* Estimating (number) of real objects vs images or videos of same objects
* Tarantula locomotion
* Time series prediction using neural networks optimized by breeding algorithms
* Minimized combinatorial logic gates for large truth tables when XOR gates allowed.
Professor Daniels' research interests include:
Time series prediction using neural networks optimized by evolutionary algorithms.
Professor Daniels works with inputs generated by the chaotic Rossler equations, and with financial data. The goal is to produce a neural network that succeeds with closed loop prediction.
Minimal logic circuit realization of large truth tables using evolutionary algorithms.
The performance of new algorithms is compared to the definitive Quine-McCluskey method, and the heuristic Espresso method, for both speed and gate count. Since both QM and Espresso end up with AND-to-OR two level logic forms, this work has attempted to create smaller realizations by allowing XOR gates in 3-level logic.
Professor Daniels has an interest in the registration of three-dimensional images from nuclear cardiology and angiography, which he has pursued in collaboration with the Fraunhofer Institute.
Other interests include: time series prediction; truth table minimization; eye movements; optimization with evolutionary algorithms; computer methods for automatic jigsaw puzzle solving; models for color vision which explain various aspects of color blindness.