Welcome to my page!


I am a student at Oklahoma Christian University majoring in Computer Science and Finance. I am a senior graduating in May 2015. For more information about me, please visit my website here.

When I am not studying, I enjoy reading, improving my singing skills, and exploring different cuisines around the world. I also work as a lab assistant for programming I and II, a job that I have found to be double rewarding; I get a chance to inspire new students while I keep my memory fresh on the basics of programming.

If I make it to graduate school, I will be one of the less than 200 Ph.Ds in my country. My dream is to use my skills to help, teach and inspire as many people as I can.
I was priviledged to be accepted in the DREU program. For 10 weeks I am doing research with Doctor Nancy Amato on motion planning and protein folding.



Probabilistic Roadmap Methods (PRMs) have proved to be very efficient in solving problems for robots with many degrees of freedom (DoFs). PRMs use randomization to build a graph of possible paths in the free configurations and connect vertices using a local planner. It is, without any doubt, impossible to hand-pick the best method for each region of a heterogeneous environment. The Adaptive Neighbor Connection (ANC) method saves the user from the tedious task of hand-selecting the best strategy and updates the selection over time. The ANC, however, can be improved by using the concept of spatial learning. In this research, we present an algorithm that allows us to use the best sampling strategy for each small region of the environment without having to subdivide the environment. Our framework will allow the success to failure ratios of connection in the immediate vicinity to have a bigger influence on the probabilities that are used to select the specific strategy to be used to add the nodes of that section to the roadmap. We perform experiments on rigid bodies of several degrees of freedoms (more than 35) in an attempt to prove that our strategy performance is indeed region-independent. By proving that our framework selects the best strategy without having to subdivide the environment, we show that it is more efficient than the ANC which it attempts to improve.


The ANC can be improved in many ways and one of them is to test it on smaller segments of the environment. This ANC_spatial algorithm would optimize the learning phase in difficult regions of the C-space.
My objective for this project is to run and evaluate experiments on the ANC-spatial and determine the best environment for comprehensive tests. At the end of this project, I should be able to prove and defend that the ANC_spatial can be more or as efficient as its previous sampling strategies in the motion planning science.

To carry this project out, I will be under mentorship of Dr. Nancy Amato, and will recieve assistance from Chinwe Ekenna, a Ph.D student in Computer Science working on Protein Folding and Motions in the Computational Biology area.

For more information, please read Adaptive Neighbor Connection for PRMs: A Natural Fit for Heterogeneous Environments and Parallelism. authored by Chinwe Ekenna.


The Best Advice for a Graduate Candidate (Given by current graduate students at the Summer Scholars Luncheon)

How do you go about to choose your advisor?

How is a day of a graduate student like?
There is no common answer to this but you have to:

Best Advice if you are considering graduate school: