Published by Proc. IEEE (download the table of contents).
II. COLLABORATIVE WORK WITH INDUSTRY
1. White paper on Cognitive Radar Information Networks for security along the Great Lakes, Collaborator: Accipiter Radar Technology.
1. S. Haykin, Adaptive Filter Theory, 5th Edition, Prentice Hall, 2013.
2. S. Haykin, Digital Communication Systems, John Wiley and Sons, 2013.
3. S. Haykin, Cognitive Dynamic Systems, Cambridge Univerity Press, March 2012. (The first book written on this integrative new field.)
1. S. Haykin and J.M. Fuster, On Cognitive Dynamic Systems: Cognitive Neuroscience and Engineering Learning From Each Other, Proc. IEEE, Special Issue on Cognitive Dynamic Systems, April 2014.
2. A. Amiri and S. Haykin, Improved Sparse Coding under the Influence of Perceptual Attention, Neural Computation Journal, MIT-Press, February 2014.
3. S. Haykin and P. Setoodeh, Cognitive Radio Networks: The Supply Chain Paradigm, second revision.
4. F. Khozeimeh and S. Haykin, Self-organized Dynamic Spectrum Management for Ad Hoc Networks, IEEE Trans. on Wireless Communications, under review.
5. S. Haykin, Y. Xue, and P. Setoodeh, "Cogitive Radar: Step Toward Bridging Gap Between Neuroscience and Engineering," Proc. IEEE, 2012, to be published.
6. S. Haykin, M. Fatemi, Y. Xue, and P. Setoodeh, "Cognitive Control," Proc. IEEE, submitted.
7. M. Fatemi and S. Haykin, "On Reinforcement Learning and Planning in Cognitive Control", IEEE Trans. on Neural Networks and Learning Systems.
8. P. Setoodeh and S. Haykin, “Double-layer dynamics of cognitive radio networks,” The First IEEE International Workshop on Emerging COgnitive Radio Applications and aLgorithms (CORAL), San Francisco, June 2012.
9. P. Setoodeh and S. Haykin, “Dynamic spectrum supply chain model for cognitive radio networks,” The First IEEE International Workshop on Emerging COgnitive Radio Applications and aLgorithms (CORAL), San Francisco, June 2012.
V. INVITED LECTURES
1. S. Haykin, "Cognitive Dynamic Systems: Cognitive Control," 2nd International Symposium on Innovative Mathematical Modeling, University of Tokyo, Japan, May 15, 2012.
2. S. Haykin, "Cognitive Control," RIKEN Institute, Tokyo, Japan, May 21, 2012
The research program on Cognitive Radio Networks was initiated in 2006, focusing on the following projects:
All of these projects have now been completed.
Building on their results, we are emboldened to start a new project, which may be viewed as "Beyond Traditional Cognitive Radio". Specifically, this new project will focus on:
The aim of this new project is to exploit femtocells as an avenue for impacting the global world of wireless networks.
Phase I onf the research project on Cognitive Radar has been completed with the following paper accepted for publication in Proc. IEEE in 2012:
To be more specific, for Phase II on Cognitive Radar, we are focusing on a new hierarchy memory systems, the aim of which is to come that much closer to the visual brain. We anticipate that this Phase II of cognitive radar research will take us beyond Phase I.
Inspired by the results obtained from Phase I of the work done on Cognitive Radar, we have discovered two-state model of the environment:
By exploiting the new notion of entropic state, we have resolved the problem of using Reinforcement Learning as the tool for action in the environment. In so doing, preliminary results of Cognitive Control have demonstrated that computational complexitiy of the controller can be improved significantly, compared to the use of Bellman's dynamic programming, albeit in a suboptimal manner.
here again, our future work on cognitive control will be aimed at coming closer to how cognitive control is performed in the human brain.
The cocktail party problem refers to a crowded environment where there are many conversations being carried out simultaneously, and the requirement is to separate out a particular voice signal of interest in an intelligible and reliable manner. To solve this problem, we have adopted a neurobiologically motivated approach, exploiting cognition. The objective here is to build on lessons learned from cognitive radar and cognitive control.
We have developed a new nonlinear sequential state estimator, which we have named the cubature Kalman ﬁlter; the term "cubature"is taken from the “cubature method” for numerical integration. This new nonlinear ﬁlter, rooted in numerical methods, is not only mathematically elegant, but also has two important attributes:
We are currently investigating application of the cubature Kalman ﬁlter in solving challenging pattern-recognition problems with particular emphasis on large-scale applications.
A sample MATLAB code that implements the square-root cubature Kalman filter can be found here.