| Project Reference: | ITP/050/20LP |
| Project Title: | Novel IoT and Multi-modal Analytic Technologies for a Smart City |
| Hosting Institution: | The Hong Kong University of Science and Technology (HKUST) |
| Abstract: | Hong Kong aspires to be a leading smart city in the world. Internet-of-things (IoT) is an indispensable element to enable such a vision. With the advent and penetration of broadband 5G, WiFi and LoRA /NB-IoT technologies, the IoT sector has been experiencing phenomenal growth in recent years. Devices sensing WiFi, BLE and visual signals are increasingly deployed in the city. Although these sensing data collected in a venue are often correlated in nature, they are usually analyzed in isolation in cloud, forming data silos. It has been a critical industrial challenge to bridge these silos for effective multi-modal learning and mining to support applications such as object recognition, people sensing, user analytics, especially customer journey. In order to achieve such objective, front-end IoT device should also be equipped with intelligence,the so-called edge AI. One such example is IoT camera for carpark, which is currently facing installation and maintenance bottleneck for large-scale deployment. With strong support from IoT industry, we propose to develop novel IoT and multi-modal analytics technologies to overcome the above challenges. To achieve cost-effective smart carpark, we will design a novel AI-based panoramic camera to recognize a large number of license plates simultaneously by employing our super-resolution technology with an ultra-low streaming rate. To address the data silo problem, we will develop an integrated data platform based on centralized or federated machine learning techniques for multi-modal, sporadic and noisy samples. We will conduct pilot cases in carparks, and malls to validate and demonstrate our technologies. These technologies will lead to cost-effective IoT deployment, and customer journey from carpark to mall. Social distancing for COVID-19 control and disinfection operations can also be achieved. |
| Project Coordinator: | Prof Gary Chan |
| Approved Funding Amount: | HK$6.59M |
| Project Period: | 31 Mar 2021 - 30 Mar 2023 |
