Initiative Exploration And Feature Maps Building In Unkonwn Environment For Indoor Robots

With the development of society, the service robots have entered family gradually. Exploring unknown interior environment independently and building maps are the basic functions of the service robots, so it is inevitable to be the research focus and the first step to realize autonomy. For the unknown interior environment, this paper carried out the following works:Based on the RRT, the new method is proposed, which is suitable to build maps for the indoor service robots. According to the indoor environment, the extraction methods of the feature are improved on the framework of point line feature maps, the new angle feature extraction method is proposed. and the extraction of dynamic factors is realized, finally, the complete active exploration and map construction system are builtThis paper analyzes the service robots platform from the three angles of the the motion model, sensor model and the odometer model, and illustrates the problems that should be paid attention to in the later works.According to the limitations of RRT in the static environment, the HSRRT algorithm is put forward for the indoor service robots in order to avoid the blindness of RRT, which is combined RRT with boundary heuristic path planning, and adds the thoughts of surrounded searching and the strategy of avoiding the dynamic factors, then its feasibility is proved by the simulation results.According to the difficulties of the extraction of corner feature points, the gradually dichotomy is put forward on the framework of point line feature maps. The Ransac algorithm is added into the process of extracting straight lines to ensure the precision of the straight line features, the precise maps are built, and dynamic factors are separated. Finally, the map construction system is completed by reducing storage space.The HSRRT algorithm is verified availability and efficiency in building maps through the experiments. Compared with the three prediction, the gradually dichotomy is proved practicality, high accuracy and repeatability. The new method is proved accuracy by the comparison of the experiment of the RANSAC line fitting. The MbICP algorithm can replace the odometer, which is verified by comparison of the maps based on the Odometer method and the MbICP algorithm. The experiments prove that the dynamic factors are excluded in building maps.