R.A. Volpe, Real and Artificial Forces in the Control of Manipulators: Theory and Experimentation, Ph.D. Thesis, Dept. of Physics, Carnegie Mellon University, Pittsburgh, Pennsylvania, September 1990.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....checks in high density circuit elements. Volpe and Khosla [34] experimentally analyzed force control strategies for direct drive robot arms, and concluded that integral force control is the best choice for explicit force control, and their result should apply to DD fingers as well. Volpe [35] also surveyed prior force control work on direct drive manipulators. Motor manufacturers offer limited angle torque (LAT) motors, which generate torque through a limited range of angles and operate at speeds and torques comparable to our mini DD actuators [12] The basic components of a LAT and ....

R. A. Volpe,Real and artificial forces in the control of manipulators: Theory and experiments, PhD thesis, Carnegie Mellon University Robotics Institute, Department of Physics, September 1990.

....compliance. The force feedforward term may or may not be used. 2 The term, mobility, here is not the same as the one used in association with locomotion. The mobility matrix is defined as a reciprocal of the inertia matrix of the system [100] 3 For this part of review, the author owes to [105, 106]. 1. Introduction 12 2. Hybrid position force control [111] This combines conventional position control and explicit force control both of which can be implemented simultaneously in orthogonal directions along the tool coordinate axes. Some instability problems, however, were reported by An ....

....Task 87 near the imaginary axis are due mainly to the environmental dynamics. Enlargement of the map near the imaginary axis is shown in Figure 5.10 which excludes the left most pole. As far as the locations of poles and zeros are concerned, the model is somewhat similar to the one used in [106], but not to [105] which drew a misleading remark in terms of stability of proportional force control although all the three use exactly the same fourth order model, namely Equation (5.3) Stability of the system is evaluated under three different force control schemes, a proportional control, an ....

R. A. Volpe. Real and Artificial Forces in the Control of Manipulators: Theory and Experiments. PhD thesis, Department of Physics, The Robotics Institute, Carnegie Mellon University, Pittsburgh, PA, September 1990. Bibliography 165

....Communication . 19 5.2. Periodic Communication . 20 6. Future Work . 21 7. Acknowledgments . 22 8. References . 23 List of Figures Reusable software control modules within a reconfigurable system . 2 Software framework for reconfigurable systems . 4 Typical target hardware for a reconfigurable sensor based control system . 6 Simple model of a port based object, ....

....a reconfigurable system. M re details of a C language specification using abstract data types for an SAI interface has been defined in Chimera and is given in [19] SAI objects have already been written for several sensors and actuators, including the Puma robot, Direct Drive Arm II Robot [7] 23] CMU Reconfigurable Modular Manipulator System [14] both Lord and Assurance Technology force torque sensors, Dimension 6 trackball, and an Imaging Technology vision subsystem. 4.6.3 Special Purpose Processor Interface and Drivers Most operating systems attempt to treat all hardware, including ....

[Article contains additional citation context not shown here]

R.A. Volpe, Real and Artificial Forces in the Control of Manipulators: Theory and Experimentation, Ph.D. Thesis, Dept. of Physics, Carnegie Mellon University, Pittsburgh, Pennsylvania, September 1990.

....A x B Figure 1: General fourth order model of the arm, sensor, and environment system. In order to have a basis for the experimentation presented herein, a fourth order plant model has been hypothesized, based on an assumption that the first mode of vibration is dominant in the system components [9, 3]. This model is shown in Figure 1, where f is the actuation force; xA is the measured position of the arm; xB is the position of the environment; and m, k, and c are the mass, stiffness, and damping parameters. This is similar to the model presented in [4] However, the following analysis will ....

....for f m is to solve Equation (4) for p exactly, and substitute these solutions into Equation (1) to obtain the relation between xA and xB . However, this brute force method would provide an algebraically complicated solution, yielding little intuitive insight. Instead, we have previously shown [9, 10] that the transfer functions for xA=xB given in Equations (1) can be used in Equation (8) to obtain an approximation of the measured force in terms of only xA (and its derivative) f m Gamma k 3 ff xA c 1 Gamma c 3 c 2 k 3 k 2 ff xA (9) GammaK 0 xA Gamma C 0 xA (10) ....

R. Volpe. Real and Artificial Forces in the Control of Manipulators: Theory and Experiments. PhD thesis, Carnegie Mellon University, Department of Physics, September 1990.

....employed, as will be shown. 4. 2 Manipulator Model Based Control Manipulator model based control involves the use of a dynamic model of the manipulator to determine the actuation torques, Bejczy (1974) Model based impedance control may be summarized by the following equations [Hogan (1985) Volpe (1990)] D( m )u h( m ; m ) g( m ) J T ( m )f m (12) u = J Gamma1 ( m ) h x u Gamma J( m ) m i (13) x u = M Gamma1 [C Delta x K Deltax Gamma f m ] 14) Deltax = x c Gamma F( m ) 15) Delta x = x c Gamma J( m ) m : 16) Equation (12) compensates for ....

Volpe, R. 1990 (September). Real and Artificial Forces in the Control of Manipulators: Theory and Experiments. PhD thesis, Carnegie Mellon University, Department of Physics.

....intrinsically require that the manipulator be force controlled. Two main approaches have been proposed for this purpose: explicit force control and impedance control. Since it has been demonstrated previously that an impedance controller with force feedback contains an explicit force controller [20, 23], the discussion in this paper is limited to explicit force control. Explicit force control involves the direct command and measurement of force values, with the goal of having the output follow the input as closely as possible. Two types of explicit force control have been proposed: force based, ....

....Inner position loop controllers, as the name suggests, have an outer force control loop that provides position commands to an inner position based controller. While many of these controllers have been analyzed before, this has not been done with an experimentally determined plant transfer function [20, 21]. As will be seen, erroneous conclusions about the stability of the system can result without a specific system model. Further, the analysis in this paper draws the force and position based strategies together, into one coherent framework. This framework provides greater understanding of how gain ....

[Article contains additional citation context not shown here]

R. Volpe. Real and Artificial Forces in the Control of Manipulators: Theory and Experiments. PhD thesis, Carnegie Mellon University, Department of Physics, September 1990.

....for impact control will be reviewed in this section [65] Previous research in force control has treated the impact phase as a transient that is dealt with by the same controller used to follow commanded force. The form of the force controller is typically an explicit force or impedance controller [70, 56]. In this section it will be shown that the best implementation of these strategies for force following is insufficient for impact control. Figure 3: The avoidance potential for a rectangle with ff = 1 and A = 1. Large values have been truncated. Figure 4: The approach potential function for a ....

....the manipulator s position measurement devices. In this case, no velocity will be sensed. Second, for fast oscillations the calculated velocity signal will lag its ideal value, and the damping force may cause instability by being applied out of phase with the true velocity of the surface [56]. Third, flexion in the links due to impact can slightly change the arm structure, thereby making the kinematics and velocity signal computation erroneous. These problems are compounded by the fact that a stiff environment which causes them will also cause a larger impact force and need stablizing ....

[Article contains additional citation context not shown here]

R. Volpe. Real and Artificial Forces in the Control of Manipulators: Theory and Experiments. PhD thesis, Carnegie Mellon University, Department of Physics, September 1990.

....of force control methods. This testing has revealed the relative efficacy of each control strategy, as well as the computational considerations necessary for the algorithms which implement each strategy. In this paper we first review the wide spectrum force control strategies that we have tested [29, 35]. Second, the computational requirements of each strategy will be presented. Third, we describe the architecture of the CMU DD Arm II system, and discuss the implementation considerations which limit the effectiveness of some schemes. Fourth, we present some experimental results from our ....

....in these computations. Finally, we will present some experimental results obtained, and discuss how computational costs have effected these results. 2 Explicit Force Control Strategies This section describes the forms of explicit force control that were implemented and experimentally evaluated [29]. Explicit force control strategies utilize direct evaluation of desired and measured forces. Two types of output are possible from this type of controller: forces or positions. In the former, forces are commanded directly, and then translated into manipulator joint torques. In the latter, ....

[Article contains additional citation context not shown here]

R. Volpe. Real and Artificial Forces in the Control of Manipulators: Theory and Experiments. PhD thesis, Carnegie Mellon University, Department of Physics, September 1990.

....Pittsburgh, Pennsylvania, 15213 Previous research in force control has treated the impact phase as a transient that is dealt with by the same controller used to follow commanded force. The form of the force controller is typically an explicit force or impedance controller [Whitney (1985) Volpe (1990)] In this paper it will be shown that the best implementation of these strategies for force following is insufficient for impact control. But the impact controller presented here still fits into the same framework. To understand this, the previous schemes will be briefly discussed and their ....

....the manipulator s position measurement devices. In this case, no velocity will be sensed. Second, for fast oscillations the calculated velocity signal will lag its ideal value, and the damping force may cause instability by being applied out of phase with the true velocity of the surface [Volpe (1990)] Third, flexion in the links due to impact can slightly change the arm structure, thereby making the kinematics and velocity signal computation erroneous. These problems are compounded by the fact that a stiff environment which causes them will also cause a larger impact force and need ....

[Article contains additional citation context not shown here]

Volpe, R. 1990 (September). Real and Artificial Forces in the Control of Manipulators: Theory and Experiments. PhD thesis, Carnegie Mellon University, Department of Physics.

....is shown in Figure 5, where H = K fp and G is the arm sensor environment plant and includes the active damping control loop. It has previously been shown that G can be considered a fourth order transfer function [3] and we have experimentally extracted parameter values for the model [16, 18]. The closed loop transfer function with the feedforward term is: F c Fm = 1 K fp )G 1 K fp G : 2) This is a Type 0 System and will have a nonzero steady state error for a step input. The root locus of this system is shown in Figures 6 and 7. The corresponding Bode plots are shown in Figure ....

....from the environment dynamics at approximately 100 rad s. After this peak there is a 40 dB decade drop off which gives a minimum phase margin of 15 ffi at K fp 1. The addition of a lowpass filter in the feedback loop can improve the response by introducing a dominant pole on the real axis [1, 16]. However, this pole placement and the behavior of the system closely match that provided by integral control. A discussion including integral force control is outside the scope of this paper [16] Therefore, lowpass filtering will not be considered further. It will prove useful later (in the ....

[Article contains additional citation context not shown here]

R. Volpe. Real and Artificial Forces in the Control of Manipulators: Theory and Experiments. PhD thesis, Carnegie Mellon University, Department of Physics, September 1990.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC