NPC

Ezer Lichtenstein

NPC is the first steps in developing an open and affordable platform for behavioral robotics.



What if robots developed their own opinions? Or felt a need to be a part of something, even if they didn’t always agree? What if robots developed basic emotions: becoming jealous; depressed; or even started developing a sense of identity?

NPC is the first steps in developing models of behavior through robotics. In this study iterative robots were developed to perform a variety to tasks. In this test, the robots are asked to work together to understand their environment. They are tasked with finding the best method of controlling an outside robot beyond their community. As they move this robot, they gain feedback and strengthen or weaken their original hypothesis of the outside world. This robotics system was designed and built as an accesible platform for other researchers, engineers, and hobbyist to create their own experiments in behavioral robotics.

Background
- Sanchez-Sinencio, Lau: Artificial Neural Networks


- Zalzala, Morris: Neural Networks for Robotic Control


- Holland: Basic Robotic Concepts


- Caudill, Butler: Understanding Neural Networks: Computer Explorations


- Blum: Neural Networks in C++


- Minsky: The Society of Mind


- Arkin: Behavior Based Robotics


- Braitenberg: Vehicles, Experiments in Synthetic Psychology


- Restack: The Brain


- Cook: Intermediate Robot Building


- Jones, Flynn: Mobile Robots


- McFarland, Bosser: Intelligent Behavior in Animals and Robots


- Hansell: Animal Architecture & Building Behavior


- Siegwart, Nourbakhsh, Scaramuzza: Introduction to Autonomous Mobile Robots


- Hawkins: On Intelligence


- Sptizer: Mind Within the Net


- Haikonen: Robot Brains


- Webb, Consi: Biorobotics


- Crick: The Astonishing Hypothesis





Papers


- http://creativemachines.cornell.edu/papers/RAM07_Yim.pdf


- http://lasa.epfl.ch/publications/uploadedFiles/ICRA2009_sproewit_preprint.pdf


- http://www.intellization.com/files/NN_thruster_control_paper_workshop_93.pdf


- http://digitalcommons.calpoly.edu/cgi/viewcontent.cgi?article=1081&context=csse_fac


- http://www.roboticsproceedings.org/rss03/p07.pdf


- http://nn.cs.utexas.edu/downloads/papers/stanley.jair04.pdf


- http://robotics.usc.edu/~maja/publications/sab90-2.pdf


- http://sbp.ri.cmu.edu/papers/sbp_papers/integrated1/mcklurkin_ants_eod-paper.pdf

Audience
Anyone interested in experimenting with intelligence or robotics.

Implementation
I am making 6 nearly identical robots that will have the ability to communicate with each other. I will be running a series of experiments with the robots to showcase their abilities. One experiment will be programming the robots to act as one of Braitenberg's vehicles. As the robots perform Braitenberg's experments I will document and observe if the his larger ideas of antropomorphization with the robots actually occurs. Another experiment will try to emulate a very simple neural network with the robots. There is an input layer which will be the robots coordinating with each other through sensors and XBees. There is the hidden layer which will be the computer filtering and weighing the inputs from the robots and outputting that to another separate robot. And the output layer being the separate robot that will get it's instruction via the 6 robots by way of the computer.