Research On Localization And Object Grasping Of Indoor Service Robot

With the improvement of people life quality and the increasing demand of the society for robots,it has become an trend to develop and research the technologies of service robots.The accuracy of localization and the ability of object recognition are the key technical indicators to measure the stability and efficiency of service robots in indoor environment.Therefore,the service robot is designed,and the mapping and localization are researched,at the same time,the objects grasping based on depth camera is researched.According to the working scene and demand of task,the system of service robot is designed.According to the characteristics of the indoor environment,differential mobile platform and the mechanical arm are constructed.The characteristics of the mechanism are analyzed,the control system scheme is determined,then the required hardware modules are selected.And the communication protocol of the control system is designed.The mapping and localization technology are researched,and an improved Monte Carlo localization algorithm is proposed.the kinematic and the odometer model of the robot is established.According to the structural characteristics of the manipulator,the kinematics of the manipulator is analyzed to lay the foundation for grasping the object.By comparing the Hector SLAM and Gmapping algorithm,the better method is choosed.Aiming at the problem of particle degradation in monte Carlo localization algorithm,the diversity of resamped particles is improved by adding optimization factor,and the improved Monte Carlo localization algorithm is verified by the simulation experiment.The objects grasping technology based on depth camera is researched.The relationship between the coordinate systems is established.The internal parameter and the hand-eye calibration parameters are obtained.An object recognition and location method based on the improved ORB(Oriented FAST and Rotated BRIEF)algorithm is proposed by using K-D tree.The experimental result show that the optimized algorithm can not only reduce the number of feature points,but also improve the matching accuracy,and use the results of feature matching to calculate the 3D coordinates of the target object.Based on the above work,the robot experiment platform is assembled and debugged,and complete its performance test,the experimental results show that the robot meets the requirements.The mapping and localization experiments are carried out by using the robot platform.And the object recognition and grasping experiments are completed.The experimental results show that the robot can achieve accurate localization and complete the object recognition and grasping in the indoor enviroment.