The results of my experiment supported my hypothesis. I found that if the robot's AI pathfinding algorithm is modified to incorporate aspects of its surrounding environment to correct its course, then the expected variance of the ending position of the robot from its target destination can be reduced by 36.5%. Robot navigation is an important topic, since improvements in artificial intelligence algorithms will enable robots to more accurately navigate autonomously and operate in a hospital environment filled with people, food tray carts, and gurneys left in the hallway. Having "smart", fully autonomous robots (rather than robots that can navigate only on pre-determined routes) would allow hospital staff to confidently send these reliable robots on errands into patient rooms and throughout the entire hospital.

My experiment could be improved by determining the turning variance factor using testing data from the type of robot that would likely be placed in the hospital setting. Robots are often engineered for specific purposes, and some navigate and negotiate turns more accurately than others. Thus, engineering a prototype robot for use in a hospital would not only make an interesting project, but allow me to do a more representative turning variance analysis. 

Working on this project has caused me to wonder about the number of decisions that a fully autonomous robot would have to make, should it have access to the entire hospital. The number of determinations that the algorithm would have to address would increase substantially. It would have to address not only hospital navigation, but obstacle avoidance, working with the flow of foot traffic, and reacting to sudden/unexpected blockages. I hope to see more AI applications used in the medical field in the future, as I see artificial intelligence software development and health care fitting together well. 

This science fair project has drawn me to work next with scalable artificial intelligence, developing an algorithm that adjusts to a larger and more complex environment. I plan to apply my AI development work to the VA hospital, again. I have started my research into the project, and I have completed some AI development work that I am testing in a Java-based robotic simulation environment. My initial testing in the simulator has gone well, so far. 

I enjoy my other programming classes, but find myself continually drawn to my on-going artificial intelligence projects. Artificial intelligence software development is very challenging, and the challenge draws me continuously to the development work. I observe AI used frequently in the world economy today, and I envision an explosion in future AI applications. AI development is my passion, and I am working to see my research work used in a real application that will help other people in the future. 


How did your experiment support or contradict your original hypothesis? How could you improve your experiment? Did everything go as planned or were there unexpected results? Does what you learned lead to a new question to ask or experiment that would answer it? If so, why would it be important, interesting, or useful to do?


Judges' Tip
An excellent conclusion will explain how the experiment answers the question or why it fails to do so and whether or not it supports the hypothesis (500 words maximum).