Research & Development
Promoting the advances of Analytics Everwhere Framework in performing automated analytical tasks capable of providing higher-level intelligence from continuous IoT/CPS data streams and generating long-term insights from accumulated IoT/CPS data streams. Exploring the possibility to adapt our framework in building prototype to support multiple Social Good use cases.
Developing and evaluating scalable and adaptive Data Analytics Platforms for the IoT/CPS. Proposing unique solution to capture, manage, process, analyze, and visualize data streams through streaming descriptive, diagnostic, and predictive analytics. Incooprating human-in-the-loop, facilitating the interaction between human and future IoT/CPS platform to deploy in different Smart City scenarios.
IoT/CPS for combatting Climate Change
Empowering current efforts to combat climate change by leveraging the most advanced AI/ML techniques to analyze the large scale of these spatial-temporal data. Solving technical challenges in connectivity, communications and protocols of IoT/CPS platforms and applying new technologies in software, analytics, architecture configurations and platform infrastructure to address Climate Change issues.
Industry 4.0Investigating and analyzing of Industrial IoT/CPS Ecosystems. Explaining the unknown and visualizing the decision making process by post-hoc analysis and build-in realtime detection mechanisms in our AE Framework. Understanding the tight relationship between the complexity of the algorithms and the data life-cycle to support the evolution of Industry 4.0.
Our Analytics Everywhere framework currently has 3 main pillar:
- Analytical Capability: The algorithms will be combined to perform descriptive, diagnostic and predictive analytics on data streams being transported through the distributed resource architecture.
- Resource Capability: This distributed resource architecture is needed to handle the continuous and accumulated computation of incoming data tuples.
- Data Life-Cycle: Data streams are usually an unbounded sequence of tuples generated with high data rates. So, depending on the types of analytical tasks and compute nodes needed by an CPS/IoT application, the data life-cycles in our framework can manage the changes that data streams go through during the automated execution of a network of analytical tasks.
An example of a 3-layers architecture to handle real-time data streams.