The Key Technology Research Of Freestyle Skiing Robot

Nowadays, the sports undertakings develops rapidly. To get good grades, not only the athletes have to take a hard sport training, but also requires a lot of science and technology’ research to support the training.In recent years, freestyle skiing aerials, which is loved by audience because of its highly ornamental. Domestic and foreign scholars have done a lot of research about it, most of them use biomechanical approach to experimental tests, and few documents relate to the establishment of the model of humanoid robot.Appeal against the problem, this paper uses freestyle skiing aerials world elite female athletes Xu Mengtao as the research object, research the impact of the athletes’ each joint angle, angular velocity, angular acceleration which on slippery stage to the athletes. Using Pro/Engineer to build a three-dimensional model of the freestyle skiing robot, and establish SimMechanics model to simulate the kinematics and dynamics. Discussing the stability criterion of the robot, and propose two stability control methods:fuzzy control and adaptive neural network fuzzy control. The specific content of the work as follows:Firstly, this paper use the idea of modular joints to build a humanoid robot model with 20 DOFs with Pro/Engineer. Based on the powerful MATLAB/Simulink, the third chapter establishe the SimMechanics model of freestyle skiing robot, which including the main torso, head, arms and legs, to analyze the kinematics and dynamics, and finally gets the theoretical curve of each joint angle, angular velocity, angular acceleration, the driving force and the driving torque about freestyle skiing robot.Secondly, by introducing the formula of the actual ZMP and expectations ZMP, this paper provides a approach to detect the real ZMP which based on the six-axis force/torque sensor. Based on the ZMP stability criterion,this paperproposed a solution which by online controlling the hip to achieve the adjustment of the desired ZMP (upper body posture) of freestyle skiing robot, and uses MATLAB fuzzy logic toolbox (FIS) to establish a fuzzy control system on freestyle skiing robot’ hip, but there is a drawback of fuzzy control:the control system has no self-learning function.Finally, through the analysis of the advantages and disadvantages about fuzzy control and adaptive neural network control, this article combine the fuzzy control and the adaptive neural network control, and uses MATLAB to auxiliary ANFIS to establish adaptive neural network fuzzy control system, to make the control system can automatically adjust the parameters of membership functions, and make the system have smaller error, and become more responsive and more stable.