen onto the railway track area. This system is served to prevent the occurrence of collision with the train when the vehicle falls onto the rail track area.
To facilitate proactive monitoring on railway safety, the RB commenced to develop an AI data analytics system for regular analysis of MTRCL's concession applications for delayed maintenance activities. The system will make use of state-of-the-art AI technologies to recognize and learn patterns, trends and correlations among key attributes in concession applications for delayed maintenance activities to identify potential maintenance planning loopholes and risks at an earlier stage. The RB could then base on the information to formulate smarter and forward-looking planning of assessments and inspections on risk-based approach.
The Tram Speed Alert System is designed to permit rea-time monitoring of the tram speed on road with different speed limits and generate real time alert to tram driver. The technical study of this on-board real-time alert system for the assistance on speed control of the tramcar commenced in September 2021. The system will make use of Global Navigation Satellite System (GNSS) for enabling a tram to compare against the speed limits along the tram lane in database. Alert would be generated to tram driver for slowing when the tram speed exceeds the pre-set speed limit, thereby enhancing the operation safety of tram.
To prevent tram from derailment incident after tram wheels hitting a small / foreign object on tram track in daytime, nighttime and bad weather, the technical study of an on-board real time object detection and alert system commenced in September 2021. The system works in real time to identify if the object in front of a tram is a potential hazard to tram operation, then give out an alarm to alert the tram driver to stop the tram. The system will make use of Light Detection and Ranging (LiDAR) technology to detect if there is an object in rail groove on ahead of a tram that may pose a potential hazard to tram operation. Alert would be generated to tram driver for action when high-risk object is detected, thereby enhancing the operation safety of tram.
SPAD Alert Device using Artificial Intelligence (AI) and video analytics technologies for Automated People Mover (APM) Driver in Hong Kong International Airport. It is a portable device and easy to install on the APM train. It is used for real-time detecting the improper driving behaviours (fatigue, yawning or inattention) of the driver, and providing the alert to the driver and APM controller when dangerous situations (speeding, red signalling light is on or any objects detected on the guideway) are detected during the manual drive by the maintenance staff. Besides, the recorded videos were encrypted and saved on the device. It can be played back in pairs with guideway conditions & driver image and device records for incident investigation.
It can prevent the occurrence of accidents caused by driver’s careless driving; the device will give early warning to the concerned driver on potential safety risks and achieve operational safety by reducing the number of accidents and minimizing downtime of the APM system. The system was launched and put on trial in July 2021.
A new Record Management System was rolled out in October 2021 using Blockchain technology for storing MTRCL’s submissions to the Safety and Security Coordinating Committee (SSCC) and Trackside Safety and Security Committee (TSSC). The decentralized approach with distributed ledgers recording structure of blockchain provides excellent features for the new BRMS to ensure authenticity, integrity, reliability, usability of digitised records and unalterable audit trail logs. The BRMS provides secure storage of the submissions and protects against unauthorised access, alteration and deletion of records.