Discovery Park B235
Dr. Parthasarathy (Partha) Guturu has Bachelor's Degree with Honors In Electronics & Electrical Communication Engineering, Postgraduate Diploma in Computer Technology and Ph.D. (Engineering) in Pattern Recognition- all from Indian Institute of Technology, Kharagpur. He has more than ten years of teaching and research experience in academia and over 7 years of research experience in design and development of complex real-time systems in industry. While in academia, he supervised 4 doctoral and a number of graduate and undergraduate dissertations. He published over 35 papers in international journals and conferences and contributed to disparate areas of Electrical and Computer Engineering including Pattern Recognition, Computer Vision/Image Processing, Artificial Intelligence, and solultions of ill-posed and combinatorial optimization problems with Neural Networks/Genetic Algorithms. In corporate research, he contributed to the areas of Intelligent Networks and 3G Wireless Systems and came up with 3 patents. The patents are in the areas of fault-tolerant message replication & database synchronization and call fail-safe wireless traffic distribution across multiple frequency bands. A paper based on the latter work has been published in IEEE GlobeCom 2003 and another one is ready for publication in IEEE Transactions on Wireless Communications. Dr. Guturu plans to integrate his past experience in Computational Intelligence and Machine Learning and the latest experience in Wireless Networks and apply to research on Wireless Sensor Networks and Systems.
My research focus over the past 15 years has been on the design of complex software systems, and the development of innovative computer algorithms for solving complex problems in the diverse fields of Computer and Communication Engineering. I contributed to both academic and corporate research. My research contributions span over the areas of Advanced Intelligent Networks, Wireless 3G Networks, Pattern Recognition, Computer Vision and Image Processing, Knowledge-based systems, Neural Networks and Genetic Algorithms. While continuing to build upon my current strengths in 3G Wireless systems, I plan to expand the scope of my research into new areas like Data Mining and Bio-informatics taking advantage of my earlier background in Pattern Analysis and Machine Intelligence. Should the planned growth our new EE department dictate the need for exploring new areas such as MEMS and BIOMEMS, I am ready to take the challenge. Because of my success in diverse areas and quick changing industrial environments, I am confident that I will be a success in the new discipline too.
As a faculty member of Computer Science & Engineering and Electronics & Electrical Communication Engineering departments at the Indian Institute of Technology, Kharagpur, I have supervised four doctoral dissertations, published extensively and coordinated many sponsored research projects in the above-mentioned areas of Computational Intelligence. My research work in these areas includes a knowledge-based system for automatic data classification, contributions to the theory and algorithms for occluded object recognition based upon diagnostic reasoning, recursive neural network theory and its application to computer vision, computer algorithms for fast recognition of patterns with a large number of attributes and new classification methods that make an effective use of limited training samples. All this research been published in international conferences and peer-reviewed international journals: such as the IEEE Transactions on Systems, Man and Cybernetics, IEE Proceedings, Pattern Recognition, Information Sciences and Pattern Recognition Letters.
My most recent R&D work at Nortel Networks, Richardson, Texas is in the area of
Software Engineering. My co-workers and I developed methods for isolation of software
objects for automated unit testing and patented the work. Prior to that, I contributed
to the Nortel’s main stream work in 3G-Wireless. I primed the BSC (Base Station Controller)
Data Call Processing software development in the 1XRTT project, which is now customer-ready.
Continuing later with the Multi-carrier Traffic Allocation feature, I primed the BSC
design and came up with a novel algorithm for traffic distribution across 800 and
1900 MHz bands. This work has been filed for a patent. A paper based on this work
has been presented at IEEE GlobeCom 2003 and an advanced version of the same has been
accepted for publication in IEEE transactions on Wireless Communications.
My earlier work at Nortel Networks, Ottawa, Canada, involves a distributed system design for an Advanced Intelligent Network (AIN) application. More specifically, the work is in the area of message replication and database synchronization. This work resulted in two US patents.
More details about these three patents are available in patents section.
I shall obtain funding from the Government and Industry by projecting my achievements in solving practical problems for the industry and continue my research work in the Wireless Network and AIN areas. In addition, I plan to expand the scope of my research to novel application areas- some related to my past background and some that fall in the overlapping region of my past and present expertise. These areas include:
I expect that, in view of the importance these areas are gaining now, funding for research in these areas should not be a problem.
These two patents are offshoots of an R&D project for design of a distributed system for Advanced Intelligent Network (AIN) applications at Nortel Networks Corporation, Ottawa, Canada. High availability requirements for AIN services (e.g., calling card services) mandate that redundant databases containing the service information (such as a customer's current balance) are maintained at numerous Service Control Points (SCPs) spread over multiple LANs and WANs and are synchronized to reflect new updates. New data update from SS7 network (the so-called network-side update or simply NSU) can be handed over to any SCP, but must be propagated to all the SCPs in the consortium, so that each database is consistently updated. This problem is called message replication, because each message is replicated and propagated to all computers in a system. To address the problem, I designed and implemented a novel high performance message asynchronous replication subsystem taking a paradigm shift from those implementing traditional 2-phase and 3-phase commit type of algorithms which are synchronous by nature and hence are slow. I also enhanced it to be fault tolerant by using for each primary source of replication one or more surrogates to take up the responsibility for replication in the event of failure of the primary. Above two inventions are by-products of this design and are related to message replication and database synchronization aspects of the problem. The first one is a fault-tolerant message routing method that minimizes the network WAN traffic during replication. The second one is a new database conflict resolution method that facilitates data base synchronization despite conflicting updates originating at different sources. The second method differs from traditional conflict resolution methods in that it addresses a number AIN peculiarities such as the distributed mode of operation of SCPs that can be overridden by one or more Service Management Systems (SMSs) acting as masters.