| Indoor Mobile Robot is the hot subject of robot technology. From a practical perspective, a newmethod of robot localization and a hierarchical path planning scenario are proposed in this thesis.Firstly, for the indoor environment usage, a new mobile structure and its driving method aredesigned to accomplish “six-direction parallel movement”,which settles the troubles that confrontrobots while rotating in a narrow space; And the robot’s movment is calculation based on the linerfeature of the ground, realizing the relative-localization; Via bringing in the infrared laser andmonocular vision technology, a distribute smart PTZ(Pan/Tilt/Zoom) system is designed, wirelesscommunication between the system and robot enables the cooperation of localization task, whichdiminishes the instability, realizing interactive absolute-localization.Secondly, based on Mahalanobis distance, the classical consensus data fusion algorithm isextened for the multi-dimensional data integration usage. By compositing an outlier elementsdetection method, the data fusion scheme turns more completed, and then used to fusion the datumfrom different position detecting systems.Thirdly, based on the OpenCV2.0platform, robot’s color flag is recognized via calculating thebhattachryya distance between color distributions of figures; On the basis of current data filterprocessing algorithm, the thesis realizes the dynamic target vision localization using EK/PF, and thefusion of the movment calculating data improves the localization result; Discusses the process thatthe robot builds hierarchical set of map with the cooperation of distribute environment smart.In the end, the FMM algorithm is improved based on the voronoi chart, a hierarchical pathplanning scenario is designed to accomplish the global optimum path searching task, the simulationexperiments are carried out on platform VC++6.0. The local path is determined by the rollingwindow algorithm in real time, and the process is simulated on the platform Matlab7.10. |
PDF Full Text Request