Self-organizing algorithm for logic circuit based on local rules

This study discusses a learning algorithm for autonomous robots that has five characteristics including autonomous exploration of effective output, low calculation costs, capability for multi-tasking, reusing past knowledge, and handling time series. We propose the use of self-organizing network elements (SONE) as a method for creating learning systems that provide these characteristics. Using this method, we created and evaluated a Self-Organizing Logic Circuit. The results of our experiments showed that this learning system met the requirements by being capable of creating a basic logic circuit, learning additional knowledge, controlling a simple robot in a simulation, and solving a maze problem.