The Environmental Perception Of Unmanned Platform Based On The 2D Laser Radar

With the acceleration of modernization, the unmanned platform is playing an increasingly important role, and the environment perception technique is one of the most significant techniques for the unmanned platform. Therefore, this paper presents a solution for the obstacle detection and the location problem of the unmanned platform based on the experimental platform, which involves Voyager Ⅱ testing car, 2D laser radar and inertial sensor integrated module.For the obstacle detection, the obstacle detection method in this paper is based on 2D laser radar data. Firstly, the data is pre-processed by the threshold filtering algorithm and the improved median filtering algorithm, and then the size of the data is richened by the context of the laser radar, finally, the obstacle is identified by the clustering algorithm. At the same time, this paper proposes to use the convex hull algorithm to solve the redundant clustering information in the clustering process, which increases the robustness of the system.For the localization problem of the unmanned platform, this paper firstly establishes the motion model of the unmanned platform, and then the factors that probably affect the accuracy of the location are determined, furthermore, the location error of the unmanned platform is predicted by the IF-THEN fuzzy model. At last, the location of the unmanned platform is optimized by the feature data of the 2D laser radar. The data result shows that the localization algorithm based on fuzzy prediction and 2D laser radar feature data is better than the localization algorithm based on the information of the inertial sensor integrated module.For the hardware and software of the unmanned platform, this paper takes the Voyager Ⅱ testing car as the travelling platform, obtains the environmental perception information through SICK 2D laser radar, receives the motion information from ASV940 inertial sensors integrated module, processes the data on the mobile computer. In terms of software, this paper realizes the laser radar data acquisition function interface, real-time control interface of the testing car, information display module interface of the inertial integrated module and algorithm selection interface. In the process of the experiment, the algorithm proposed in this paper is verified by the dynamic and static environment. Experiments show that the performance of CBFD(Clustering by finding peak density) is better than K-means clustering algorithm and ISODATA clustering algorithm(Iterative self-organizing data analysis technique) in both two kinds of environment.