Study On Control Technique Of Airship Random Fly

Indoor airship, one kind of unmanned autonomous aircrafts, is a aircraft that is able tohover in the air with very little power. It has the advantages of environmental protection,energy saving and the long time of hovering. It can also attract people’s attention by using theupper space of the room. The goal of this paper is to realize the autonomous stroll without thesupport of positioning system of indoor airship, and to shelter obstacle autonomously andhave ergodic ability of some upper space of the room.According to different traversal strategy of unknown environment, the paper will classifythe strategy into three categories. The first category is completely simple randomizedstrategies; the second category is based on the strategies that at the same time to build theenvironment map and positioning, which is called Simultaneous Localization and Mapping,short for SLAM, instead of building environment map, the priority of the algorithm is todetecting, which is called the detection of the Next-Bext-View path planning, the main idea isto find the next best candidate viewpoint; First of all，the third type of strategy called "Anglelearning" completes global framework，whose purpose is building the interior space. Then,it is used to complete the environmental traversal combined with the global information andlocal path.Method of fuzzy artificial is improved with the characteristics of higher inertia, fuzzyrandom, low speed, difficulty of positioning and scanning the data of current condition ofindoor airship. The related gravity and repulsion potential field are designed, and the maininfluence of potential field are fuzzy distance of obstacle and forward space detection. Theautonomous stroll and sheltering obstacle can be realized with the algorithm considering thefuzziness of airship state. In consideration of the actual flight of airship can be simplified tothe flight in specific space, and high complicacy and low efficiency of path planning in3Dspace, the simulation using Matlab in2D space is conducted, and feasibility of algorithm isverified.