Research On The Construction Method Of Indoor Dynamic Environment Map

As one of the representatives of intelligent wave,mobile robot has been gradually applied to many fields including factory transportation,medical treatment,service and so on.Due to its complex and changeable working environment,mobile robot has been given higher requirements.As the basis of mobile robot motion and path planning,its self-positioning and reconstruction of the surrounding environment map have become one of the current research hotspots.At present,both at home and abroad for complex indoor environment of mobile robot localization and map reconstruction technology remains to be improved in accuracy and practical aspects,especially how to use a single sensor for the positioning of the mobile robot and the reconstruction of the surrounding environment map,or a weak adaptability to dynamic information,or build environment map is too thin,or function of map building in cost is larger,can not meet the requirements of action guide mobile robot,which greatly reduces the mobile robot is practical,therefore have important research value.In this paper,the mapping problem of mobile robot based on single sensor in complex indoor environment is studied as follows.By analyzing the working environment of the indoor mobile robot,the area is relatively small,the ground is relatively smooth and even some ground is too smooth and lack of texture,and there are many dynamic objects,etc.For the mobile robot working in the dynamic environment,the corresponding dynamic information filtering algorithm and the construction method of dense passable area map are designed.First compares the basic development system and the selection,selection of the ORB-SLAM2 system as the basis of system was improved,the environment of dynamic information design the dynamic information filtering algorithm based on local window,the current frame image respectively with a frame and reference frame image matching,to determine the dynamic area and the dynamic area information filtering,have pure static environment information,realized the function of dynamic information filtering and validated.Secondly,dense and regional map construction algorithm is designed algorithm based on key frames,first is to put the system to get pure static environment point cloud information in the separation of point cloud on the ground,to get preliminary ground plane fitting,then to screening of key frames,will be the key frames after the screening surface area of the break up,after the split end projection to the ground plane fitting,get the final dense ground passable regions point cloud map and validated.Finally,the overall feasibility of the algorithm is verified by the experimental platform and the influencing factors are analyzed.It is proved that the improved system can construct a dense passable regional point cloud map in the environment with the intervention of dynamic objects and has excellent practical performance.The realization of dense map construction of passable areas in complex indoor environment makes mobile robot's cognition of the surrounding environment more clear and accurate,which has great reference value for mobile robot's path planning,navigation and other functions,and improves the practicability of map reconstruction.