Research On Technology For Indoor Flight Localization And Obstacle Avoidance Based On Laser Range Scanning

On the application background of using micro quadrotor to explore the unknown indoor environment, the indoor flight localization and obstacle avoidance technologies based on laser ranging scanning have been studied in this thesis. As to the problem of calculation complexity of the Simultaneous Localization And Mapping(SLAM) algorithm, the research has the goal that it can be implemented on ARM-based embedded flight control system for micro quadrotor.The Iterative Closest Point(ICP) matching algorithm is used to realize indoor localization. To reduce the calculation complexity during the matching process, the standard ICP matching algorithm is improved. First, the real feature points in environment but not the scanning points of laser ranging scanner is used to accomplish the matching process. Second, the searching range of the corresponding point is limited. In order to get real feature points in environment, scanning datum are processed, such as data filtering, region segmentation, feature line sections and feature points extraction. And a screening mechanism of real feature points is given. Considering the variation range of yaw angle of micro quadrotor between two scans, there is no need to search all points in a point set when seeking a point's corresponding point, but to search the points in specific angle range.In order to make micro quadrotor fly in unknown environment, a strategy which is based on rolling windows planning and Rapidly-exploring Random Tree(RRT) algorithm to realize obstacle avoidance and path planning in unknown environment is given. Because of unknown information of global environment, global obstacle avoidance planning is regarded as local obstacle avoidance planning and simulated by repeatable local obstacle avoidance planning. Determination method of local sub-goals, which utilizes boundary points information gained from region segmentation to calculate the routing length between two regions' adjacent boundary points and uses the midpoint of the longest routing length of boundary points as local sub-goal is given.Last, an experiment data acquisition system was set up to verify indoor localization and obstacle avoidance algorithm, and indoor localization and obstacle avoidance off-line experiments were conducted. Off-line experiment results indicate that the algorithm given in this thesis can accomplish indoor localization and obstacle avoidance in unknown environment.