Research
The Center for Intelligent Networked Systems (CINS) promotes collaborative research in the following areas.
-
Advanced Sensors
The ability to sense the environment is the first phase of the CINS approach. Research on advanced sensors (e.g., millimeter-wave, infrared, seismic, UAV-attached, biochemical, acoustic, nuclear) will lead to richer, heterogeneous, multi-dimensional sensor data.
-
Wireless sensor networks
To ease their deployment, the sensors will need to be wireless and mobile. Research on wireless mobile and ad-hoc sensors networks is developing robust, secure and self-localizing wireless sensor networks.
-
Data fusion
The data obtained from a heterogeneous sensor network must be fused in order to correlate data from different sensors, reduce the scale of the data, and provide higher-level information for learning and decision making. Research on data fusion is developing methods to extract maximum information capable of optimizing decision-making performance for the target environment.
-
Data integration
Other sources of data (e.g., intelligence databases) are equally important in providing external information to improve decision-making performance. Research in databases and data integration is developing methods for dealing with the massive, almost unlimited, amount of data that could be relevant to the task.
-
Data mining
The ability to mine the information resulting from the fusion and integration phase is critical in order to detect patterns in the environment to support better understanding of the domain, to describe normative behavior against which anomalies can be identified, and to assist in prediction and decision making. Data mining research in the CINS is investigating methods for detecting patterns and anomalies in massive networks of nodes and links representing the information from sensors and external data sources.
-
Artificial intelligence
In addition to integrating intelligence into the entire CINS process, AI plays a particular role in using the information, patterns and anomalies to decide which actions to take in the environment. AI research in the areas of knowledge representation, learning, reasoning, prediction and decision making are crucial to the utility, adaptability and effectiveness of the CINS approach in any domain.
-
Modeling and simulation
Deploying and testing a system in the target environment is the best method to measure the performance of the system. However, in many environments of interest (e.g., homeland security and defense) deployment and even acquisition of test data, is not possible due issues of security, privacy and disruption of operations. For this reason a significant component of the CINS approach is the modeling and simulation of the environment, including the sensors, sensor data, external data and the environment itself. Research in modeling and simulation is developing models of the target environments and the relevant entities and data in the environment in order to provide a high-fidelity simulation of the target environment that can be used to accurately measure and improve the performance of a system prior to deployment.
-
Software engineering
The CINS approach requires the integration of multiple heterogeneous dynamic components into a coherent, real-time system that can take timely actions based on its perception and understanding of the environment. Research in software engineering of agent-based systems is a viable approach to facilitate the development of such a system and analyze its performance.
-
High-performance computing
The need to process massive amounts of real-time data requires an unprecedented level of computing power. Research in high-performance computing, especially in the new breed of massive shared memory, massively hyper-threaded architectures, is critical to achieving the necessary computing power.
-
Security and privacy
Whenever an environment is being sensed and automated actions taken, issues of security and privacy must be considered. The sensed and fused data, the external data, the integrated information, the learned patterns and anomalies all raise privacy concerns for the elements being observed in the target environment. Safety must also be maintained in environments in which automated actions are possible. Security of the data, systems and environment is also important for the integrity of and trust in such systems. Research in security and privacy of systems and data are developing methods to achieve maximum effectiveness while minimizing security and privacy risks.
-
Bioinformatics
Biological systems (e.g., humans) are one of the CINS target environments in order to help fulfil the CINS mission of making our world healthier. Biological systems are not unlike other environments in that sensors can be used to perceive the state of biological systems and automated actions can be taken to diagnose and treat specific conditions before they become a threat. Research in bioinformatics is developing models of such systems, methods for sensing them, and algorithms for predicting their behavior. The results of this research can be used to develop CINS processes to revolutionize our approach to health.