Applying GA as Autonomous Landing Methodology to a Computer-Simulated UAV

The ultimate purpose of this study is to truly make unmanned aerial vehicles to be autonomous. As the starting point, in this paper, we have selected genetic algorithm as the method to achieve autonomy and will check the possibility of self-regulated autonomous unmanned aerial vehicle by applying the genetic algorithm. Landing has always been one of the most important tasks for aerial vehicles. Particularly, self-regulated autonomous landing is essential when it comes to unmanned aerial navigation. Since researching the autonomy of landing through falling body installed with a lunar Lander-like propulsion system would be more efficient for attaining the generalization of autonomous landing, it is applied on the simulation of computer-simulated falling body. When applying genetic algorithm, by first encoding genome into only 4 types of actions(left turn, right turn, thrust, free fall) then applying it on unmanned falling body, and finally combining the major computations of genetic algorithm to the unmanned falling body, we have made a successful progress in experiments. Meanwhile, previous studies have relied on various sensors to correct vertical, rotational, and horizontal errors. However, the use of measurement sensors has to be minimized in order to achieve true autonomy. The greatest accomplishment in this study was implemented a computer-simulated unmanned aerial vehicle, in order to minimize reliance of sensors, which can achieve the true meaning of autonomy of unmanned aerial vehicle and establishing a test bed for verifying the possibility by using genetic algorithm.