Bhaskar Saha – ICSL

BhaskarSaha.JPG Office:
Room 227
Intelligent Systems Control laboratory (ICSL)
Manufacturing Research Center (MaRC)
Georgia Institute of Technology (GaTech)
813 Ferst Drive, N.W.
Atlanta, Georgia 30332-0560 U.S.A.

Phone:
+1(404) 276-7364 (cell)
+1(650) 604-4379 (office)

Email:
bsaha@ece.gatech.edu
bhaskar.saha@gmail.com

About

I am a PhD student in the School of Electrical and Computer Engineering at GeorgiaTech. I am working on applying Model-Based Reasoning techniques for fault diagnosis of complex electro-mechanical systems. I joined Georgia Tech in Fall 2002 as a direct PhD student after receiving a Bachelor’s degree from the Indian Institute of Technology, Kharagpur in Instrumentation Engineering in 2001.

Current Work

• Algorithm development for advanced prognostics (Internship: NASA Ames Research Center)
Applied state-of-the-art Bayesian techniques of classification, regression and state estimation to the critical problem of estimating remaining useful life of systems. Developed an integrated Relevance Vector Machine (RVM) and Particle Filter (PF) framework to estimate present damage and damage growth rates as a function of operational parameters to determine time-to-failure probability densities. Also, helped set up an experimental testbed to validate various prognostics algorithms on charge, health and life estimates of small capacity Li-ion cells.

Research

• Application of Model-Based Reasoning to System-level Fault Diagnosis
Formulated a model-based reasoning architecture for a two fold purpose: to detect and classify component faults from observable system behavior, and to generate fault propagation models so as to make a more accurate estimation of current operational risks. The validity of this method has been proven in helicopter gearbox diagnostics. Successful detection of gear pinion crack was achieved from experimental test data from an H-60 intermediate gearbox.

• Functional modeling and Verification of a Monopropellant Propulsion System
Developed a system-level functional model for a monopropellant propulsion system presented in the NASA Fault Tree Handbook with Aerospace Applications, 2002. This model facilitates validation and verification of the MATLAB Simulink representation of the system along with its fault tolerant controller under nominal as well as faulty conditions. For verification purpose the system has been modeled in temporal logic (TLA+) so as to derive system behavioral requirements from its specifications.

• Unmanned Aerial Vehicle (UAV) diagnostics using neural networks
Developed a neural network based fault classifier that uses state variable measurements to detect loss of collective pitch control authority in the main rotor system of the GTMAX, an UAV based on the Yamaha RMax platform. This fault detection and identification routine enables a reconfigurable fault tolerant controller to stabilize the system and prevent catastrophic failure.

Publications

  • Bhaskar Saha, Kai Goebel, Scott Poll and Jon Christophersen, “An Integrated Approach to Battery Health Monitoring using Bayesian Regression and State Estimation”, accepted for publication in IEEE Autotestcon 2007, Baltimore, MD, September 2007.
  • Bhaskar Saha and George Vachtsevanos, “A Novel Model-Based Reasoning Approach to System-Level Diagnostics of a Helicopter Intermediate Gearbox”, Proc. of the 60th Meeting of the Society for Machinery Failure Prevention Technology, pp. 281–290 , April 2006.
  • Bhaskar Saha and George Vachtsevanos, “A Model-Based Reasoning Approach to System Fault Diagnosis”, 10th WSEAS Intl. Conf. On SYSTEMS, July 2006 & WSEAS Trans. on Systems, Issue 8, Volume 5, pp. 1997–2004, August 2006.
  • Bhaskar Saha and George Vachtsevanos, “A Model Abstraction and Representation Approach to Helicopter Powertrain Fault Diagnosis”, accepted for poster presentation at Integrated Systems Health Management Conf., Cincinnati, Aug 2006.
  • Kai Goebel and Bhaskar Saha, “Estimating Remaining Useful Life using Data-driven Techniques”, accepted for poster presentation at Data Mining in Aeronautics, Science, and Exploration Systems 2007 Conference, Mountain View, CA, June 2007.
  • George Vachtsevanos, Bin Zhang, Marcos Orchard, Bhaskar Saha, Abhinav Saxena and Young Jin Lee, “Verification and Validation of Fault Tolerant Systems: A Propulsion System Case Study”, submitted to Reliability Engineering and System Safety.
  • Graham Drozeski, Bhaskar Saha and George Vachtsevanos, “A Fault Detection and Reconfigurable Control Architecture for Unmanned Aerial Vehicles”, 2005 IEEE Aerospace Conf. Proc., pp. 2955–2963, March 2005.
  • Graham Drozeski, Bhaskar Saha and George Vachtsevanos, “A Fault-Tolerant Architecture for Unmanned Rotorcraft”, Proc. of the AHS International Specialists’ Meeting on Unmanned Rotorcraft, January 2005.

Links

Sankalita Saha (my wife’s webpage)

—Bhaskar 18:00, 25 October 2006 (EDT)