Genetic algorithms for self-spreading autonomous and holonomic unmanned vehicles in a three-dimensional space

We present a genetic algorithm, called 3D-GA, as a decentralized spatial control mechanism autonomously running in holonomic unmanned vehicles (HUVs) to achieve a uniform distribution in a three dimensional space. In aerial and underwater military theatres this task is difficult due to dynamic, harsh and bandwidth limited conditions, and lack of a centralized controller. Using only near neighbor information, our 3D-GA guides each HUV to select a velocity vector with a higher fitness among exponentially large number of choices, converging towards a uniform spatial distribution. We demonstrate that the HUVs running our 3D-GA create a highly resilient network that can adapt to changing conditions such as the addition or loss of HUVs due to replenishment, malfunction or destruction. If HUVs are added or removed, the rest of the HUVs will reposition themselves using our 3D-GA to maximize their volumetric coverage of an aerial or underwater space. Our simulation software results verify that our 3D-GA can be an effective tool for providing a robust solution for volumetric spatial control of HUVs in military applications.