Random Trajectory Tracking Control For A Class Of Manipulator Systems

For a class of manipulator systems in the random vibration environment,the problems of stochastic modeling and trajectory tracking control are considered in this paper.Firstly,the stochastic noise is reasonably introduced into the system to establish the dynamic model of the stochastic manipulator system.Then on the basis of some reasonable assumptions,two kinds of different tracking control strategies are given for the cases in which the velocity is unmeasured and the parameters are unknown.Finally,the control strategies are applied to specific models.The effectiveness of the control strategies are verified by simulation experiments.(1)Modeling of random manipulator To design the controller,the stochastic manipulator system is modeled.First based on the Lagrange mechanics methods,the deterministic models of manipulator systems are established.Then by the principle of dynamic-static method and relative movement principle,the influences of random disturbances in the environment along link transformed into interferences with force or moment,so the random dynamic models of manipulators are builded.(2)Output feedback tracking control When the velocity is not measurable in the random manipulator system,an output feedback tracking controller is designed.Firstly,a novel high-gain observer is constructed by introducing internal dynamic variables.Then the output feedback tracking controller is designed to make the state of the closed-loop system has an asymptotic gain in the first moment.The effectiveness of the output feedback control strategy is illustrated by a random two-link manipulator model.(3)Adaptive tracking control The mass of the object grabbed by the manipulator and the air resistance coefficients are unknown under the random surroundings,an adaptive tracking controller is designed.Firstly,the unknown parameters are separated from the coupling terms by the separation principle.Then an adaptive tracking controller is designed such that the closed-loop system is the noise-to-state-practically stable in probability.Taking random M1 manipulator as an example,the rationality of adaptive control strategy is further illustrated.