Research On The Dynamics And Control Algorithm Of The Multi-link Flexible-joint Robot

With the expansion of the application field of robot, the control accuracy and performance of the robot are increasingly demanded. A large number of experiments and studies show that the joint flexibility is the key factor to control the high precision of the robot. Meanwhile, the high precision control must be based on the accurate dynamic model. Therefore, it is very important to study the dynamics and control algorithm of the flexible-joint robot.In this paper, the multi-link flexible-joint robot is taken as the research object and several high precision control algorithms are studied progressively based on the dynamic model. Firstly, based on Spong's hypothesis, the dynamics equation of the multi-link flexible-joint robot where the transmission ratio is not 1 is established by using the Kane method and the virtual prototype model is established by using the equivalent ratio method in Adams, which provides the basis for the control of the multi-link flexible-joint robot.Secondly, the backstepping control algorithm is studied to eliminate the limitation of joint flexibility and the need of the acceleration feedback information. The basic backstepping control algorithm is proposed by taking full advantage of dynamic property on condition that the dynamic model is precisely available. A new kind of neural network adaptive backstepping control is proposed on condition that the dynamic model is unknown or inaccurate. The proposed control algorithm separates the nonlinear unknowns from the known terms and uses Radial Basic Function to approximate the nonlinear unknowns, and designs a control law and parameter update laws according to the Lyapunov function. Consequently, the control of the joint trajectory tracking of multi-link flexible-joint robots can be realized.Thirdly, the dynamic surface control algorithm is studied to eliminate the phenomenon of ‘differential explosion' in the process of backstepping control. The basic dynamic surface control algorithm is proposed on the basis of the backsteping method on condition that the dynamic model is precisely available. An adaptive dynamic surface control algorithm without neural network on condition that the dynamic model is unknown or inaccurate. The proposed control algorithm represents the unknown model partial as a linear form of inertia parameters and estimats the parameters, introduces the first order filter to eliminate higher order differential terms by using dynamic surface method, and designs the parameter update laws according to the Lyapunov function. Consequently, the control of the joint trajectory tracking of multi-link flexible-joint robots can be realized.Lastly, the dynamic simulation is made by using the software Adams, and the simulation results prove the correctness of the kane dynamic equation of the multi-link flexible-robot; The control schemes are constructed by using the S function to make the control simulation for the proposed control algorithms in the Simulink module of Matlab, and the simulation results show the effectiveness of the proposed algorithm; The combined simulation is carried out by using Matlab and Adams, and the simulation result shows that the backstepping can be used to the actual control and has a satisfactory effect.