Research On Dynamic Modeling And Brachiating Control Of Primate Bionic Robot Under Elastic Support Condition

With the emergence of a series of problems in space,navigation and other applications,underactuated systems have become one of the hot topics of control theory.Underactuated system is a kind of nonlinear system with independent control variables less than the number of system degrees of freedom,which is superior to complete drive system in energy saving,cost reduction,weight reduction and system flexibility.Simply put,that is,less than the amount of input to control the system.Therefore,the study of underactuated control system is of great significance.In this paper,the underactuated system of primate-like bionic robots under elastic support is studied.The system is derived from the underactuation of acrobot,with more number of underactuated degrees of freedom,moreover,and with the introduction of elasticity,dynamic modeling and control of the system have become more complicated.The study of such problems will help to extend the application of under-actuated systems.In this paper,the vibration of the cantilever beam is simplified by two orthogonal spring dampers.First,the Lagrangian equation is used to establish the dynamic model of the primate bionic robot under elastic support.The dynamic model of the system is built in Simulink.Then,a simple motor torque input is used to observe the change of q1 and q2 and verify the correctness of the deduced dynamic model.Based on the Lyapunov stability theory and the partial feedback linearization method,the controllability of the vertical equilibrium point under the elastic support constraint is analyzed.By setting the LQR linear controller,it is found that the system is observable at the vertical equilibrium point Not completely controllable.In this paper,genetic algorithm is used to control the primordial bionic robot under elastic support.A simulation model based on genetic algorithm is set up.The simulation results show that the dynamic model and genetic algorithm control strategy Effectiveness.In order to optimize the control method,the fitness function is the sum of the distance from the end point of the pole 2 to the target tree branch and the energy consumed in the process.The experimental results verify the effectiveness of the control strategy.Then,the virtual constraint is introduced in the process of cantilever swing,and the model is controlled by LQR linear controller,so that the linear state is maintained between the rod 1 and the rod 2 during the swing.Finally,the experiment verifies the correctness of the virtual constraint.Finally,a semi-simulation platform is developed,which includes the mechanical system,the electronic control system and the software system.The computer acquired sensor information through PCI-6014 data acquisition card.Real-time simulation module is used to control the system.The simulation results show that the control model is correct.