Research And Implementation Of Motion Control Of Quadruped Bionic Robot With Multiple Motion Modes

With the continuous development of artificial intelligence technology,the country has taken the development and promotion of artificial intelligence technology as an important strategy for national development.As a very important branch of the artificial intelligence industry chain,robots are more and more valued by the state,universities and enterprises.Robots have great needs and scenarios in medical,transportation,aerospace,military,civilian,and education.Based on the needs of the national and market for the intelligent robot industry and the research direction of the laboratory,this paper focuses on how to realize a four-legged intelligent bionic robot.In practice,the body structure of the robot,the motion simulation of the robot and the robot are carried out.Then I build the hardware and software system,and the four-legged bionic intelligent robot was explored through the trilogy.Finally,the intelligent bionic four-legged robot with cute shape,stable structure and strong system was successfully developed.In the research process of robots,it is necessary to first design the body structure of the quadruped robot according to the body structure of the tetrapod in biology,establish a simulation model,and then build a complete robot control system.The intelligent robot responds to external commands,processes the environmental signals around it in real time,receives various sensors from the body and the control signals transmitted from the outside.The brain of the robot must accurately make judgments and control the robot's motion system.Finally,the paper also realizes the landing of artificial intelligence algorithm,which enables the robot to perform visual recognition function and realize different motion modes by sensing different objects.In the research and development process of the project,I focused on the motion law of the quadruped robot.The cooperative control of 11 active freedom steering gears in the robot steering system is studied by the nonlinear oscillator-based Hopf harmonic oscillation model in the Central Pattern Generate network.Regularity,the simulation of the signal motion of the hip joint and the knee joint is realized,and the motion law of the walk state and the trot state is verified.Then the virtual prototype model is built in Automatic Dynamic Analysis of Mechanical Systems to simulate the motion mode of the robot.I set up a real four-legged bionic robot,developed a robot hardware and software system,trained the motion data in different sports modes,and realized the robot's walk,trot,back,turn left,turn right,and self-balancing after pushing down.The motion mode:by loading the six-axis sensor and then using the attitude-aware fusion algorithm to sense the robot's own pose,allowing the robot to adjust its motion pattern according to the spatial pose;use Tensorflow framework and CNN neural network model on the intelligent system of the robot The AlexNet network structure trains the visual model algorithm of the quadruped robot to realize the visual recognition function of the robot.The products of this project have entered the robot competition and artificial intelligence programming education courses of several experimental high schools in Qingdao and Jiangsu Province,as a research education and hardware platform for students to use.