Indoor Exploration Of MAV Using Simultaneous Localization And Mapping

Using MAVs to explore unknown environments instead of human can avoid the threat from the uncertain factors in unknown environment.It will be the focus of future military MAVs.In a GPS denied indoor environment,it is difficult for MAV to locate itself by using onboard sensors only.This paper takes the MAV's exploration in an unknown indoor environment as the research background.,developed a method for MAV's localization in indoor environment and the strategy for unknown environment exploration.The localization method and the exploration strategy are proved to be feasible and effective by flight test.And the research results have important reference value for engineering application.First of all,in order to achieve MAV's indoor localization based on its onboard sensors,we use a scan matching method based on Gauss Newton iteration to get the pose of MAV and the local map model,which lays the foundation for the research of autonomous exploration.Secondly,to reduce the cumulative error produce by scan matching,a method of optimizing the pose of MAV is proposed by using submap as key frame.Once the scan-pose and the submap-pose has a good match,the loop closure will be done and the cumulative error will be reduced.On this basis,to improve the performance of localization in long straight corridor using laser range finder,a multi-sensor fusion location algorithm based on optical flow and IMU is designed.Then,to achieve the purpose of autonomous exploration in unknown indoor environment using MAV,the existing SRT algorithm is further improved by augmenting a frontier based guidance strategy with wall-following logic.By adding a wall-following logic,the defect of laser range finder can not locate because of lack of environmental characteristics is overcome.And the frontier based guidance strategy chose new nodes near the frontier of the unexplored area and the explored area can reduce the randomness of SRT algorithm and improve the searching efficiency.Finally,the experiment of MAV's exploration in unknown environment are carried out by using our designed MAV with onboard SLAM system.The test results show that the indoor location and mapping method presented in this paper can provide accurate localization for MAV,and the improved SRT algorithm can provide effective navigation for MAV in unknown environment.