CONDUCTOR is a digital construction management platform for safer, smarter, and more sustainable construction works. Consisting of 3 modules, it adopts emerging technologies such as AI, IoT, digital twin with AR integration. The Progress Management module enables users to monitor construction progress through digital twin. Calibration sticker helps positioning and reality mapping and reveals the digital twin with AR and computer vision technology. It greatly enhances the transparency of construction progress and ensures quality. The Safety module combines CCTV and AI analytics to expediate site safety inspection. AI helps detect worker safety by identifying individuals without proper personal protective equipment. SOS function and danger detection in the smart watch allows immediate assistance. The Sustainability module utilises IoT technology to consolidate environmental and biometric data in sites. Climate risks including heat stress and their impacts can be detected. The smart watches collect workers’ biometric data for monitoring their health status. The module can also track the material supply chain, real-time pollution and carbon emission. As a collaboration SaaS platform, it allows the adoption of emerging technologies into construction like plug-and-play, facilitating the digital transformation of the construction industry and supporting sustainable construction.
The Escalator Clearance Inspection Device (ECID) is a patented design that offers a standardized, accurate and automated inspection solution to ensure all clearances between escalator steps are within the required standard. Conventionally, maintainers measure clearances manually with a feeler gauge, which is time-consuming and may yield inconsistent results. The novelty of the device lies in its vision module, algorithm, illumination module, and fixture. The vision module, empowered by the multi-view algorithm, enables 3D vision for clearance measurement at various locations on an escalator. The global disparity matching and fitting algorithm is designed dedicated for old escalators, which often trapped excessive dust, debris and are affected by wear and tear issues. The illumination module adopts the co-focus lens optical design to provide sufficient lighting for image capturing, enabling the exposure time for each image frame down to 200 μs, while optimizing the power consumption allowing the portable device to run on battery. The adaptive and adjustable fixture enables all modes of operation for various escalator brands. The ECID brings smart maintenance to escalator safety as a computer vision based renovation that is capable of automatically generating measurement results on a moving escalator in a contactless manner accurately and efficiently.
MTR Light Rail network is unique as it shares space with road traffic, train captains are required to pay much attention to crossing pedestrians and obey stop signals to avoid serious incidents. To address this human factor risk, the Smart Railway Signalling and Foreign Object Detection System is a fresh set of eyes, providing precautionary warnings of approaching stop signals, pedestrians, and obstacles in the track area. This is a lightweight yet impactful innovation, it has a camera, edge-computing processor, and alert indicator box installed in the driving cab. Leveraging the video analytics (VA) technologies with a substantial amount of dataset training, it could effectively adapt to the highly variable trackside environment, allowing the model to accurately recognize people, signal aspects, and trackside objects. Then, provide timely alerts to drivers when dangerous situations arise, i.e., a stop signal is ahead, the presence of obstacles or jaywalkers in front of a train set. Additionally, when detecting pedestrians or animals ahead, the automatic sounding of the train bell can prevent accidents by alerting them of incoming Light Rail Vehicles. Overall, this application increases staff awareness and promotes MTR’s safety culture, ultimately enhancing the train operational and public safety within the MTR network.
“Depot shunting and coupling are critical activities in the operation of a railway depot. This involves relocating rail vehicles between different pits for maintenance and stabling purposes. As a unique characteristic of railway business, such operation could only be performed during non-traffic hours from 1:30 to 4:00 am. Physical fatigue, blind spots, and lighting issues are the major challenges of depot shunting which could even result in damages and injuries. By utilizing video analytics and depth sensing technology, our application scans for the kinematic envelope and detects obstacles and track configuration in a railway environment. Compared with other existing technologies, such as radar, various applications have been deployed for road vehicles of similar nature. However, the direct application remains a challenge in the railway context. Some objects can be close to the train and tracks but still considered safe, e.g., signs, posts, and other infrastructure. In such an environment, our application can separate detection from generic objects and generate accurate warnings in real time, enhancing the overall safety and awareness of depot shunting and coupling in railway depots.”