Project

2023

1. 

   Depth Camera-Based Fall Detection

   In CUHK AIoT Lab, 2023

   Abs

   Our system incorporates a depth camera that creates a 3D representation of the environment, detecting falls based on changes in a person’s relative position. By capturing and analyzing depth information, we can identify the spatial relationships and movements associated with a fall event. This depth-based approach enables us to accurately discern falls from other activities, providing a robust solution for real-time fall detection. Furthermore, the depth camera’s functionality is not significantly impaired by variations in lighting conditions, making it a reliable component for 24/7 monitoring within settings like nursing homes.

2. 

   Elderly Fall Detection System Based on Thermal Sensor

   In CUHK AIoT Lab, 2023

   Abs

   In addition to the depth camera, our system employs a thermal sensor to further enhance the accuracy of fall detection. The thermal sensor identifies heat signature changes when a person falls and remains on the floor. It can detect the sudden change in thermal energy associated with a fall, as well as the continued presence of a person on the floor, thereby signaling a potential fall incident. The thermal sensor is highly reliable and, importantly, it respects personal privacy as it does not capture identifiable images of individuals. This feature has been instrumental in enhancing elderly care, enabling swift responses to falls while respecting the privacy of the individuals under care.

3. 

   Hydrological and Reservoir Monitoring System of Hunan Province

   In Hunan Provincial Hydrological and Water Resources Survey Center, China, 2023

   Abs

   Our team has successfully deployed a network of 10,000 reservoir monitoring nodes across Hunan Province. This large-scale undertaking has significantly enhanced our ability to collect and manage multimodal sensor data, including RGB, osmotic pressure, and rainfall. This data provides the backbone for comprehensive analysis of various hydrological tasks such as flood warning, reservoir water level regulation, water quality monitoring, and surface or groundwater resource monitoring. In addition, we have addressed challenges related to scheduling and distributed system coordination in this large-scale system, which has enhanced the efficiency and stability of data collection and processing.

2022

1. 

   Railway Bridge and Culvert Collision Warning System with Overhead Clearance Protection

   In Hunan Zhongke Haili Technology Co., Ltd, China, 2022

   Abs

   Overheight and oversized vehicles frequently collide with overhead clearance frames and bridges, posing significant safety risks to railway tunnels and bridges, and jeopardizing railway traffic safety. Our solution tackles this issue head-on. We install LiDAR sensors and vibration detectors on lamp posts to monitor vehicle speed and height. Key features of our system include: Collision Warning: Alerts are issued upon collision risk detection. Accurate Detection: The system ensures no incidents are missed. Overheight Pre-warning: Proactive warnings are sent to overheight vehicles. Off-grid Capability: Operates in areas without network or electricity, powered by solar energy. System Self-check: Regular self-checks are performed by the main control board. This solution protects railway infrastructure and enhances safety by preventing potential accidents.

2021

1. 

   Automated Warehouse Packaging Solution

   In Hunan Zhongke Haili Technology Co., Ltd, China, 2021

   Abs

   Our Automated Warehouse Packaging Solution revolutionizes the packaging process in warehouses by replacing manual labor with a fully automated, unmanned system. The solution integrates Automated Guided Vehicles (AGVs), automatic hooks, spray coders, robotic arms, labeling and wrapping machines, and strapping devices. The AGVs transport goods from the semi-finished goods warehouse to the packaging area and deliver the packaged goods to the finished goods warehouse. The automatic hooks, labeling, and wrapping machines streamline the loading, labeling, palletizing, and wrapping processes. The system also employs machine vision technology to ensure the accuracy of label application and spray coding. This comprehensive automation solution significantly enhances warehouse efficiency and packaging accuracy.