Research On Design And Navigation Methods Of An Indoor Mecanumwheeled Mobile Robot

The aged people support problem is exposed as China is gradually stepping into the aging society.The support of the aged supplies an enormous market for service robot field.Also it is a big challenge for human-robot interaction.Although indoor robot navigation has been studied for years,it is not available in family environment.Because the human-beings and robots have different understanding about environment,the human-robot interaction is difficult to be achieved.In consequence,the semantic navigation is getting hot in recent years.In dynamic indoor environment,the obstacle avoidance is also important for a mobile robot.Firstly,the Mecanum-wheed mobile robot system is designed.In mechanical part,the robot use the independent suspension to adopt the changeable road.The mecanumwheed structure can achieve 3-DOF motion control to pass the narrow space in indoor environment.The self-designed motor driver module supply the power for each wheel.The battery management system(BMS),laser scan and Kinect sensor make up the 3D sense system of environment.Based on Robot Operating System(ROS),the motion control and navigation is achieved.Secondly,the paper proposed kinematics model of mecanum-wheeled robot and establish the environment map of semantic and grid.Considering the omni-directional moving performance,it is different from differential-wheeled robot in drive control and odometry model.The laser scan model is established to correct the sensor data.The semantic toplogical map is established to serve the semantic global navigation.And the 2D grid metric map is calculated to maintain the basic partical navigation in houseland.Thirdly,the Mento Carlo Localization algorithm and A* path planning algorithm are extended to available in mecanum-wheeled robot.In initialization of Mento Carlo Localization,the distrubition of particles is pre-located to semantic area.So as to reduce the time of iteration and reduce the probabilistic of wrong matching.In semantic global navigation,the connect area is proposed to enhance the pass-abality of robot in narrow area.Considering the dynamic environment,the A* algorithm is used to accomplish the second time regulation of routes.Finally,taking some experiments to verify the correctness of above algorithm.The experiments includes the stability test of robot,the semantic navigation test,the Mento Carlo Localization algorithm test and the dynamic A* algorithm test.The successful experiments can verify the effiency and stability of the algorithm.