People at CTMS

» Search People

Silvia Ferrari, Adjunct Professor of Mechanical Engineering and Materials Science and Faculty Network Member of Duke Institute for Brain Sciences

 

Silvia Ferrari

Professor Ferrari's research aims at providing intelligent control systems with a higher degree of mathematical structure to guide their application and improve reliability. Decision-making processes are automated based on concepts drawn from control theory and the life sciences. Recent efforts have focused on the development of reconfigurable controllers implementing neural networks with procedural long-term memories. Full-scale simulations show that these controllers are capable of learning from new and unmodeled aircraft dynamics in real time, improving performance and even preventing loss of control in the event of control failures, nonlinear and near-stall dynamics, and parameter variations. New optimal control problems and methods based on computational geometry are being investigated to improve the effectiveness of integrated surveillance systems by networks of autonomous vehicles, such as, underwater gliders and ground robots.

Contact Info:
Office Location:  144 Hudson Hall, Box 90300, Durham, NC 27708
Office Phone:  (919) 660-5310
Email Address: send me a message
Web Page:  http://fred.egr.duke.edu/

Education:

Ph.D.Princeton University2002
M.A.Princeton University1999
B.S.Embry-Riddle Aeronautical University1997

Specialties:

neural networks
Bayesian networks
Controls
Smart Technology

Research Interests:

Design and analysis of methods and algorithms for learning and computational intelligence. Theory and approximation properties of network models, such as neural and probabilistic networks, for the purpose of enhancing their learning abilities and improving reliability. Approximate dynamic programming and optimal control techniques, with applications in adaptive flight control and mobile sensor networks. Application of expert systems and systems theory to psychological and cognitive modeling from data.

Keywords:

Adaptation, Psychological • Algorithms • Approximate dynamic programming
• Artificial Intelligence • Cluster Analysis • Computer Communication Networks • Computer Simulation • Computing Methodologies • Data Mining • Decision Support Techniques • Feedback • Humans • Hybrid systems • Information Theory • Intelligent systems for criminal profiling • Knowledge • Learning • Models, Statistical • Models, Theoretical • Neural Networks (Computer) • Nonlinear Dynamics • Numerical Analysis, Computer-Assisted • On-line learning in neural and Baysian networks
• Pattern Recognition, Automated • Planning • Poisson Distribution • Programming, Linear • Reconfigurable control of aircraft
• Sensor planning for integrated surveillance systems
• Systems Theory • Transducers • Virtual reality

Current Ph.D. Students  
  • Guoxian Zhang  
  • Gianluca DiMuro  
  • Chenghui Cai  

Representative Publications   (More Publications)

  1. Cai, C; Ferrari, S, Information-driven sensor path planning by approximate cell decomposition., IEEE Transactions on Systems, Man and Cybernetics - Part B: Cybernetics, vol. 39 no. 3 (Submitted, in revision), pp. 672-689 [19193512], [doi]  [abs]
  2. Ferrari, S, Multiobjective algebraic synthesis of neural control systems by implicit model following., IEEE Transactions on Neural Networks, vol. 20 no. 3 (Submitted, in press), pp. 406-419 [19203887], [doi]  [abs]
  3. Ferrari, S; Steck, JE; Chandramohan, R, Adaptive feedback control by constrained approximate dynamic programming., IEEE Transactions on Systems, Man and Cybernetics - Part B: Cybernetics, vol. 38 no. 4 (August 2008), pp. 982-987 [18632388], [doi]  [abs]
  4. Baumgartner, K; Ferrari, S, A geometric transversal approach to analyzing track coverage in sensor networks, IEEE Transactions on Computers, vol. 57 no. 8 (2008), pp. 1113-1128, ISSN 0018-9340 [doi]  [abs]
  5. Baumgartner, K; Ferrari, S; Palermo, G, Constructing Bayesian networks for criminal profiling from limited data, Knowledge-Based Systems, vol. 21 no. 7 (Accepted, in press), pp. 563-572, ISSN 0950-7051 (Available online at: http://dx.doi.org/10.1016/j.knosys.2008.03.019.) [doi]  [abs]
  6. Ferrari, S., and Jensenius M., A Constrained Optimization Approach to Preserving Prior Knowledge During Incremental Training, IEEE Transactions on Neural Networks, vol. 19 no. 6 (June 2008)
  7. Ferrari, S; Vaghi, A, Demining sensor modeling and feature-level fusion by bayesian networks, IEEE Sensors Journal, vol. 6 no. 2 (2006), pp. 471-483, ISSN 1530-437X [JSEN.2006.870162], [doi]  [abs]
  8. Ferrari, S; Stengel, RF, Smooth function approximation using neural networks., IEEE Transactions on Neural Networks, vol. 16 no. 1 (January, 2005), pp. 24-38, ISSN 1045-9227 [15732387], [doi]  [abs]
  9. S. Ferrari and R. F. Stengel, Online adaptive critic flight control, Journal Of Guidance Control And Dynamics, vol. 27 no. 5 (2004), pp. 777 -- 786, ISSN 0731-5090
  10. Ferrari, S; Stengel, RF, Classical/neural synthesis of nonlinear control systems, Journal of Guidance, Control, and Dynamics, vol. 25 no. 3 (2002), pp. 442-448  [abs]