Li Tiemin

Calibration method and experiment of Stewart platform using a laser tracker

The focus position of Five-hundred-meter Aperture Spherical Telescope may be changed unexpectedly because of the environmental condition such as strong winds. So a Fine Feed Cabin Model based on Stewart platform is used to get rid of the unexpected change by controlling the position and orientation of the platform. To improve the accuracy of the platform, a calibration method by a laser tracker LTD500 is presented and an error compensation experiment is carried out. The position and orientation of the platform can be figured out by LTD500. The measurement configurations can be selected from the boundary of the joint space. After calibration the accuracy can improve to about 0.2 mm and meet the requirement. The experiment result proves that the method using a laser tracker is very useful to calibrate Stewart platform, even if there are no enough measurement configurations.

Petri-net-based coordination motion control for legged robot

Legged robot generally moves according to the preplanned gait, which cannot coordinate the motions among the legs with environment. Petri-net is a graphic assembling net model that can be applied to describe and control the dynamic behaviors in multi-agent system. Considering the motion characters of legged robot, this paper proposed a single-leg Petri-net control model for legged robot, and defined the corresponding model parameters. As an example, five such models were assembled to construct the coordination motion control model for the five-legged robot we developed. On this base, the reachable marking graph including all moving states of the five-legged robot was calculated and expressed. Applying the Petri-net model, a new free gait was presented. At last, the motion of five-legged robot crossing grooves in discrete area was simulated. The results show that the five-legged robot can adapt to the environment automatically with the Petri-net control model in this paper.

Kinematic Calibration and Forecast Error Compensation of a 2-DOF Planar Parallel Manipulator

Due to large workspace, heavy-duty and over-constrained mechanism, a small deformation is caused and the precision of the 2-DOF planar parallel manipulator is affected. The kinematic calibration cannot compensate the end-effector errors caused by the small deformation. This paper presents a method combined step kinematic calibration and linear forecast real-time error compensation in order to enhance the precision of a two degree-of-freedom (DOF) planar parallel manipulator of a hybrid machine tool. In the step kinematic calibration phase of the method, the end-effector errors caused by the errors of major constant geometrical parameters is compensated. The step kinematic calibration is based on the minimal linear combinations (MLCs) of the error parameters. All simple and feasible measurements in practice are given, and identification analysis of the set of the MLCs for each measurement is carried out. According to identification analysis results, both measurement costs and observability are considered, and a step calibration including step measurement, step identification and step error compensation is determined. The linear forecast real-time error compensation is used to compensate the end-effector errors caused by other parameters after the step kinematic calibration. Taking the advantages of the step kinematic calibration and the linear forecast real-time error compensation, a method for improving the precision of the 2-DOF planar parallel manipulator is developed. Experiment results show that the proposed method is robust and effective, so that the position errors are kept to the same order of the measurement noise. The presented method is attractive for the 2-DOF planar parallel manipulator and can be also applied to other parallel manipulators with fewer than six DOFs.

Observability index of identified parameter for kinematic calibration of parallel mechanism

This paper presents a new observability index of identified parameter to quantify the selection of best configuration set in kinematic calibration of parallel mechanisms. This proposed observability index has the following advantages: (1) it is a detailed index for each identified parameters, that is, one index corresponds to one identified parameter; (2) it has a real significance: the variations of estimated value of the identified parameter; (3) it includes the information of both identification Jacobian and measurement noise. Simulation results of a three degrees-of-freedom (DOF) parallel mechanism are given to illustrate the validity and effectiveness of the proposed index.

Analysis and control of a 4-DOF manipulator to operate inside narrow and long pipes

To manipulate inside a narrow and long pipes, a 4-DOF manipulator which possesses a PRRR mechanism is designed. Geometrical method is employed to deduce the kinematic equations of the mechanism, then the condition number is obtained and a group of optimal link magnitudes is determined. Then the Newton-Euler method is used to derive the dynamic model and the drive torques of all joints are calculated. According to the dynamic model, a robust control method is proposed on the basis of linear feedback theory of non-linear systems. In contrast with traditional PID control the control method presented in this paper can better reduce the tracking error of the joints, moreover the convergence of error is faster