Research On The System Of Environment Recognition And Mapping For Rescue Robot

Rescue robot is a kind of the robot with special function. The most important part in theresearch on this field is environment feature recognition and mapping. The accurate map isthe key to the rescue work. According to the particularity of rescue environment, mechanicalsystem and control system were designed firstly.This paper focused on the research of rescueenvironment feature recognition and mapping for rescue robot. The rescue robot made canbe used to do the experiments about the system of environment recognition and mapping.According to the feature of environment, the mechanical structure of rescue robot isdesigned. Then, the control system is designed that uses the master-slave control mode tocontro rescue robot. In order to solve the problems of the environmental informationacquisition, the sensor frame is designed and appropriate sensors are chosen. The method ofthe laser radar data processing is provided and stable platform which is used to scan is made.In order to solve the problem of the environmental identity of rescue robot, regionsegmentation and feature extraction are used to realize the extraction of the environmentinformation. An improved Split-and-Merge algorithm is presented to extract feature line fromenvironmental information more effectively. In order to solve the environment recognitionand mapping problems for rescue robot. The matching method based on feature line ispresented that improves the system speed. When rescue robot is moving, the system canupdate global map timely and accurately, at the same time the simultaneous localization isrealized.In the rescue environment, rescue robot was used to do the environment recognition andmapping so that a relatively complete environment map can be got. The experimental resultsshow that the matching method based on feature line to build map can completeenvironmental feature recognition and mapping，and that its method is of very goodpracticality and accuracy.