This focus area explores new methods and platforms to perform multi-modality sensing based on mobile swarm agents. We study ways to improve sensing performance by letting the swarm find areas of interest using emergent behavior and autonomous navigation.

We have developed a novel method for mobile agent navigation based on potential fields generated using radio beacons, without the use of GPS or other localization methods. This simple, yet robust, navigation scheme has been successfully implemented on a mobile robot platform using wireless nodes equipped with IEEE 802.15.4 / Zigbee based radios. The mobile agents are also equipped with multiple sensors to detect various signal modalities, towards obtaining more complete information on the scenario under observation. Processing such information concurrently over multiple sensor modalities is studied under the focus area 2 described below.

Our research interests in this area include swarm robustness, contaminant detection and tracking algorithms, spatio-temporal sampling, and formation forming using mobile agents. This work ranges from theoretical analysis and computer simulations to actual implementations on real-world mobile sensor platforms.

We study wireless sensor and actuator networks with a special focus on distributed sensing, information processing, and control. An important problem is the handling of heterogeneous signal sources and sensor data in a notoriously unreliable and uncertain wireless sensing environments.

Areas of particular interest are the Dempster-Schafer theory of evidence and its generalizations, spatio-temporally selective fusion, and energy efficient transport and distributed processing of sensor information. A typical application of this work is the "needle in a hay stack" problem in real-time data mining in sensor networks. In this direction, we developed a novel method named "Evidence Filtering" that can directly process imperfect sensor data over multiple signal modalities.

Also we have developed a simple and modular state-space model for distributed sensing and actuation in grid sensor networks where nodes are placed in a regular spatial grid. In this direction, we study local multidimensional (m-D) state-space models for distributed operations.

During synchronization of networked dynamical systems, the problem of drifting clocks and non-identical sampling times is known to cause problems regarding stability and performance. We concentrate on accurate modeling of these systems as well as analyzing their behavior. A particular focus is the problem of robust stability with respect to sampling rates. Methods for robust design and fault detection in such systems in the presence of synchronization errors are also a key concern.

Network congestion control using control theoretic methods are also been investigated under this area. Most of the work has concentrated on explicit rate feedback in ATM networks (ABR option). A particular focus of this work is the accurate modeling of the communication link using dynamical system models.

This focus area presents a framework for the combination of evidence in an environment where data are generated from heterogeneous sources possessing partial or incomplete knowledge about the global network scenario. The approach taken is based on the conditional belief and plausibility notions in Dempster–Shafer evidence theory that allow one to condition these partial knowledge bases so that only that portion of the incoming evidence that is relevant is utilized for updating an existing knowledge base. The strategy proposed enables one to accommodate some of the most challenging, yet essential, features that are encountered when evidence is generated from possibly a large numbers of sources. These include heterogeneity and reliability of incoming evidence, inertia and integrity of evidence already gathered, and potentially limited resources at the nodes where evidence updating is carried out. The proposed framework is applied in robot map discovery using ultrasonic sensors and a real-world scenario where sensor data generated by heterogeneous sensors are used for potential threat carrier-type detection.