Critical Node Analysis and Mitigation in Industrial Wireless Sensor Networks (Applications of Spectral Graph theory)

Friday, December 9, 2016 - 1:25pm - 2:25pm
Lind 305
Soumitri Kolavennu (Honeywell)
Honeywell’s automation and control solutions unit consists of a number businesses where wireless sensor networks are utilized. These wireless networks provide total install cost savings for the business and its customers. In addition, during operation the systems are more robust to link and node failures. Wireless sensor networks are employed in industrial control and data acquisition systems, hotel management networks, automated meter reading networks, building scale fire and life safety systems etc.

The wireless sensor networks are deployed as mesh networks where the sensor and actuator nodes can talk to each other neighbors and reach destination nodes which are typically gateways where the data is consumed. Mesh networks are desirable because of the inherent ability of the network to reconfigure itself around a faulted node or a link and still be able to reach the gateway. However such guarantees are only possible when there are no critical nodes (nodes that split the network into disjoint clusters). This talk concentrates on ability of the network management functions to detect critical nodes quickly and provide mitigation for the critical nodes and links using spectral graph theory. Other applications of spectral graph theory for these networks are also discussed. The main motivation of the presentation is to present the typical kinds of problems in practical implementation of sensor networks in the automation and control industries.

Soumitri Kolavennu joined Honeywell in 1999, initially as a student intern. His areas of expertise include wireless, voice, and advanced control.  His early work in wireless mesh networking and wireless localization enabled ACS products like OneWireless™ and Impact Xtreme to be differentiated in the marketplace. Recently, He has been a key contributor to Honeywell’s Connected Home offerings, specifically the Voice recognition and Home Kit based thermostats. He was the primary contributor and editor for the Networking and Provisioning Layers of the ISA100.11a standard for industrial wireless sensor networks and he has helped draft standards relevant to the smart grid and energy management systems. He is a member of the Standards Subcommittee of IEEE Control Systems Society and a past member of the Smart Grid Architecture Committee (SGAC) for the NIST Smart Grid Interoperability Panel (SGIP). He is the fellows leader for Homes and Building technologies with in Honeywell. He has over 50 issued patents. His publications include a recent edited volume, “Industrial Wireless Sensor Networks: Monitoring, Control and Automation,”among others. Soumitri holds a B.Tech. degree from Andhra University, India, and a Ph.D. from Florida State University, both in chemical engineering.