Simultaneous Localization And Mapping And Experimental Research On Amphibious Hexapod Robot

With the rapid development of the ocean,the research on coastal defense,military deployment and reconnaissance and communications in complex environments such as land-sea junction and shoal has become a hot spot.Traditional robots are often only capable of working on land or underwater.The amphibious environment is complex.On the land are often soft gravel and rugged rocks.Underwater is full of plankton and undercurrent,all of which are critical to robot performance put forward higher requirements.Legged-robots fall off feet enough to be able to cross obstacles like an animal,as is evident by the superiority of soft and rough terrain.In this paper,an amphibious hexapod robot is developed with the aim of amphibious homework,and SLAM theoretical analysis and experimental research are carried out to lay the foundation for robotic autonomy.After analyzing the working environment and the contents of the robot,this paper presents a highly efficient dual joint-driven amphibious six-legged robot with the analysis of the robot gait as the starting point,including the structure of the robot leg,fuselage and drive joint Design scheme and a scheme of three-layer control system.The robot leg is composed of a designed six-circular arc cam mechanism with a constant width,a sliding gyro and a rotating four-bar mechanism.A gait is planned by changing the phase difference of the legs,The main parameters of the transmission and the structure of the fuselage structure,the analysis of the leg joint torque,the design of a small submerged drive joint and key components selection;planning robot control system structure and the whole three-dimensional modeling.Kinematic analysis and simulation verification of the walking foot of the robot are carried out.The trajectory of the foot end is obtained by simulation and the parameters of the footing are analyzed.According to the diversity of the working environment and the tasks of the robot,a joint control method is planned,a mathematical model of the robot joint is established,The fuzzy PID controller shows that this method has more superior control performance than the traditional method.The robot kinematics model and the environment observation model are established to provide the model basis for the EKF-SLAM research.Based on the extended Kalman filter algorithm,the SLAM system is modeled by therobot,the system state is predicted under the point feature environment,the features of environment points are extracted,the point feature data is related based on the nearest neighbor data association method,and the system state is observed by using the observed landmarks Updating and augmenting,using Matlab to simulate EKF-SLAM,analyzing the influence of the number of environmental road signs on robot positioning and construction errors,extracting the environmental features based on the least squares method,adopting NNDA method for feature matching,The EKF-SLAM line characteristics were studied,and the data extracted by lidar were used to verify the line feature extraction and matching algorithm.Completed the development of amphibious hexapod robot prototype and robot experimental control platform,carried out the straightforward robot and load experiment,verified the rationality of structural design,in order to verify the robot's position tracking performance,the steering and position control experiments were carried out to build the experimental environment,Carried out SLAM experiments of six-legged robot under different signpost environment,and finally Experiments were conducted on sandy beach and grassland to verify the performance of the robot under actual conditions.