Research On Robot Localization And Navigation In Emergency Events

In the activities of emergency rescue,search and rescue robots need to enter into unknown building environments to complete the search and rescue tasks.The traditional methods of robot localization and navigation cannot build the navigation map and realize the localization and navigation of the robot in a short time.In order to solve this problem,this paper proposes a quick way of building a robot navigation map based on the doorplate semantic.In order to get the map information of the accident environment quickly after emergencies happened,this paper studies how to build navigation map quickly and easily by using building fire evacuation maps which can be easily acquired.Using the technologies of image processing,the fire map is quickly transformed into a double-layer raster map that can be applied to robot localization and navigation.By extracting the information of the doorplate in the fire evacuation map,a semantic map based on the doorplate is constructed,and the semantics of the doorplate is associated with the global position.In order to solve the problem of localization difficulty caused by map imprecision,environment similarity and environment structure singleness,this paper proposes a method for robot location and navigation based on doorplate semantics.Using the image processing technology,this paper realizes the fast and accurate localization and identification of the doorplate information in the building environment.In combination with the semantic information in the map,fast localization of the robot is realized.In order to test the effect and performance of robot localization and navigation through the doorplate semantic map,a simulation environment and research platform were built respectively,in which the algorithm and theory in this paper were analyzed and optimized.The validity and practicability of this study are verified.