Manipulators (Mechanism)

A user friendly graphical interface was developed to control a Stewart platform which is a six degree-of-freedom in-parallel mechanism. The interface allows the user to define the platform motion relative to various coordinate systems: base, platform and joint. The velocity/position reference to the platform's controller can be provided by the following ways: preprogrammed data file, serial communication RS-232, 6 degrees of freedom joystick and soft teach pendant. The platform was designed to be used as "Space Emulator" and therefore a 6 degrees of freedom force/torque sensor was needed. Two different models of such sensors were designed and analyzed using finite element analysis techniques. Based on the results one particular model was selected, fabricated, instrumented with strain gages and calibrated in order to obtain its stiffness matrix. The effect of drifting of the sensor output due to self heating of the strain gages and the electronic components of the strain gage amplifiers was also studied.

Traditional industrial manipulators possess fixed configuration and are widely used in manufacturing application in which the manipulator base is fixed. However, some applications exist which would require the robotic manipulators to function in non-stationary environment especially in space. In this thesis, a six degree of freedom parallel-series hybrid manipulator is described. It consists of a 3 d.o.f. in-series manipulator mounted on a 3 d.o.f. in-parallel manipulator. A compatibility equation is found to govern the relationship between in-series component angular velocity and linear velocity; a constraint equation is added to the Jacobian of in-parallel component. Using these two equations, a decomposition strategy is proposed for solving the inverse velocity problem of the hybrid manipulator together with the simulation examples of inverse position tracking and straight line trajectory planning. Effectiveness of this method and factors affecting the simulation result are examined.

A Stewart platform is a six degree of freedom robot manipulator with its six links arranged in a parallel configuration. A dynamic model for the plant of each link, which consists of an amplifier, an electrohydraulic servo valve, and a hydraulic actuator, is found from open-loop step and frequency responses. To determine a model for the complete closed loop system, integrators located in the link input and feedback paths were added to the plant's model. PID controllers were designed to increase the system's bandwidth. Once control of the individual links was achieved, control algorithms were developed to control the motion of Stewart platform. The algorithm would move the platform through its initialization sequence, then control platform velocity, as dictated by the user.

The I.B.M. Electric Drive Robot (E.D.R.) is a six-link manipulator originally controlled by a classical analog P.I.D. controller. Its performance is not satisfactory because of its poor tracking capabilities and a considerable vibration during arm movement. This is the central motivation for designing an adaptive computed torque controller for this system. In order to accomplish this the physical model of the robot is first reparameterized such that it is linear with respect to a set of uncertain parameters. Once this is accomplished the adaptive controller is then formulated. Next methods of computer simulation are developed and employed. These simulation results show the superior performance of the proposed scheme over both a classical computed torque controller and the current P.I.D. controller.

In this thesis, a novel two-dimensional learning control scheme for robot manipulators is proposed. The convergence of the scheme for a general n-degree of freedom robot is shown. In the next part of the thesis, an algorithm for the approximation of a two-dimensional causal, recursive, separable-in-denominator (CRSD) filter, using the impulse response and autocorrelation data, is presented. The stability of the designed filter is discussed and it is shown that the approximated filter is always stable. The simulation results for the approximation technique as well as the two-dimensional learning control scheme are also included in the thesis.

Central to many manipulator positional control schemes is a requirement to invert the forward kinematic equations which model the given manipulator. It is shown in this thesis that for manipulator types where a common wrist center exists, a simplified Jacobian form is feasible and its inversion can be used in place of inverse kinematic solutions for positional control. The Jacobian simplification is obtained by decoupling of the wrist member from the positional member, resulting in a Jacobian inversion involving the solution of two sets of three equations with three unknowns. Within the development of the alternate Jacobian form, a technique for substituting incremental rotations with incremental translations is introduced yielding better insight into the Jacobian structure. A requirement for small moves is validated with a discussion of a proposed positional control strategy and a comprehensive example is presented as a summary of the results.

This thesis presents several algorithms to treat the problem
of closed-loop near minimum-time controls with
discontinuities. First, a neighboring control algorithm is
developed to solve the problem in which controls are bounded
by constant constraints. Secondly, the scheme is extended
to account for state-dependent control constraints. And
finally, a path tracking algorithm for robotic manipulators
is presented, which is also a neighboring control algorithm.
These algorithms are suitable for real time controls because
the on-line computations involved are relatively simple.
Simulation results show that these algorithms work well
despite the fact that the prescribed final points can not be
reached exactly.

The concept of "generalized manipulator" is introduced, and
the closed form and recursive form dynamical models of the
generalized manipulator are presented in Newton-Euler
formulation. The physical meaning of each term in the
dynamical model is explained.
The dynamical models formulated by the Newton-Euler method
and the Lagrangian-Euler method are proved equivalent. The
dynamical model of the generalized manipulator is reduced to
ordinary manipulators. The reduced dynamical model is shown
identical to existing models. Furthermore, the reduced
dynamical model of the generalized manipulator can be used
to compute forces and torques components along any
direction.
Application of the model to problems of mobile robots and
flexible manipulators is shown.

Robot calibration is a software-based accuracy enhancement process. It is normally implemented in a well-controlled environment. However, for a system that function in a natural environment, it is desirable that the system is capable of performing a calibration task without any external expensive calibration apparatus and elaborate setups, i.e., system self-calibration. Vision systems have become standard automation components as cameras are normally integral components of most robotic manipulators. This research focuses on camera-aided robot self-calibration. Unlike classical vision-based robot calibration methods, which need both image coordinates and precise 3D world coordinates of calibration points, the self-calibration algorithms proposed in the dissertation only require a sequence of images of objects in a natural environment and a known scale. A new robot self-calibration algorithm using a known scale at every camera pose is proposed in the dissertation. It has been known that, the extrinsic parameters of the camera along with its intrinsic parameters can be obtained up to a scale factor by using the corresponding image points of objects due to the factor that the system is inherently under-determined. Now, if the camera is treated as the tool of the robot, one is then able to compute the corresponding robot pose directly from the camera, extrinsic parameters once the scale factor is available. This scale factor, which changes from one camera pose to another, can be uniquely determined from the known scale at each robot pose. The limitation of the above approach for robot self-calibration is that the known scale has to be utilized at every robot measurement pose. A new algorithm is proposed by using the known scale only once in the entire self-calibration procedure. The prerequisite of this calibration algorithm is a carefully planned optimal measurement trajectory for the estimation of the scale factor. By taking into consideration of the observability of the link error parameters, the problem can be formulated either as a constrained or a weighted minimization problem that can be solved by an optimization procedure. A new method for camera lens distortion calibration by using only point correspondences of two images without knowing the camera movement is described in the dissertation. The images for robot calibration can be shared for lens distortion coefficient calibration. This characteristic saves the user much effort in collecting image data and makes it possible to conduct a robot calibration task on line. Extensive simulations and experiment studies on a PUMA 560 robot at FAU Robotics Center reveal the convenience and effectiveness of the proposed self-calibration approaches. Compared to other robot calibration algorithms, the proposed algorithms in this dissertation are more autonomous and can be applied to a natural environment.

One emerging application of parallel manipulators is to use them as CNC (Computerized Numerical Controlled) machine tools and recently several prototypes of such CNC machines have been developed, all based on hexapod machine--a type of parallel manipulators similar to a Stewart platform. The goal of this research is to develop an effective control scheme, cross-coupling control, for this type of CNC machine tools, which will reduce the contouring errors and thus further enhance their advantages. This dissertation describes the research work as follows. Firstly. based on the analysis of the kinematics and dynamics, a PID (Proportional-Integral-Derivative) controller was designed for each leg of the hexapod CNC machine. Secondly, real-time contour error models were developed and verified to determine not only for the calculation of the contour errors of the hexapod CNC machine but also for the general case of any machine tools. Thirdly, the contour errors of the hexapod CNC machine were investigated for a conventional PID controller. The results indicate that the accuracy of the hexapod machine is better than the conventional CNC machine tools even for mismatched axes and load exertion. Finally, a cross-coupling control scheme was proposed for the purpose to enhance the contour accuracy of this new type of hexapod CNC machine tools. A cross-coupling controller design for a 2-DOF platform was performed to provide the guidelines. Then, a cross-coupling controller for the new type of hexapod CNC machine tools was designed by feeding back the contour error to each axis. The efficiency of the proposed cross-coupling controller was verified through simulations. The result shows that the proposed cross-coupling controller is very effective in reducing the contouring errors. While cross-coupling controllers were originally proposed for conventional CNC machine tools, this research is the first attempt of expanding this concept to the new type of hexapod CNC machine tools.