Large-scale deployment of 3S visualization will require extensive sensing capabilities. For this purpose, we will leverage the island-wide sensor network infrastructure and data warehousing capability envisioned in A*STAR’s Urban Systems Initiative.
In addition, we will carry out research that is beyond the near-term scope of the Urban Systems Initiative, focusing on scalable real-time analytics, especially for difficult data types.
Some types of sensor data are very difficult to analyze automatically. Most video data falls into this category, as does much audio, including human conversations. Analysis of text data is time-consuming and error-prone, though it is much more advanced than audio and video. It can even be a challenge to analyze simple numerical sensor data when readings are frequent. Add in the need for real-time analytics and the ability to adjust to dynamic needs, and we are immediately beyond the capabilities of today's technology. Yet the ability to understand what people are doing and saying in the moment is key for urban safety and security.
ADSC researchers are working to develop new methods that enable real-time analytics of selected kinds of audio, video, text, and conventional data.
Data Analytics Projects:
- Data-Driven Modelling and Real Time Optimization for Diversified Chiller Plants, Dr. Zhenjie Zhang
- Next-Generation Compilers and Architectures for Computation Acceleration with Energy Efficiency, Professor Deming Chen
- Scalable, Real-time Analytics for Challenging Data, Professor Marianne Winslett